[ { "path": ".agent/sync_async_type_hints_overload_guide.md", "content": "# Agent Overload Implementation Guide\n\n## Purpose\nThis document provides instructions for implementing `@overload` signatures for redis-py command methods.\n\n## How to Use This Guide\n\n### Step 1: Select a Batch\nWork on one batch at a time. Each batch contains methods grouped by command class.\n\n### Step 2: For Each Method in the Batch\n1. **Verify the return type** by checking Redis docs at https://redis.io/commands\n2. **Add two `@overload` signatures** before the method when the sync and async return types differ:\n - One for sync (`self: SyncClientProtocol`) returning the sync type\n - One for async (`self: AsyncClientProtocol`) returning `Awaitable[sync_type]`\n - If both clients return the same concrete type, keep a single typed method instead of adding redundant overloads\n3. **Mirror the full input signature exactly** in both overloads:\n - Same parameter names, order, defaults, positional/keyword shape, `*args`, and `**kwargs` as the implementation\n - Use the same modern annotation style as return types: prefer `X | Y` and `T | None` over `Union[...]` / `Optional[...]`\n4. **Keep the original implementation** unchanged except for signature annotation normalization when needed\n - For implementation return unions, prefer the readable order `SyncType | Awaitable[SyncType]` (for example `(dict | None) | Awaitable[dict | None]`)\n - Runtime decorators such as deprecation or experimental markers must stay on the real implementation, not on `@overload` stubs. Put those decorators immediately above the implementation `def` after the overload block.\n5. **Run type checker** to verify no errors\n\n### Step 3: Mark Batch Complete\nAfter implementing all methods in a batch, update status in inventory.\n\n---\n\n## Overload Pattern Template\n\n```python\nfrom typing import Awaitable, overload\n\nfrom redis.typing import AsyncClientProtocol, SyncClientProtocol\n\nclass SomeCommands:\n @overload\n def method(self: SyncClientProtocol, arg: str | None = None) -> SyncReturnType: ...\n @overload\n def method(\n self: AsyncClientProtocol, arg: str | None = None\n ) -> Awaitable[SyncReturnType]: ...\n def method(\n self, arg: str | None = None\n ) -> SyncReturnType | Awaitable[SyncReturnType]:\n # original implementation unchanged\n ...\n```\n\n### Protocol Definitions (in `redis/typing.py`)\n\n```python\nclass SyncClientProtocol(Protocol):\n \"\"\"Marker for sync clients.\"\"\"\n _is_async_client: Literal[False]\n\nclass AsyncClientProtocol(Protocol):\n \"\"\"Marker for async clients.\"\"\"\n _is_async_client: Literal[True]\n```\n\n**Note:** We use Protocol-based discrimination (not `Redis[bytes]` / `AsyncRedis[str]`) to avoid multiplying overloads by `decode_responses` setting.\n\n---\n\n## Status Legend\n\n| Symbol | Meaning |\n|--------|---------|\n| ✅ | Can use standard overload pattern |\n| ⚠️ | Has separate async implementation - SKIP |\n| 🔄 | Returns iterator - SKIP |\n| ❌ | Dunder method - SKIP |\n| 📋 | Already has explicit types - may still need overloads |\n\n---\n\n## CRITICAL: Response Callback System\n\nUnderstanding the callback system is essential for determining accurate return types.\n\n### Three-Tier Callback Architecture\n\nThe `redis-py` library uses three callback dictionaries in `redis/_parsers/helpers.py`:\n\n| Dictionary | Lines | Purpose |\n|------------|-------|---------|\n| `_RedisCallbacks` | 754-844 | **Base callbacks** - shared by both RESP2 and RESP3 |\n| `_RedisCallbacksRESP2` | 847-896 | **RESP2-specific** overrides/additions |\n| `_RedisCallbacksRESP3` | 899-947 | **RESP3-specific** overrides/additions |\n\n### How Callbacks Are Applied\n\n1. When a command is executed, the library checks for a callback in this order:\n - If using RESP2: Check `_RedisCallbacksRESP2` first, then fall back to `_RedisCallbacks`\n - If using RESP3: Check `_RedisCallbacksRESP3` first, then fall back to `_RedisCallbacks`\n\n2. If no callback exists, the raw response is returned (depends on `decode_responses` setting)\n\n### Key Callback Functions\n\n| Callback | Return Type | Notes |\n|----------|-------------|-------|\n| `bool_ok` | `bool` | Converts \"OK\" to `True` |\n| `bool` | `bool` | Converts int to bool |\n| `str_if_bytes` | `str` | Converts bytes to str regardless of decode_responses |\n| `float_or_none` | `float \\| None` | Parses float, returns None if null |\n| `parse_scan` | `tuple[int, list]` | Cursor + keys |\n| `parse_info` | `dict[str, Any]` | Parses INFO output |\n| `zset_score_pairs` | `list[tuple[..., float]]` | For sorted set with scores (RESP2) |\n| `zset_score_pairs_resp3` | `list[tuple[..., float]]` | For sorted set with scores (RESP3) |\n\n### Protocol-Specific Return Types\n\n**IMPORTANT**: Some commands have different return types depending on protocol version:\n\n| Command | RESP2 Return | RESP3 Return | Reason |\n|---------|--------------|--------------|--------|\n| `acl_cat` | `list[str]` | `list[bytes \\| str]` | RESP2 has `str_if_bytes`, RESP3 has no callback |\n| `acl_genpass` | `str` | `bytes \\| str` | RESP2 has `str_if_bytes`, RESP3 has no callback |\n| `client_getname` | `str \\| None` | `bytes \\| str \\| None` | RESP2 has `str_if_bytes`, RESP3 has no callback |\n| `geohash` | `list[str]` | `list[bytes \\| str]` | RESP2 has `str_if_bytes`, RESP3 has no callback |\n| `hgetall` | `dict` (pairs_to_dict) | `dict` (identity) | Different parsing logic |\n| `zincrby`, `zscore` | `float` (float_or_none) | `float` (raw) | RESP2 parses, RESP3 returns directly |\n\n### How to Determine Return Type\n\n1. **Check `_RedisCallbacks`** (base) - Is there a callback for the command?\n2. **Check `_RedisCallbacksRESP2`** - Is there a RESP2-specific override?\n3. **Check `_RedisCallbacksRESP3`** - Is there a RESP3-specific override?\n4. **If no callback** - Return type depends on `decode_responses`:\n - Bulk string: `bytes | str`\n - Integer: `int`\n - Array: `list[...]`\n - Null: `None`\n\n### Recommended Typing Strategy\n\nFor commands with protocol-specific differences, use the **most permissive union type**:\n- If RESP2 returns `str` and RESP3 returns `bytes | str`, use `bytes | str`\n- This ensures type safety regardless of protocol version\n\n---\n\n## Batches\n\n### BATCH 1: ACLCommands (core.py)\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 1 | `acl_cat` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 2 | `acl_dryrun` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback |\n| 3 | `acl_deluser` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 4 | `acl_genpass` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 5 | `acl_getuser` | `dict \\| None` | `Awaitable[dict \\| None]` | ✅ | Base: parse_acl_getuser |\n| 6 | `acl_help` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 7 | `acl_list` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 8 | `acl_log` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | Base: parse_acl_log / RESP3: lambda |\n| 9 | `acl_log_reset` | `bool` | `Awaitable[bool]` | ✅ | Via acl_log with RESET |\n| 10 | `acl_load` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 11 | `acl_save` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 12 | `acl_setuser` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 13 | `acl_users` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 14 | `acl_whoami` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n\n### BATCH 2: ManagementCommands Part 1 (core.py) - Methods 15-50\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 15 | `auth` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 16 | `bgrewriteaof` | `bool \\| bytes \\| str` | `Awaitable[bool \\| bytes \\| str]` | ✅ | RESP2: True / RESP3: raw |\n| 17 | `bgsave` | `bool \\| bytes \\| str` | `Awaitable[bool \\| bytes \\| str]` | ✅ | RESP2: True / RESP3: raw |\n| 18 | `role` | `list` | `Awaitable[list]` | ✅ | No callback - mixed types |\n| 19 | `client_kill` | `bool \\| int` | `Awaitable[bool \\| int]` | ✅ | Base: parse_client_kill |\n| 20 | `client_kill_filter` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 21 | `client_info` | `dict[str, str]` | `Awaitable[dict[str, str]]` | ✅ | Base: parse_client_info |\n| 22 | `client_list` | `list[dict[str, str]]` | `Awaitable[list[dict[str, str]]]` | ✅ | Base: parse_client_list |\n| 23 | `client_getname` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 24 | `client_getredir` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 25 | `client_reply` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback |\n| 26 | `client_id` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 27 | `client_tracking_on` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - returns raw OK |\n| 28 | `client_tracking_off` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - returns raw OK |\n| 29 | `client_tracking` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - returns raw OK |\n| 30 | `client_trackinginfo` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 31 | `client_setname` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 32 | `client_setinfo` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 33 | `client_unblock` | `bool` | `Awaitable[bool]` | ✅ | Base: bool (converts int) |\n| 34 | `client_pause` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 35 | `client_unpause` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - returns raw OK |\n| 36 | `client_no_evict` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - returns raw OK |\n| 37 | `client_no_touch` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - returns raw OK |\n| 38 | `command` | `list` | `Awaitable[list]` | ✅ | Base: parse_command / RESP3: parse_command_resp3 |\n| 39 | `command_info` | `None` | `None` | ⚠️ SKIP | N/A |\n| 40 | `command_count` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 41 | `command_list` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw |\n| 42 | `command_getkeysandflags` | `list[list[bytes \\| str \\| list[bytes \\| str]]]` | `Awaitable[list[list[bytes \\| str \\| list[bytes \\| str]]]]` | ✅ | No callback - mixed [key, flags] shape |\n| 43 | `command_docs` | `dict` | `Awaitable[dict]` | ✅ | No callback - raw dict |\n| 44 | `config_get` | `dict[str, str]` | `Awaitable[dict[str, str]]` | ✅ | RESP2: parse_config_get / RESP3: lambda |\n| 45 | `config_set` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 46 | `config_resetstat` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 47 | `config_rewrite` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - returns raw OK |\n| 48 | `dbsize` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 49 | `debug_object` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | RESP2: parse_debug_object / RESP3: raw |\n| 50 | `debug_segfault` | `None` | `None` | ⚠️ SKIP | N/A |\n\n### BATCH 3: ManagementCommands Part 2 (core.py) - Methods 51-93\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 51 | `echo` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 52 | `flushall` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 53 | `flushdb` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 54 | `sync` | `bytes` | `Awaitable[bytes]` | ✅ | No callback - raw bytes |\n| 55 | `psync` | `bytes` | `Awaitable[bytes]` | ✅ | No callback - raw bytes |\n| 56 | `swapdb` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 57 | `select` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 58 | `info` | `dict[str, Any]` | `Awaitable[dict[str, Any]]` | ✅ | Base: parse_info |\n| 59 | `lastsave` | `datetime` | `Awaitable[datetime]` | ✅ | Base: timestamp_to_datetime |\n| 60 | `latency_doctor` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 61 | `latency_graph` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 62 | `lolwut` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 63 | `reset` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 64 | `migrate` | `bool \\| bytes \\| str` | `Awaitable[bool \\| bytes \\| str]` | ✅ | No callback - NOKEY or OK |\n| 65 | `object` | `Any` | `Awaitable[Any]` | ✅ | Varies by subcommand |\n| 66 | `memory_doctor` | `None` | `None` | ⚠️ SKIP | N/A |\n| 67 | `memory_help` | `None` | `None` | ⚠️ SKIP | N/A |\n| 68 | `memory_stats` | `dict[str, Any]` | `Awaitable[dict[str, Any]]` | ✅ | RESP2: parse_memory_stats / RESP3: lambda |\n| 69 | `memory_malloc_stats` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 70 | `memory_usage` | `int \\| None` | `Awaitable[int \\| None]` | ✅ | Integer or nil reply |\n| 71 | `memory_purge` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 72 | `latency_histogram` | `dict` | `Awaitable[dict]` | ✅ | No callback - raw dict |\n| 73 | `latency_history` | `list[tuple[int, int]]` | `Awaitable[list[tuple[int, int]]]` | ✅ | No callback - array of arrays |\n| 74 | `latency_latest` | `list[tuple[bytes \\| str, int, int, int]]` | `Awaitable[list[tuple[bytes \\| str, int, int, int]]]` | ✅ | First element is key name |\n| 75 | `latency_reset` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 76 | `ping` | `bool` | `Awaitable[bool]` | 📋 | Base: lambda - returns True if PONG |\n| 77 | `quit` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 78 | `replicaof` | `bool` | `Awaitable[bool]` | ✅ | No callback - OK |\n| 79 | `save` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 80 | `shutdown` | `None` | `None` | ⚠️ SKIP | N/A |\n| 81 | `slaveof` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 82 | `slowlog_get` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | Base: parse_slowlog_get |\n| 83 | `slowlog_len` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 84 | `slowlog_reset` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 85 | `time` | `tuple[int, int]` | `Awaitable[tuple[int, int]]` | ✅ | Base: lambda - tuple(secs, usecs) |\n| 86 | `wait` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 87 | `waitaof` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 88 | `hello` | `dict` | `Awaitable[dict]` | ✅ | No callback - raw dict |\n| 89 | `failover` | `bool` | `Awaitable[bool]` | ✅ | No callback - OK |\n| 90 | `hotkeys_start` | `bytes \\| str` | `Awaitable[bytes \\| str]` | 📋 | No callback - raw |\n| 91 | `hotkeys_stop` | `bytes \\| str` | `Awaitable[bytes \\| str]` | 📋 | No callback - raw |\n| 92 | `hotkeys_reset` | `bytes \\| str` | `Awaitable[bytes \\| str]` | 📋 | No callback - raw |\n| 93 | `hotkeys_get` | `list[dict]` | `Awaitable[list[dict]]` | 📋 | RESP2: lambda pairs_to_dict |\n\n### BATCH 4: BasicKeyCommands Part 1 (core.py) - Methods 94-130\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 94 | `append` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 95 | `bitcount` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 96 | `bitfield` | `BitFieldOperation` | `BitFieldOperation` | 📋 | Returns operation builder |\n| 97 | `bitfield_ro` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 98 | `bitop` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 99 | `bitpos` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 100 | `copy` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 101 | `decrby` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 102 | `delete` | `int` | `Awaitable[int]` | ✅ | Integer reply - count deleted |\n| 103 | `__delitem__` | `None` | N/A | ❌ SKIP | Dunder |\n| 104 | `delex` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 105 | `dump` | `bytes \\| None` | `Awaitable[bytes \\| None]` | ✅ | Always bytes (serialized) |\n| 106 | `exists` | `int` | `Awaitable[int]` | ✅ | Integer reply - count |\n| 107 | `expire` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 108 | `expireat` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 109 | `expiretime` | `int` | `Awaitable[int]` | 📋 | Integer - timestamp |\n| 110 | `digest_local` | `bytes \\| str` | `Awaitable[bytes \\| str]` | 📋 | No callback - raw |\n| 111 | `digest` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 112 | `get` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ DONE | No callback - raw |\n| 113 | `getdel` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 114 | `getex` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 115 | `__getitem__` | `bytes \\| str` | N/A | ❌ SKIP | Dunder |\n| 116 | `getbit` | `int` | `Awaitable[int]` | ✅ | Integer reply - 0 or 1 |\n| 117 | `getrange` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 118 | `getset` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 119 | `incrby` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 120 | `incrbyfloat` | `float` | `Awaitable[float]` | ✅ | Base: float |\n| 121 | `keys` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw array |\n| 122 | `lmove` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 123 | `blmove` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 124 | `mget` | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | No callback - raw array |\n| 125 | `mset` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 126 | `msetex` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 127 | `msetnx` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 128 | `move` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 129 | `persist` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 130 | `pexpire` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n\n### BATCH 5: BasicKeyCommands Part 2 (core.py) - Methods 131-156\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 131 | `pexpireat` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 132 | `pexpiretime` | `int` | `Awaitable[int]` | ✅ | Integer - timestamp in ms |\n| 133 | `psetex` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 134 | `pttl` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 135 | `hrandfield` | `bytes \\| str \\| list \\| None` | `Awaitable[bytes \\| str \\| list \\| None]` | ✅ | Varies by COUNT param |\n| 136 | `randomkey` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 137 | `rename` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 138 | `renamenx` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 139 | `restore` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw OK |\n| 140 | `set` | `bool \\| None` | `Awaitable[bool \\| None]` | ✅ DONE | Base: parse_set_result |\n| 141 | `__setitem__` | `None` | N/A | ❌ SKIP | Dunder |\n| 142 | `setbit` | `int` | `Awaitable[int]` | ✅ | Integer reply - prev bit |\n| 143 | `setex` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 144 | `setnx` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 145 | `setrange` | `int` | `Awaitable[int]` | ✅ | Integer reply - new length |\n| 146 | `stralgo` | `dict \\| str \\| int` | `Awaitable[dict \\| str \\| int]` | ✅ | RESP2: parse_stralgo / RESP3: lambda |\n| 147 | `strlen` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 148 | `substr` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 149 | `touch` | `int` | `Awaitable[int]` | ✅ | Integer reply - count |\n| 150 | `ttl` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 151 | `type` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 152 | `watch` | `bool` | `Awaitable[bool]` | ⚠️ SKIP | Transaction |\n| 153 | `unwatch` | `bool` | `Awaitable[bool]` | ⚠️ SKIP | Transaction |\n| 154 | `unlink` | `int` | `Awaitable[int]` | ✅ | Integer reply - count |\n| 155 | `lcs` | `bytes \\| str \\| int \\| list \\| dict` | `Awaitable[bytes \\| str \\| int \\| list \\| dict]` | ✅ | Raw reply varies by options and protocol |\n| 156 | `__contains__` | `bool` | N/A | ❌ SKIP | Dunder |\n\n### BATCH 6: ListCommands (core.py) - Methods 157-178\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 157 | `blpop` | `tuple[bytes \\| str, bytes \\| str] \\| list[bytes \\| str] \\| None` | `Awaitable[tuple[bytes \\| str, bytes \\| str] \\| list[bytes \\| str] \\| None]` | ✅ | RESP2: tuple / RESP3: raw list |\n| 158 | `brpop` | `tuple[bytes \\| str, bytes \\| str] \\| list[bytes \\| str] \\| None` | `Awaitable[tuple[bytes \\| str, bytes \\| str] \\| list[bytes \\| str] \\| None]` | ✅ | RESP2: tuple / RESP3: raw list |\n| 159 | `brpoplpush` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 160 | `blmpop` | `list[bytes \\| str \\| list[bytes \\| str]] \\| None` | `Awaitable[list[bytes \\| str \\| list[bytes \\| str]] \\| None]` | ✅ | No callback - nested [key, values] shape |\n| 161 | `lmpop` | `list[bytes \\| str \\| list[bytes \\| str]] \\| None` | `Awaitable[list[bytes \\| str \\| list[bytes \\| str]] \\| None]` | ✅ | No callback - nested [key, values] shape |\n| 162 | `lindex` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 163 | `linsert` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 164 | `llen` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 165 | `lpop` | `bytes \\| str \\| list[bytes \\| str] \\| None` | `Awaitable[bytes \\| str \\| list[bytes \\| str] \\| None]` | ✅ | Varies by COUNT |\n| 166 | `lpush` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 167 | `lpushx` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 168 | `lrange` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw array |\n| 169 | `lrem` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 170 | `lset` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 171 | `ltrim` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 172 | `rpop` | `bytes \\| str \\| list[bytes \\| str] \\| None` | `Awaitable[bytes \\| str \\| list[bytes \\| str] \\| None]` | ✅ | Varies by COUNT |\n| 173 | `rpoplpush` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 174 | `rpush` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 175 | `rpushx` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 176 | `lpos` | `int \\| list[int] \\| None` | `Awaitable[int \\| list[int] \\| None]` | ✅ | Varies by COUNT/RANK |\n| 177 | `sort` | `list[bytes \\| str] \\| list[tuple[bytes \\| str, ...]] \\| int` | `Awaitable[list[bytes \\| str] \\| list[tuple[bytes \\| str, ...]] \\| int]` | ✅ | Base: sort_return_tuples incl. grouped tuples |\n| 178 | `sort_ro` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw array |\n\n### BATCH 7: ScanCommands + SetCommands (core.py) - Methods 179-202\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 179 | `scan` | `tuple[int, list[bytes \\| str]]` | `Awaitable[tuple[int, list[bytes \\| str]]]` | ✅ | Base: parse_scan |\n| 180 | `scan_iter` | `Iterator` | `AsyncIterator` | 🔄 SKIP | Iterator |\n| 181 | `sscan` | `tuple[int, list[bytes \\| str]]` | `Awaitable[tuple[int, list[bytes \\| str]]]` | ✅ | Base: parse_scan |\n| 182 | `sscan_iter` | `Iterator` | `AsyncIterator` | 🔄 SKIP | Iterator |\n| 183 | `hscan` | `tuple[int, dict[bytes \\| str, bytes \\| str] \\| list[bytes \\| str]]` | `Awaitable[tuple[int, dict[bytes \\| str, bytes \\| str] \\| list[bytes \\| str]]]` | ✅ | Base: parse_hscan, `NOVALUES` returns key list |\n| 184 | `hscan_iter` | `Iterator` | `AsyncIterator` | 🔄 SKIP | Iterator |\n| 185 | `zscan` | `tuple[int, list[tuple[bytes \\| str, float]]]` | `Awaitable[tuple[int, list[tuple[bytes \\| str, float]]]]` | ✅ | Base: parse_zscan |\n| 186 | `zscan_iter` | `Iterator` | `AsyncIterator` | 🔄 SKIP | Iterator |\n| 187 | `sadd` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 188 | `scard` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 189 | `sdiff` | `set[bytes \\| str]` | `Awaitable[set[bytes \\| str]]` | ✅ | RESP2+RESP3: lambda set |\n| 190 | `sdiffstore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 191 | `sinter` | `set[bytes \\| str]` | `Awaitable[set[bytes \\| str]]` | ✅ | RESP2+RESP3: lambda set |\n| 192 | `sintercard` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 193 | `sinterstore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 194 | `Literal[0] \\| Literal[1]` | `Awaitable[Literal[0] \\| Literal[1]]` | ✅ | Integer 0/1 - no callback |\n| 195 | `smembers` | `set[bytes \\| str]` | `Awaitable[set[bytes \\| str]]` | ✅ | RESP2+RESP3: lambda set |\n| 196 | `smismember` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of 0/1 |\n| 197 | `smove` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 198 | `spop` | `bytes \\| str \\| set[bytes \\| str] \\| None` | `Awaitable[bytes \\| str \\| set[bytes \\| str] \\| None]` | ✅ | Varies by COUNT, count form returns a set |\n| 199 | `srandmember` | `bytes \\| str \\| list[bytes \\| str] \\| None` | `Awaitable[bytes \\| str \\| list[bytes \\| str] \\| None]` | ✅ | Varies by NUMBER |\n| 200 | `srem` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 201 | `sunion` | `set[bytes \\| str]` | `Awaitable[set[bytes \\| str]]` | ✅ | RESP2+RESP3: lambda set |\n| 202 | `sunionstore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n\n### BATCH 8: StreamCommands (core.py) - Methods 203-226\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 203 | `xack` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 204 | `xackdel` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 205 | `xadd` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - stream ID |\n| 206 | `xcfgset` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw OK |\n| 207 | `xautoclaim` | `list[Any]` | `Awaitable[list[Any]]` | ✅ | Base: parse_xautoclaim / JUSTID special case |\n| 208 | `xclaim` | `list[tuple[bytes \\| str \\| None, dict \\| None]] \\| list[bytes \\| str]` | `Awaitable[list[tuple[bytes \\| str \\| None, dict \\| None]] \\| list[bytes \\| str]]` | ✅ | Base: parse_xclaim / JUSTID returns ID list |\n| 209 | `xdel` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 210 | `xdelex` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 211 | `xgroup_create` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 212 | `xgroup_delconsumer` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 213 | `xgroup_destroy` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 214 | `xgroup_createconsumer` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 215 | `xgroup_setid` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 216 | `xinfo_consumers` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | RESP2: parse_list_of_dicts / RESP3: lambda |\n| 217 | `xinfo_groups` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | RESP2: parse_list_of_dicts / RESP3: lambda |\n| 218 | `xinfo_stream` | `dict[str, Any]` | `Awaitable[dict[str, Any]]` | ✅ | Base: parse_xinfo_stream |\n| 219 | `xlen` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 220 | `xpending` | `dict[str, Any]` | `Awaitable[dict[str, Any]]` | ✅ | Base: parse_xpending |\n| 221 | `xpending_range` | `list[dict[str, bytes \\| str \\| int]]` | `Awaitable[list[dict[str, bytes \\| str \\| int]]]` | ✅ | parse_xpending_range detail rows |\n| 222 | `xrange` | `list[tuple[bytes \\| str \\| None, dict \\| None]] \\| None` | `Awaitable[list[tuple[bytes \\| str \\| None, dict \\| None]] \\| None]` | ✅ | Base: parse_stream_list |\n| 223 | `xread` | `list \\| dict` | `Awaitable[list \\| dict]` | ✅ | RESP2: parse_xread / RESP3: parse_xread_resp3 |\n| 224 | `xreadgroup` | `list \\| dict` | `Awaitable[list \\| dict]` | ✅ | RESP2: parse_xread / RESP3: parse_xread_resp3 |\n| 225 | `xrevrange` | `list[tuple[bytes \\| str \\| None, dict \\| None]] \\| None` | `Awaitable[list[tuple[bytes \\| str \\| None, dict \\| None]] \\| None]` | ✅ | Base: parse_stream_list |\n| 226 | `xtrim` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n\n### BATCH 9: SortedSetCommands Part 1 (core.py) - Methods 227-260\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 227 | `zadd` | `int \\| float` | `Awaitable[int \\| float]` | ✅ | RESP2: parse_zadd |\n| 228 | `zcard` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 229 | `zcount` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 230 | `zdiff` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | `WITHSCORES`: RESP2 tuple pairs / RESP3 raw nested lists |\n| 231 | `zdiffstore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 232 | `zincrby` | `float \\| None` | `Awaitable[float \\| None]` | ✅ | RESP2: `float_or_none` / RESP3: raw float |\n| 233 | `zinter` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | `score_cast_func` means scored branch uses `Any` |\n| 234 | `zinterstore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 235 | `zintercard` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 236 | `zlexcount` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 237 | `zpopmax` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | RESP2 tuple pairs / RESP3 nested lists |\n| 238 | `zpopmin` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | RESP2 tuple pairs / RESP3 nested lists |\n| 239 | `zrandmember` | `ZRandMemberResponse` | `Awaitable[ZRandMemberResponse]` | ✅ | COUNT / WITHSCORES shape varies by protocol |\n| 240 | `bzpopmax` | `BlockingZSetPopResponse` | `Awaitable[BlockingZSetPopResponse]` | ✅ | RESP2 tuple / RESP3 raw list / `None` |\n| 241 | `bzpopmin` | `BlockingZSetPopResponse` | `Awaitable[BlockingZSetPopResponse]` | ✅ | RESP2 tuple / RESP3 raw list / `None` |\n| 242 | `zmpop` | `ZMPopResponse` | `Awaitable[ZMPopResponse]` | ✅ | Raw `[key, [[member, score], ...]]` or `None` |\n| 243 | `bzmpop` | `ZMPopResponse` | `Awaitable[ZMPopResponse]` | ✅ | Raw `[key, [[member, score], ...]]` or `None` |\n| 244 | `zrange` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | `score_cast_func` means scored branch uses `Any` |\n| 245 | `zrevrange` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | `score_cast_func` means scored branch uses `Any` |\n| 246 | `zrangestore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 247 | `zrangebylex` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw array |\n| 248 | `zrevrangebylex` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw array |\n| 249 | `zrangebyscore` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | `score_cast_func` means scored branch uses `Any` |\n| 250 | `zrevrangebyscore` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | `score_cast_func` means scored branch uses `Any` |\n| 251 | `zrank` | `ZRankResponse` | `Awaitable[ZRankResponse]` | ✅ | `WITHSCORE` returns `[rank, score]`, not tuple |\n| 252 | `zrem` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 253 | `zremrangebylex` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 254 | `zremrangebyrank` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 255 | `zremrangebyscore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 256 | `zrevrank` | `ZRankResponse` | `Awaitable[ZRankResponse]` | ✅ | `WITHSCORE` returns `[rank, score]`, not tuple |\n| 257 | `zscore` | `float \\| None` | `Awaitable[float \\| None]` | ✅ | RESP2: `float_or_none` / RESP3: raw float |\n| 258 | `zunion` | `ZSetRangeResponse` | `Awaitable[ZSetRangeResponse]` | ✅ | `score_cast_func` means scored branch uses `Any` |\n| 259 | `zunionstore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 260 | `zmscore` | `list[float \\| None]` | `Awaitable[list[float \\| None]]` | ✅ | RESP2: `parse_zmscore` / RESP3: raw float list |\n\n### BATCH 10: HyperlogCommands + HashCommands (core.py) - Methods 261-289\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 261 | `pfadd` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 262 | `pfcount` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 263 | `pfmerge` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 264 | `hdel` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 265 | `hexists` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 266 | `hget` | `bytes \\| str \\| None` | `Awaitable[bytes \\| str \\| None]` | ✅ | No callback - raw |\n| 267 | `hgetall` | `dict[bytes \\| str, bytes \\| str]` | `Awaitable[dict[bytes \\| str, bytes \\| str]]` | ✅ | RESP2: pairs_to_dict / RESP3: identity |\n| 268 | `hgetdel` | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | No callback - one result per requested field |\n| 269 | `hgetex` | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | No callback - one result per requested field |\n| 270 | `hincrby` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 271 | `hincrbyfloat` | `float` | `Awaitable[float]` | ✅ | Base: float |\n| 272 | `hkeys` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw array |\n| 273 | `hlen` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 274 | `hset` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 275 | `hsetex` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 276 | `hsetnx` | `int` | `Awaitable[int]` | ✅ | No callback - integer 0/1 reply |\n| 277 | `hmset` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 278 | `hmget` | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | No callback - raw array |\n| 279 | `hvals` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw array |\n| 280 | `hstrlen` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 281 | `hexpire` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 282 | `hpexpire` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 283 | `hexpireat` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 284 | `hpexpireat` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 285 | `hpersist` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 286 | `hexpiretime` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 287 | `hpexpiretime` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 288 | `httl` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n| 289 | `hpttl` | `list[int]` | `Awaitable[list[int]]` | ✅ | Array of integers |\n\n### BATCH 11: PubSubCommands + ScriptCommands (core.py) - Methods 290-306\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 290 | `publish` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 291 | `spublish` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 292 | `pubsub_channels` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw |\n| 293 | `pubsub_shardchannels` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw |\n| 294 | `pubsub_numpat` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 295 | `pubsub_numsub` | `list[tuple[bytes \\| str, int]]` | `Awaitable[list[tuple[bytes \\| str, int]]]` | ✅ | Base: parse_pubsub_numsub |\n| 296 | `pubsub_shardnumsub` | `list[tuple[bytes \\| str, int]]` | `Awaitable[list[tuple[bytes \\| str, int]]]` | ✅ | Base: parse_pubsub_numsub |\n| 297 | `eval` | `Any` | `Awaitable[Any]` | ✅ | Script-dependent |\n| 298 | `eval_ro` | `Any` | `Awaitable[Any]` | ✅ | Script-dependent |\n| 299 | `evalsha` | `Any` | `Awaitable[Any]` | ✅ | Script-dependent |\n| 300 | `evalsha_ro` | `Any` | `Awaitable[Any]` | ✅ | Script-dependent |\n| 301 | `script_exists` | `list[bool]` | `Awaitable[list[bool]]` | ✅ | Base: lambda map bool |\n| 302 | `script_debug` | `None` | `None` | ⚠️ SKIP | N/A |\n| 303 | `script_flush` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 304 | `script_kill` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 305 | `script_load` | `str` | `Awaitable[str]` | ✅ | Base: str_if_bytes |\n| 306 | `register_script` | `Script` | `AsyncScript` | ⚠️ SKIP | Different classes |\n\n### BATCH 12: GeoCommands + ModuleCommands + ClusterCommands + FunctionCommands (core.py) - Methods 307-335\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 307 | `geoadd` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 308 | `geodist` | `float \\| None` | `Awaitable[float \\| None]` | ✅ | Base: float_or_none |\n| 309 | `geohash` | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 310 | `geopos` | `list[tuple[float, float] \\| None]` | `Awaitable[list[tuple[float, float] \\| None]]` | ✅ | RESP2: lambda float tuple / RESP3: raw |\n| 311 | `georadius` | `list[Any] \\| int` | `Awaitable[list[Any] \\| int]` | ✅ | `store` / `store_dist` return int, otherwise parsed list |\n| 312 | `georadiusbymember` | `list[Any] \\| int` | `Awaitable[list[Any] \\| int]` | ✅ | `store` / `store_dist` return int, otherwise parsed list |\n| 313 | `geosearch` | `list[Any]` | `Awaitable[list[Any]]` | ✅ | Base: parse_geosearch_generic |\n| 314 | `geosearchstore` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 315 | `module_load` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 316 | `module_loadex` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw OK reply |\n| 317 | `module_unload` | `bool` | `Awaitable[bool]` | ✅ | Base: bool |\n| 318 | `module_list` | `list[dict[Any, Any]]` | `Awaitable[list[dict[Any, Any]]]` | ✅ | RESP2: lambda pairs_to_dict / RESP3: raw dict list |\n| 319 | `command_info` | `None` | `None` | ⚠️ SKIP | N/A |\n| 320 | `command_count` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 321 | `command_getkeys` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 322 | `command` | `dict[str, dict[str, Any]]` | `Awaitable[dict[str, dict[str, Any]]]` | ✅ | Base: parse_command / RESP3: parse_command_resp3 |\n| 323 | `cluster` | `Any` | `Awaitable[Any]` | ✅ | Subcommand-dependent |\n| 324 | `readwrite` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 325 | `readonly` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 326 | `function_load` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 327 | `function_delete` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 328 | `function_flush` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 329 | `function_list` | `list[Any]` | `Awaitable[list[Any]]` | ✅ | No callback - raw protocol-dependent list |\n| 330 | `fcall` | `Any` | `Awaitable[Any]` | ✅ | Function-dependent |\n| 331 | `fcall_ro` | `Any` | `Awaitable[Any]` | ✅ | Function-dependent |\n| 332 | `function_dump` | `bytes` | `Awaitable[bytes]` | ✅ | No callback - raw bytes |\n| 333 | `function_restore` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 334 | `function_kill` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw OK reply |\n| 335 | `function_stats` | `Any` | `Awaitable[Any]` | ✅ | No callback - raw protocol-dependent structure |\n\n### BATCH 13: ClusterCommands (cluster.py) - Methods 336-378\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 336 | `mget_nonatomic` | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ⚠️ SKIP | Complex multi-node |\n| 337 | `mset_nonatomic` | `list[bool]` | `Awaitable[list[bool]]` | ⚠️ SKIP | Complex multi-node |\n| 338 | `exists` | `int` | `Awaitable[int]` | 📋 | Integer reply |\n| 339 | `delete` | `int` | `Awaitable[int]` | 📋 | Integer reply |\n| 340 | `touch` | `int` | `Awaitable[int]` | 📋 | Integer reply |\n| 341 | `unlink` | `int` | `Awaitable[int]` | 📋 | Integer reply |\n| 342 | `slaveof` | `NoReturn` | `NoReturn` | 📋 SKIP | Raises error |\n| 343 | `replicaof` | `NoReturn` | `NoReturn` | 📋 SKIP | Raises error |\n| 344 | `swapdb` | `NoReturn` | `NoReturn` | 📋 SKIP | Raises error |\n| 345 | `cluster_myid` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 346 | `cluster_addslots` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 347 | `cluster_addslotsrange` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 348 | `cluster_countkeysinslot` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 349 | `cluster_count_failure_report` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 350 | `cluster_delslots` | `list[bool]` | `Awaitable[list[bool]]` | ⚠️ SKIP | Complex multi-node |\n| 351 | `cluster_delslotsrange` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 352 | `cluster_failover` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 353 | `cluster_info` | `dict[str, str]` | `Awaitable[dict[str, str]]` | ✅ | Base: parse_cluster_info |\n| 354 | `cluster_keyslot` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 355 | `cluster_meet` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 356 | `cluster_nodes` | `str` | `Awaitable[str]` | ✅ | Base: parse_cluster_nodes |\n| 357 | `cluster_replicate` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 358 | `cluster_reset` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 359 | `cluster_save_config` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 360 | `cluster_get_keys_in_slot` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | RESP2: str_if_bytes / RESP3: raw |\n| 361 | `cluster_set_config_epoch` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 362 | `cluster_setslot` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 363 | `cluster_setslot_stable` | `bool` | `Awaitable[bool]` | ✅ | No callback - OK |\n| 364 | `cluster_replicas` | `str` | `Awaitable[str]` | ✅ | Base: parse_cluster_nodes |\n| 365 | `cluster_slots` | `list` | `Awaitable[list]` | ✅ | No callback - raw |\n| 366 | `cluster_shards` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | No callback - raw |\n| 367 | `cluster_myshardid` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 368 | `cluster_links` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | No callback - raw |\n| 369 | `cluster_flushslots` | `bool` | `Awaitable[bool]` | ✅ | No callback - OK |\n| 370 | `cluster_bumpepoch` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 371 | `client_tracking_on` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ⚠️ SKIP | Cluster-specific, no callback |\n| 372 | `client_tracking_off` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ⚠️ SKIP | Cluster-specific, no callback |\n| 373 | `hotkeys_start` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ⚠️ SKIP | Cluster-specific |\n| 374 | `hotkeys_stop` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ⚠️ SKIP | Cluster-specific |\n| 375 | `hotkeys_reset` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ⚠️ SKIP | Cluster-specific |\n| 376 | `hotkeys_get` | `list[dict]` | `Awaitable[list[dict]]` | ⚠️ SKIP | Cluster-specific |\n| 377 | `stralgo` | `dict \\| bytes \\| str \\| int` | `Awaitable[dict \\| bytes \\| str \\| int]` | ✅ | RESP2: parse_stralgo / RESP3: lambda |\n| 378 | `scan_iter` | `Iterator` | `AsyncIterator` | 🔄 SKIP | Iterator |\n\n### BATCH 14: SentinelCommands (sentinel.py) - Methods 379-391\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 379 | `sentinel` | `Any` | `Awaitable[Any]` | ⚠️ SKIP | Async differs |\n| 380 | `sentinel_get_master_addr_by_name` | `tuple[str, int] \\| None` | `Awaitable[tuple[str, int] \\| None]` | ✅ | Base: parse_sentinel_get_master |\n| 381 | `sentinel_master` | `dict` | `Awaitable[dict]` | ✅ | RESP2: parse_sentinel_master / RESP3: parse_sentinel_state_resp3 |\n| 382 | `sentinel_masters` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | RESP2: parse_sentinel_masters / RESP3: parse_sentinel_masters_resp3 |\n| 383 | `sentinel_monitor` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 384 | `sentinel_remove` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 385 | `sentinel_sentinels` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | RESP2: parse_sentinel_slaves_and_sentinels / RESP3: _resp3 |\n| 386 | `sentinel_set` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 387 | `sentinel_slaves` | `list[dict]` | `Awaitable[list[dict]]` | ✅ | RESP2: parse_sentinel_slaves_and_sentinels / RESP3: _resp3 |\n| 388 | `sentinel_reset` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 389 | `sentinel_failover` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 390 | `sentinel_ckquorum` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n| 391 | `sentinel_flushconfig` | `bool` | `Awaitable[bool]` | ✅ | Base: bool_ok |\n\n### BATCH 15: SearchCommands (search/commands.py) - Methods 392-424\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 392 | `batch_indexer` | `BatchIndexer` | `BatchIndexer` | 📋 SKIP | Builder pattern |\n| 393 | `create_index` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 394 | `alter_schema_add` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 395 | `dropindex` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 396 | `add_document` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 397 | `add_document_hash` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 398 | `delete_document` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 399 | `load_document` | `Document` | `Awaitable[Document]` | ⚠️ SKIP | Async result parsing |\n| 400 | `get` | `Document` | `Awaitable[Document]` | ✅ | Module-specific |\n| 401 | `info` | `dict` | `Awaitable[dict]` | ⚠️ SKIP | Async result parsing |\n| 402 | `get_params_args` | `list` | `list` | 📋 SKIP | Helper method |\n| 403 | `search` | `Result` | `Awaitable[Result]` | ⚠️ SKIP | Async result parsing |\n| 404 | `hybrid_search` | `HybridResult` | `Awaitable[HybridResult]` | ⚠️ SKIP | Async result parsing |\n| 405 | `explain` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | No callback - raw |\n| 406 | `explain_cli` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw |\n| 407 | `aggregate` | `AggregateResult` | `Awaitable[AggregateResult]` | ⚠️ SKIP | Async result parsing |\n| 408 | `profile` | `tuple` | `Awaitable[tuple]` | ✅ | Module-specific |\n| 409 | `spellcheck` | `dict` | `Awaitable[dict]` | ⚠️ SKIP | Async result parsing |\n| 410 | `dict_add` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 411 | `dict_del` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 412 | `dict_dump` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw |\n| 413 | `config_set` | `bool` | `Awaitable[bool]` | ⚠️ SKIP | Async result parsing |\n| 414 | `config_get` | `dict` | `Awaitable[dict]` | ⚠️ SKIP | Async result parsing |\n| 415 | `tagvals` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | No callback - raw |\n| 416 | `aliasadd` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 417 | `aliasupdate` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 418 | `aliasdel` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 419 | `sugadd` | `int` | `Awaitable[int]` | ⚠️ SKIP | Async result parsing |\n| 420 | `suglen` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 421 | `sugdel` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 422 | `sugget` | `list` | `Awaitable[list]` | ⚠️ SKIP | Async result parsing |\n| 423 | `synupdate` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 424 | `syndump` | `dict[bytes \\| str, list[bytes \\| str]]` | `Awaitable[dict[bytes \\| str, list[bytes \\| str]]]` | ✅ | No callback - raw |\n\n### BATCH 16: JSONCommands (json/commands.py) - Methods 425-452\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 425 | `arrappend` | `list[int \\| None]` | `Awaitable[list[int \\| None]]` | ✅ | Module int array |\n| 426 | `arrindex` | `list[int \\| None]` | `Awaitable[list[int \\| None]]` | ✅ | Module int array |\n| 427 | `arrinsert` | `list[int \\| None]` | `Awaitable[list[int \\| None]]` | ✅ | Module int array |\n| 428 | `arrlen` | `list[int \\| None]` | `Awaitable[list[int \\| None]]` | ✅ | Module int array |\n| 429 | `arrpop` | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | Depends on decode_responses |\n| 430 | `arrtrim` | `list[int \\| None]` | `Awaitable[list[int \\| None]]` | ✅ | Module int array |\n| 431 | `type` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | Depends on decode_responses |\n| 432 | `resp` | `list` | `Awaitable[list]` | ✅ | JSON structure |\n| 433 | `objkeys` | `list[list[bytes \\| str] \\| None]` | `Awaitable[list[list[bytes \\| str] \\| None]]` | ✅ | Depends on decode_responses |\n| 434 | `objlen` | `list[int \\| None]` | `Awaitable[list[int \\| None]]` | ✅ | Module int array |\n| 435 | `numincrby` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | Depends on decode_responses |\n| 436 | `nummultby` | `bytes \\| str` | `Awaitable[bytes \\| str]` | ✅ | Depends on decode_responses |\n| 437 | `clear` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 438 | `delete` | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 439 | `get` | `Any` | `Awaitable[Any]` | ✅ | JSON parsed |\n| 440 | `mget` | `list[Any]` | `Awaitable[list[Any]]` | ✅ | JSON parsed |\n| 441 | `set` | `bool \\| None` | `Awaitable[bool \\| None]` | ✅ | OK or None |\n| 442 | `mset` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 443 | `merge` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 444 | `set_file` | `bool \\| None` | `Awaitable[bool \\| None]` | 📋 | Explicit |\n| 445 | `set_path` | `dict[str, bool]` | `Awaitable[dict[str, bool]]` | 📋 | Explicit |\n| 446 | `strlen` | `list[int \\| None]` | `Awaitable[list[int \\| None]]` | ✅ | Module int array |\n| 447 | `toggle` | `bool \\| list[bool]` | `Awaitable[bool \\| list[bool]]` | ✅ | Module bool |\n| 448 | `strappend` | `int \\| list[int \\| None]` | `Awaitable[int \\| list[int \\| None]]` | ✅ | Module int |\n| 449 | `debug` | `int \\| list[bytes \\| str]` | `Awaitable[int \\| list[bytes \\| str]]` | ✅ | Module mixed |\n| 450 | `jsonget` | `Any` | `Awaitable[Any]` | ✅ | Deprecated alias |\n| 451 | `jsonmget` | `Any` | `Awaitable[Any]` | ✅ | Deprecated alias |\n| 452 | `jsonset` | `Any` | `Awaitable[Any]` | ✅ | Deprecated alias |\n\n### BATCH 17: TimeSeriesCommands (timeseries/commands.py) - Methods 453-469\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 453 | `create` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 454 | `alter` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 455 | `add` | `int` | `Awaitable[int]` | ✅ | Timestamp |\n| 456 | `madd` | `list[int]` | `Awaitable[list[int]]` | ✅ | Timestamps |\n| 457 | `incrby` | `int` | `Awaitable[int]` | ✅ | Timestamp |\n| 458 | `decrby` | `int` | `Awaitable[int]` | ✅ | Timestamp |\n| 459 | `delete` | `int` | `Awaitable[int]` | ✅ | Count deleted |\n| 460 | `createrule` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 461 | `deleterule` | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 462 | `range` | `list[tuple[int, float]]` | `Awaitable[list[tuple[int, float]]]` | ✅ | Parsed samples |\n| 463 | `revrange` | `list[tuple[int, float]]` | `Awaitable[list[tuple[int, float]]]` | ✅ | Parsed samples |\n| 464 | `mrange` | `list` | `Awaitable[list]` | ✅ | Complex structure |\n| 465 | `mrevrange` | `list` | `Awaitable[list]` | ✅ | Complex structure |\n| 466 | `get` | `tuple[int, float] \\| list` | `Awaitable[tuple[int, float] \\| list]` | ✅ | Parsed sample |\n| 467 | `mget` | `list` | `Awaitable[list]` | ✅ | Complex structure |\n| 468 | `info` | `TSInfo` | `Awaitable[TSInfo]` | 📋 | Explicit type |\n| 469 | `queryindex` | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | Depends on decode_responses |\n\n### BATCH 18: BloomFilter Commands (bf/commands.py) - Methods 470-491\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 470 | `create` (BF) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 471 | `add` (BF) | `int` | `Awaitable[int]` | ✅ | 0 or 1 |\n| 472 | `madd` (BF) | `list[int]` | `Awaitable[list[int]]` | ✅ | 0s and 1s |\n| 473 | `insert` (BF) | `list[int]` | `Awaitable[list[int]]` | ✅ | 0s and 1s |\n| 474 | `exists` (BF) | `int` | `Awaitable[int]` | ✅ | 0 or 1 |\n| 475 | `mexists` (BF) | `list[int]` | `Awaitable[list[int]]` | ✅ | 0s and 1s |\n| 476 | `scandump` (BF) | `tuple[int, bytes \\| None]` | `Awaitable[tuple[int, bytes \\| None]]` | ✅ | Cursor + data |\n| 477 | `loadchunk` (BF) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 478 | `info` (BF) | `dict` | `Awaitable[dict]` | ✅ | Parsed info |\n| 479 | `card` (BF) | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 480 | `create` (CF) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 481 | `add` (CF) | `int` | `Awaitable[int]` | ✅ | 0 or 1 |\n| 482 | `addnx` (CF) | `int` | `Awaitable[int]` | ✅ | 0 or 1 |\n| 483 | `insert` (CF) | `list[int]` | `Awaitable[list[int]]` | ✅ | 0s and 1s |\n| 484 | `insertnx` (CF) | `list[int]` | `Awaitable[list[int]]` | ✅ | 0s and 1s |\n| 485 | `exists` (CF) | `int` | `Awaitable[int]` | ✅ | 0 or 1 |\n| 486 | `mexists` (CF) | `list[int]` | `Awaitable[list[int]]` | ✅ | 0s and 1s |\n| 487 | `delete` (CF) | `int` | `Awaitable[int]` | ✅ | 0 or 1 |\n| 488 | `count` (CF) | `int` | `Awaitable[int]` | ✅ | Integer reply |\n| 489 | `scandump` (CF) | `tuple[int, bytes \\| None]` | `Awaitable[tuple[int, bytes \\| None]]` | ✅ | Cursor + data |\n| 490 | `loadchunk` (CF) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 491 | `info` (CF) | `dict` | `Awaitable[dict]` | ✅ | Parsed info |\n\n### BATCH 19: TOPKCommands + TDigestCommands + CMSCommands (bf/commands.py) - Methods 492-518\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 492 | `reserve` (TOPK) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 493 | `add` (TOPK) | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | Depends on decode_responses |\n| 494 | `incrby` (TOPK) | `list[bytes \\| str \\| None]` | `Awaitable[list[bytes \\| str \\| None]]` | ✅ | Depends on decode_responses |\n| 495 | `query` (TOPK) | `list[int]` | `Awaitable[list[int]]` | ✅ | 0s and 1s |\n| 496 | `count` (TOPK) | `list[int]` | `Awaitable[list[int]]` | ✅ | Counts |\n| 497 | `list` (TOPK) | `list[bytes \\| str]` | `Awaitable[list[bytes \\| str]]` | ✅ | Depends on decode_responses |\n| 498 | `info` (TOPK) | `dict` | `Awaitable[dict]` | ✅ | Parsed info |\n| 499 | `create` (TD) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 500 | `reset` (TD) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 501 | `add` (TD) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 502 | `merge` (TD) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 503 | `min` (TD) | `float` | `Awaitable[float]` | ✅ | Float reply |\n| 504 | `max` (TD) | `float` | `Awaitable[float]` | ✅ | Float reply |\n| 505 | `quantile` (TD) | `list[float]` | `Awaitable[list[float]]` | ✅ | Floats |\n| 506 | `cdf` (TD) | `list[float]` | `Awaitable[list[float]]` | ✅ | Floats |\n| 507 | `info` (TD) | `dict` | `Awaitable[dict]` | ✅ | Parsed info |\n| 508 | `trimmed_mean` (TD) | `float` | `Awaitable[float]` | ✅ | Float reply |\n| 509 | `rank` (TD) | `list[int]` | `Awaitable[list[int]]` | ✅ | Integers |\n| 510 | `revrank` (TD) | `list[int]` | `Awaitable[list[int]]` | ✅ | Integers |\n| 511 | `byrank` (TD) | `list[float]` | `Awaitable[list[float]]` | ✅ | Floats |\n| 512 | `byrevrank` (TD) | `list[float]` | `Awaitable[list[float]]` | ✅ | Floats |\n| 513 | `initbydim` (CMS) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 514 | `initbyprob` (CMS) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 515 | `incrby` (CMS) | `list[int]` | `Awaitable[list[int]]` | ✅ | Integers |\n| 516 | `query` (CMS) | `list[int]` | `Awaitable[list[int]]` | ✅ | Integers |\n| 517 | `merge` (CMS) | `bool` | `Awaitable[bool]` | ✅ | Module OK |\n| 518 | `info` (CMS) | `dict` | `Awaitable[dict]` | ✅ | Parsed info |\n\n### BATCH 20: VectorSetCommands (vectorset/commands.py) - Methods 519-530 ✅ DONE\n| # | Method | Sync Type | Async Type | Status | Notes |\n|---|--------|-----------|------------|--------|-------|\n| 519 | `vadd` | `int` | `Awaitable[int]` | ✅ DONE | Integer reply |\n| 520 | `vsim` | `VSimResult` | `Awaitable[VSimResult]` | 📋 DONE | Explicit type |\n| 521 | `vdim` | `int` | `Awaitable[int]` | ✅ DONE | Integer reply |\n| 522 | `vcard` | `int` | `Awaitable[int]` | ✅ DONE | Integer reply |\n| 523 | `vrem` | `int` | `Awaitable[int]` | ✅ DONE | Integer reply |\n| 524 | `vemb` | `VEmbResult` | `Awaitable[VEmbResult]` | 📋 DONE | Explicit type |\n| 525 | `vlinks` | `VLinksResult` | `Awaitable[VLinksResult]` | 📋 DONE | Explicit type |\n| 526 | `vinfo` | `dict` | `Awaitable[dict]` | ✅ DONE | Standard dict |\n| 527 | `vsetattr` | `int` | `Awaitable[int]` | ✅ DONE | Integer reply |\n| 528 | `vgetattr` | `VGetAttrResult` | `Awaitable[VGetAttrResult]` | 📋 DONE | Explicit type |\n| 529 | `vrandmember` | `VRandMemberResult` | `Awaitable[VRandMemberResult]` | 📋 DONE | Explicit type |\n| 530 | `vrange` | `list[str]` | `Awaitable[list[str]]` | ✅ DONE | Standard list |\n\n---\n\n## Batch Summary\n\n| Batch | File | Command Class | Methods | Count |\n|-------|------|---------------|---------|-------|\n| 1 | core.py | ACLCommands | 1-14 | 14 |\n| 2 | core.py | ManagementCommands (Part 1) | 15-50 | 36 |\n| 3 | core.py | ManagementCommands (Part 2) | 51-93 | 43 |\n| 4 | core.py | BasicKeyCommands (Part 1) | 94-130 | 37 |\n| 5 | core.py | BasicKeyCommands (Part 2) | 131-156 | 26 |\n| 6 | core.py | ListCommands | 157-178 | 22 |\n| 7 | core.py | ScanCommands + SetCommands | 179-202 | 24 |\n| 8 | core.py | StreamCommands | 203-226 | 24 |\n| 9 | core.py | SortedSetCommands | 227-260 | 34 |\n| 10 | core.py | HyperlogCommands + HashCommands | 261-289 | 29 |\n| 11 | core.py | PubSubCommands + ScriptCommands | 290-306 | 17 |\n| 12 | core.py | GeoCommands + ModuleCommands + ClusterCommands + FunctionCommands | 307-335 | 29 |\n| 13 | cluster.py | ClusterCommands | 336-378 | 43 |\n| 14 | sentinel.py | SentinelCommands | 379-391 | 13 |\n| 15 | search/commands.py | SearchCommands | 392-424 | 33 |\n| 16 | json/commands.py | JSONCommands | 425-452 | 28 |\n| 17 | timeseries/commands.py | TimeSeriesCommands | 453-469 | 17 |\n| 18 | bf/commands.py | BFCommands + CFCommands | 470-491 | 22 |\n| 19 | bf/commands.py | TOPKCommands + TDigestCommands + CMSCommands | 492-518 | 27 |\n| 20 | vectorset/commands.py | VectorSetCommands | 519-530 | 12 ✅ DONE |\n\n---\n\n## Implementation Checklist\n\nFor each batch:\n- [ ] Review all methods in the batch\n- [ ] Skip methods marked with ⚠️ SKIP, 🔄 SKIP, ❌ SKIP\n- [ ] For each ✅ method:\n - [ ] Check the Notes column for callback info\n - [ ] Verify return type against `redis/_parsers/helpers.py` callback dictionaries\n - [ ] Add two `@overload` signatures\n - [ ] Run type checker\n- [ ] Test changes\n- [ ] Mark batch complete\n\n---\n\n## Notes\n\n1. **Callback System is Critical** - Always check the three-tier callback hierarchy in `redis/_parsers/helpers.py`:\n - `_RedisCallbacks` (base) - shared by RESP2 and RESP3\n - `_RedisCallbacksRESP2` - RESP2-specific overrides\n - `_RedisCallbacksRESP3` - RESP3-specific overrides\n2. **Protocol-Specific Types** - When RESP2 and RESP3 have different callbacks, use the most permissive union type\n3. **Import considerations** - Make sure `TYPE_CHECKING`, `overload`, and `Awaitable` are imported\n4. **Self-type discrimination** - Use `self: SyncClientProtocol` and `self: AsyncClientProtocol`\n5. **Keep original method** - Only add overloads, don't modify the actual implementation\n6. **No callback = raw response** - If a command has no callback, return type depends on `decode_responses`\n" }, { "path": ".dockerignore", "content": "**/__pycache__\n**/*.pyc\n.coverage\n.coverage.*\n" }, { "path": ".github/CODEOWNERS", "content": "doctests/* @dmaier-redislabs\n" }, { "path": ".github/ISSUE_TEMPLATE.md", "content": "Thanks for wanting to report an issue you've found in redis-py. Please delete this text and fill in the template below. \nIt is of course not always possible to reduce your code to a small test case, but it's highly appreciated to have as much data as possible. Thank you!\n\n**Version**: What redis-py and what redis version is the issue happening on?\n\n**Platform**: What platform / version? (For example Python 3.5.1 on Windows 7 / Ubuntu 15.10 / Azure)\n\n**Description**: Description of your issue, stack traces from errors and code that reproduces the issue\n" }, { "path": ".github/PULL_REQUEST_TEMPLATE.md", "content": "### Description of change\n\n_Please provide a description of the change here._\n\n### Pull Request check-list\n\n_Please make sure to review and check all of these items:_\n\n- [ ] Do tests and lints pass with this change?\n- [ ] Do the CI tests pass with this change (enable it first in your forked repo and wait for the github action build to finish)?\n- [ ] Is the new or changed code fully tested?\n- [ ] Is a documentation update included (if this change modifies existing APIs, or introduces new ones)?\n- [ ] Is there an example added to the examples folder (if applicable)?\n\n_NOTE: these things are not required to open a PR and can be done\nafterwards / while the PR is open._\n" }, { "path": ".github/actions/run-tests/action.yml", "content": "name: 'Run redis-py tests'\ndescription: 'Runs redis-py tests against different Redis versions and configurations'\ninputs:\n python-version:\n description: 'Python version to use for running tests'\n default: '3.12'\n parser-backend:\n description: 'Parser backend to use: plain or hiredis'\n required: true\n redis-version:\n description: 'Redis version to test against'\n required: true\n hiredis-version:\n description: 'hiredis version to test against'\n required: false\n default: '>3.0.0'\n hiredis-branch:\n description: 'hiredis branch to test against'\n required: false\n default: 'master'\n event-loop:\n description: 'Event loop to use'\n required: false\n default: 'asyncio'\n protocol:\n description: 'RESP protocol version to use'\n required: false\n default: 'all'\n repository:\n description: 'Repository to checkout'\n required: false\n ref:\n description: 'Branch to checkout'\n required: false\n client-libs-test-image:\n description: 'Client libs test image tag'\n required: false\nruns:\n using: \"composite\"\n steps:\n - uses: actions/checkout@v4\n with:\n repository: ${{ inputs.repository }}\n ref: ${{ inputs.ref }}\n\n - uses: actions/setup-python@v5\n with:\n python-version: ${{ inputs.python-version }}\n cache: 'pip'\n\n - uses: actions/checkout@v4\n if: ${{ inputs.parser-backend == 'hiredis' && inputs.hiredis-version == 'unstable' }}\n with:\n repository: redis/hiredis-py\n submodules: true\n path: hiredis-py\n ref: ${{ inputs.hiredis-branch }}\n\n - name: Setup Test environment\n env:\n REDIS_VERSION: ${{ inputs.redis-version }}\n CLIENT_LIBS_TEST_IMAGE_TAG: ${{ inputs.client-libs-test-image || inputs.redis-version }}\n CLIENT_LIBS_TEST_STACK_IMAGE_TAG: ${{ inputs.client-libs-test-image }}\n run: |\n set -e\n\n echo \"::group::Installing dependencies\"\n pip install -r dev_requirements.txt\n pip uninstall -y redis # uninstall Redis package installed via redis-entraid\n pip install -e .[jwt] # install the working copy\n if [ \"${{inputs.parser-backend}}\" == \"hiredis\" ]; then\n if [[ \"${{inputs.hiredis-version}}\" == \"unstable\" ]]; then\n echo \"Installing unstable version of hiredis from local directory\"\n pip install -e ./hiredis-py\n else\n pip install \"hiredis${{inputs.hiredis-version}}\"\n fi\n echo \"PARSER_BACKEND=$(echo \"${{inputs.parser-backend}}_${{inputs.hiredis-version}}\" | sed 's/[^a-zA-Z0-9]/_/g')\" >> $GITHUB_ENV\n else\n echo \"PARSER_BACKEND=${{inputs.parser-backend}}\" >> $GITHUB_ENV\n fi\n echo \"::endgroup::\"\n\n echo \"::group::Starting Redis servers\"\n # Check if REDIS_VERSION is in the custom map\n mapped_version=\"\"\n if [[ -n \"${REDIS_VERSION_CUSTOM_MAP:-}\" ]]; then\n for mapping in $REDIS_VERSION_CUSTOM_MAP; do\n tag=\"${mapping%%:*}\"\n version=\"${mapping##*:}\"\n if [[ \"$REDIS_VERSION\" == \"$tag\" ]]; then\n mapped_version=\"$version\"\n echo \"Found custom mapping: $REDIS_VERSION -> $mapped_version\"\n break\n fi\n done\n fi\n # Use mapped version if found, otherwise use REDIS_VERSION\n version_to_parse=\"${mapped_version:-$REDIS_VERSION}\"\n redis_major_version=$(echo \"$version_to_parse\" | grep -oP '^\\d+')\n echo \"REDIS_MAJOR_VERSION=${redis_major_version}\" >> $GITHUB_ENV\n\n if (( redis_major_version < 8 )); then\n echo \"Using redis-stack for module tests\"\n\n # Mapping of redis version to stack version\n declare -A redis_stack_version_mapping=(\n [\"7.4.4\"]=\"rs-7.4.0-v5\"\n [\"7.2.9\"]=\"rs-7.2.0-v17\"\n )\n\n if [[ -v redis_stack_version_mapping[$REDIS_VERSION] ]]; then\n if [[ -z \"${CLIENT_LIBS_TEST_STACK_IMAGE_TAG}\" ]]; then\n export CLIENT_LIBS_TEST_STACK_IMAGE_TAG=${redis_stack_version_mapping[$REDIS_VERSION]}\n fi\n echo \"REDIS_MOD_URL=redis://127.0.0.1:6479/0\" >> $GITHUB_ENV\n else\n echo \"Version not found in the mapping.\"\n exit 1\n fi\n\n if (( redis_major_version < 7 )); then\n export REDIS_STACK_EXTRA_ARGS=\"--tls-auth-clients optional --save ''\"\n export REDIS_EXTRA_ARGS=\"--tls-auth-clients optional --save ''\"\n fi\n\n invoke devenv --endpoints=all-stack\n\n else\n echo \"Using redis CE for module tests\"\n if [[ -z \"${CLIENT_LIBS_TEST_STACK_IMAGE_TAG}\" ]]; then\n export CLIENT_LIBS_TEST_STACK_IMAGE_TAG=$REDIS_VERSION\n fi\n echo \"REDIS_MOD_URL=redis://127.0.0.1:6379\" >> $GITHUB_ENV\n invoke devenv --endpoints all\n fi\n\n sleep 10 # time to settle\n echo \"::endgroup::\"\n shell: bash\n\n - name: Run tests\n run: |\n set -e\n\n run_tests() {\n local protocol=$1\n local eventloop=\"\"\n\n if [ \"${{inputs.event-loop}}\" == \"uvloop\" ]; then\n eventloop=\"--uvloop\"\n fi\n\n echo \"::group::RESP${protocol} standalone tests\"\n echo \"REDIS_MOD_URL=${REDIS_MOD_URL}\"\n\n if (( $REDIS_MAJOR_VERSION < 7 )) && [ \"$protocol\" == \"3\" ]; then\n echo \"Skipping module tests: Modules doesn't support RESP3 for Redis versions < 7\"\n invoke standalone-tests --redis-mod-url=${REDIS_MOD_URL} $eventloop --protocol=\"${protocol}\" --extra-markers=\"not redismod and not cp_integration\"\n else\n invoke standalone-tests --redis-mod-url=${REDIS_MOD_URL} $eventloop --protocol=\"${protocol}\"\n fi\n\n echo \"::endgroup::\"\n\n echo \"::group::RESP${protocol} cluster tests\"\n invoke cluster-tests $eventloop --protocol=${protocol}\n echo \"::endgroup::\"\n }\n\n if [ \"${{inputs.protocol}}\" == \"all\" ]; then\n run_tests 2 \"${{inputs.event-loop}}\"\n run_tests 3 \"${{inputs.event-loop}}\"\n else\n run_tests \"${{inputs.protocol}}\" \"${{inputs.event-loop}}\"\n fi\n shell: bash\n\n - name: Debug\n if: failure()\n run: |\n sudo apt-get install -y redis-tools\n echo \"Docker Containers:\"\n docker ps\n echo \"Cluster nodes:\"\n redis-cli -p 16379 CLUSTER NODES\n shell: bash\n\n - name: Upload test results and profiling data\n uses: actions/upload-artifact@v4\n with:\n name: pytest-results-redis_${{inputs.redis-version}}-python_${{inputs.python-version}}-parser_${{env.PARSER_BACKEND}}-el_${{inputs.event-loop}}-protocol_${{inputs.protocol}}\n path: |\n *-results.xml\n prof/**\n profile_output*\n if-no-files-found: error\n retention-days: 10\n\n - name: Upload codecov coverage\n uses: codecov/codecov-action@v4\n with:\n fail_ci_if_error: false\n" }, { "path": ".github/dependabot.yml", "content": "version: 2\nupdates:\n - package-ecosystem: \"github-actions\"\n directory: \"/\"\n labels:\n - \"maintenance\"\n schedule:\n interval: \"monthly\"\n" }, { "path": ".github/release-drafter-config.yml", "content": "name-template: '$NEXT_MINOR_VERSION'\ntag-template: 'v$NEXT_MINOR_VERSION'\nfilter-by-commitish: true\ncommitish: master\nautolabeler:\n - label: 'maintenance'\n files:\n - '*.md'\n - '.github/*'\n - label: 'bug'\n branch:\n - '/bug-.+'\n - label: 'maintenance'\n branch:\n - '/maintenance-.+'\n - label: 'feature'\n branch:\n - '/feature-.+'\ncategories:\n - title: '🔥 Breaking Changes'\n labels:\n - 'breakingchange'\n - title: '🧪 Experimental Features'\n labels:\n - 'experimental'\n - title: '🚀 New Features'\n labels:\n - 'feature'\n - 'enhancement'\n - title: '🐛 Bug Fixes'\n labels:\n - 'fix'\n - 'bugfix'\n - 'bug'\n - 'BUG'\n - title: '🧰 Maintenance'\n labels:\n - 'maintenance'\n - 'dependencies'\n - 'documentation'\n - 'docs'\n - 'testing'\nchange-template: '- $TITLE (#$NUMBER)'\nexclude-labels:\n - 'skip-changelog'\ntemplate: |\n # Changes\n\n $CHANGES\n\n ## Contributors\n We'd like to thank all the contributors who worked on this release!\n\n $CONTRIBUTORS\n\n" }, { "path": ".github/spellcheck-settings.yml", "content": "matrix:\n- name: Markdown\n expect_match: false\n apsell:\n lang: en\n d: en_US\n ignore-case: true\n dictionary:\n wordlists:\n - .github/wordlist.txt\n output: wordlist.dic\n pipeline:\n - pyspelling.filters.markdown:\n markdown_extensions:\n - markdown.extensions.extra:\n - pyspelling.filters.html:\n comments: false\n attributes:\n - alt\n ignores:\n - ':matches(code, pre)'\n - code\n - pre\n - blockquote\n - img\n sources:\n - '*.md'\n - 'docs/*.rst'\n - 'docs/*.ipynb'\n" }, { "path": ".github/wordlist.txt", "content": "APM\nARGV\nBFCommands\nbalancer\nCacheImpl\ncancelling\nCAS\nCFCommands\nCMSCommands\nClusterNode\nClusterNodes\nClusterPipeline\nClusterPubSub\nConnectionPool\nconfig\nCoreCommands\nCSC\nDatabaseConfig\nDNS\nEchoHealthCheck\nEVAL\nEVALSHA\nfailover\nFQDN\nGrokzen's\nhandoff\nHealthcheck\nHealthCheckPolicies\nhealthcheck\nhealthchecks\nINCR\ninit\nIOError\nInstrumentations\nJSONCommands\nJaeger\nLudovico\nMagnocavallo\nMeterProvider\nMetricGroup\nmillis\nMultiDbConfig\nMultiDBClient\nMcCurdy\nNOSCRIPT\nNoValidDatabaseException\nNUMPAT\nNUMPT\nNUMSUB\nOSS\nOpenCensus\nOpenTelemetry\nOpenTracing\nOtel\notel\nOTelConfig\notlp\nOTLPMetricExporter\nPeriodicExportingMetricReader\nproto\nPubSub\nREADONLY\nRediSearch\nRedisBloom\nRedisCluster\nRedisClusterCommands\nRedisClusterException\nRedisClusters\nRedisInstrumentor\nRedisJSON\nRedisTimeSeries\nSHA\nSLA\nSearchCommands\nSentinelCommands\nSentinelConnectionPool\nSharded\nSolovyov\nSpanKind\nSpecfiying\nStatusCode\nTCP\nTemporaryUnavailableException\nTLS\nTOPKCommands\nTimeSeriesCommands\nUptrace\nValueError\nWATCHed\nWatchError\napi\nargs\nasync\nasyncio\nautoclass\nautomodule\nbackoff\nbdb\nbehaviour\nbool\nboolean\nbooleans\nbysource\ncharset\ndel\ndev\ndocstring\ndocstrings\neg\nexc\nfirsttimersonly\nfo\ngenindex\ngmail\nhiredis\nhttp\nidx\niff\nini\njson\nkeyslot\nkeyspace\nkwarg\nkwargs\nlinters\nlocalhost\nlua\nmakeapullrequest\nmaxdepth\nmget\nmicroservice\nmicroservices\nmset\nmultikey\nmykey\nnonatomic\nobservability\nopentelemetry\noss\nperformant\npmessage\npng\npre\npsubscribe\npubsub\npunsubscribe\npy\npypi\nquickstart\nreadonly\nreadwrite\nredis\nredismodules\nreinitialization\nreplicaof\nrepo\nruntime\nsedrik\nsharded\nsdk\nssl\nstr\nstunnel\nsubcommands\nthevalueofmykey\ntimeseries\ntoctree\ntopk\ntriaging\ntxt\nun\nunicode\nurl\nvirtualenv\nwww\nXREAD\nXREADGROUP\nyaml\n" }, { "path": ".github/workflows/codeql-analysis.yml", "content": "# For most projects, this workflow file will not need changing; you simply need\n# to commit it to your repository.\n#\n# You may wish to alter this file to override the set of languages analyzed,\n# or to provide custom queries or build logic.\n#\n# ******** NOTE ********\n# We have attempted to detect the languages in your repository. Please check\n# the `language` matrix defined below to confirm you have the correct set of\n# supported CodeQL languages.\n#\nname: \"CodeQL\"\n\non:\n push:\n branches: [ master ]\n pull_request:\n # The branches below must be a subset of the branches above\n branches: [ master ]\n\njobs:\n analyze:\n name: Analyze\n runs-on: ubuntu-latest\n permissions:\n actions: read\n contents: read\n security-events: write\n\n strategy:\n fail-fast: false\n matrix:\n language: [ 'python' ]\n # CodeQL supports [ 'cpp', 'csharp', 'go', 'java', 'javascript', 'python', 'ruby' ]\n # Learn more about CodeQL language support at https://git.io/codeql-language-support\n\n steps:\n - name: Checkout repository\n uses: actions/checkout@v6\n\n # Initializes the CodeQL tools for scanning.\n - name: Initialize CodeQL\n uses: github/codeql-action/init@v4\n with:\n languages: ${{ matrix.language }}\n # If you wish to specify custom queries, you can do so here or in a config file.\n # By default, queries listed here will override any specified in a config file.\n # Prefix the list here with \"+\" to use these queries and those in the config file.\n # queries: ./path/to/local/query, your-org/your-repo/queries@main\n\n # Autobuild attempts to build any compiled languages (C/C++, C#, or Java).\n # If this step fails, then you should remove it and run the build manually (see below)\n - name: Autobuild\n uses: github/codeql-action/autobuild@v4\n\n # ℹ️ Command-line programs to run using the OS shell.\n # 📚 https://git.io/JvXDl\n\n # ✏️ If the Autobuild fails above, remove it and uncomment the following three lines\n # and modify them (or add more) to build your code if your project\n # uses a compiled language\n\n #- run: |\n # make bootstrap\n # make release\n\n - name: Perform CodeQL Analysis\n uses: github/codeql-action/analyze@v4\n" }, { "path": ".github/workflows/docs.yaml", "content": "name: Docs CI\n\non:\n push:\n branches:\n - master\n - '[0-9].[0-9]'\n pull_request:\n branches:\n - master\n - '[0-9].[0-9]'\n schedule:\n - cron: '0 1 * * *' # nightly build\n\nconcurrency:\n group: ${{ github.event.pull_request.number || github.ref }}-docs\n cancel-in-progress: true\n\npermissions:\n contents: read # to fetch code (actions/checkout)\n\njobs:\n\n build-docs:\n name: Build docs\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v6\n - uses: actions/setup-python@v6\n with:\n python-version: \"3.10\"\n cache: 'pip'\n - name: install deps\n run: |\n sudo apt-get update -yqq\n sudo apt-get install -yqq pandoc make\n - name: run code linters\n run: |\n pip install -r dev_requirements.txt -r docs/requirements.txt\n invoke build-docs\n\n - name: upload docs\n uses: actions/upload-artifact@v7\n with:\n name: redis-py-docs\n path: |\n docs/_build/html\n" }, { "path": ".github/workflows/hiredis-py-integration.yaml", "content": "name: Hiredis-py integration tests\n\non:\n workflow_dispatch:\n inputs:\n redis-py-branch:\n description: 'redis-py branch to run tests on'\n required: true\n default: 'master'\n hiredis-branch:\n description: 'hiredis-py branch to run tests on'\n required: true\n default: 'master'\n\nconcurrency:\n group: ${{ github.event.pull_request.number || github.ref }}-hiredis-integration\n cancel-in-progress: true\n\npermissions:\n contents: read # to fetch code (actions/checkout)\n\nenv:\n CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}\n # this speeds up coverage with Python 3.12: https://github.com/nedbat/coveragepy/issues/1665\n COVERAGE_CORE: sysmon\n CURRENT_CLIENT_LIBS_TEST_STACK_IMAGE_TAG: '8.4.0'\n CURRENT_REDIS_VERSION: '8.4.0'\n\njobs:\n redis_version:\n runs-on: ubuntu-latest\n outputs:\n CURRENT: ${{ env.CURRENT_REDIS_VERSION }}\n steps:\n - name: Compute outputs\n run: |\n echo \"CURRENT=${{ env.CURRENT_REDIS_VERSION }}\" >> $GITHUB_OUTPUT\n\n hiredis-tests:\n runs-on: ubuntu-latest\n needs: [redis_version]\n timeout-minutes: 60\n strategy:\n max-parallel: 15\n fail-fast: false\n matrix:\n redis-version: [ '${{ needs.redis_version.outputs.CURRENT }}' ]\n python-version: [ '3.10', '3.14']\n parser-backend: [ 'hiredis' ]\n hiredis-version: [ 'unstable' ]\n event-loop: [ 'asyncio' ]\n env:\n ACTIONS_ALLOW_UNSECURE_COMMANDS: true\n name: Redis ${{ matrix.redis-version }}; Python ${{ matrix.python-version }}; RESP Parser:${{matrix.parser-backend}} (${{ matrix.hiredis-version }}); EL:${{matrix.event-loop}}\n steps:\n - uses: actions/checkout@v6\n with:\n ref: ${{ inputs.redis-py-branch }}\n - name: Run tests\n uses: ./.github/actions/run-tests\n with:\n python-version: ${{ matrix.python-version }}\n parser-backend: ${{ matrix.parser-backend }}\n redis-version: ${{ matrix.redis-version }}\n hiredis-version: ${{ matrix.hiredis-version }}\n hiredis-branch: ${{ inputs.hiredis-branch }}\n" }, { "path": ".github/workflows/install_and_test.sh", "content": "#!/bin/bash\n\nset -e\n\nSUFFIX=$1\nif [ -z ${SUFFIX} ]; then\n echo \"Supply valid python package extension such as whl or tar.gz. Exiting.\"\n exit 3\nfi\n\nscript=`pwd`/${BASH_SOURCE[0]}\nHERE=`dirname ${script}`\nROOT=`realpath ${HERE}/../..`\n\ncd ${ROOT}\nDESTENV=${ROOT}/.venvforinstall\nif [ -d ${DESTENV} ]; then\n rm -rf ${DESTENV}\nfi\npython -m venv ${DESTENV}\nsource ${DESTENV}/bin/activate\npip install --upgrade --quiet pip\npip install --quiet -r dev_requirements.txt\npip uninstall -y redis # uninstall Redis package installed via redis-entraid\ninvoke devenv --endpoints=all-stack\ninvoke package\n\n# find packages\nPKG=`ls ${ROOT}/dist/*.${SUFFIX}`\nls -l ${PKG}\n\nTESTDIR=${ROOT}/STAGETESTS\nif [ -d ${TESTDIR} ]; then\n rm -rf ${TESTDIR}\nfi\nmkdir ${TESTDIR}\ncp -R ${ROOT}/tests ${TESTDIR}/tests\ncd ${TESTDIR}\n\n# install, run tests\npip install ${PKG}\n# Redis tests\npytest -m 'not onlycluster' --ignore=tests/test_scenario --ignore=tests/test_asyncio/test_scenario\n# RedisCluster tests\nCLUSTER_URL=\"redis://localhost:16379/0\"\nCLUSTER_SSL_URL=\"rediss://localhost:27379/0\"\npytest -m 'not onlynoncluster and not redismod and not ssl' \\\n --ignore=tests/test_scenario \\\n --ignore=tests/test_asyncio/test_scenario \\\n --redis-url=\"${CLUSTER_URL}\" \\\n --redis-ssl-url=\"${CLUSTER_SSL_URL}\"\n" }, { "path": ".github/workflows/integration.yaml", "content": "name: CI\n\non:\n push:\n paths-ignore:\n - 'docs/**'\n - '**/*.rst'\n - '**/*.md'\n branches:\n - master\n - '[0-9].[0-9]'\n pull_request:\n branches:\n - master\n - '[0-9].[0-9]'\n schedule:\n - cron: '0 1 * * *' # nightly build\n\nconcurrency:\n group: ${{ github.event.pull_request.number || github.ref }}-integration\n cancel-in-progress: true\n\npermissions:\n contents: read # to fetch code (actions/checkout)\n\nenv:\n CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}\n # this speeds up coverage with Python 3.12: https://github.com/nedbat/coveragepy/issues/1665\n COVERAGE_CORE: sysmon\n # patch releases get included in the base version image when they are published\n # for example after 8.2.1 is published, 8.2 image contains 8.2.1 content\n CURRENT_CLIENT_LIBS_TEST_STACK_IMAGE_TAG: '8.4.0'\n CURRENT_REDIS_VERSION: '8.4.0'\n REDIS_VERSION_CUSTOM_MAP: 'custom-21860421418-debian-amd64:8.6'\n\njobs:\n dependency-audit:\n name: Dependency audit\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v6\n - uses: pypa/gh-action-pip-audit@v1.0.8\n with:\n inputs: dev_requirements.txt\n ignore-vulns: |\n GHSA-w596-4wvx-j9j6 # subversion related git pull, dependency for pytest. There is no impact here.\n CVE-2026-26007 # dependency for entraid tests\n CVE-2026-32597 # PyJWT does not validate the crit (Critical) Header Parameter defined in RFC 7515, this will be fixed in the next release\n\n lint:\n name: Code linters\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v6\n - uses: actions/setup-python@v6\n with:\n python-version: \"3.10\"\n cache: 'pip'\n - name: run code linters\n run: |\n pip install -r dev_requirements.txt\n pip uninstall -y redis # uninstall Redis package installed via redis-entraid\n invoke linters\n\n redis_version:\n runs-on: ubuntu-latest\n outputs:\n CURRENT: ${{ env.CURRENT_REDIS_VERSION }}\n steps:\n - name: Compute outputs\n run: |\n echo \"CURRENT=${{ env.CURRENT_REDIS_VERSION }}\" >> $GITHUB_OUTPUT\n\n tests:\n runs-on: ubuntu-latest\n timeout-minutes: 60\n needs: redis_version\n strategy:\n max-parallel: 30\n fail-fast: false\n matrix:\n redis-version: ['custom-21860421418-debian-amd64', '${{ needs.redis_version.outputs.CURRENT }}', '8.2', '8.0.2' ,'7.4.4', '7.2.9']\n python-version: ['3.10', '3.14']\n parser-backend: ['plain']\n event-loop: ['asyncio']\n protocol: ['2', '3']\n env:\n ACTIONS_ALLOW_UNSECURE_COMMANDS: true\n name: Redis ${{ matrix.redis-version }}; Python ${{ matrix.python-version }}; RESP Parser:${{matrix.parser-backend}}; EL:${{matrix.event-loop}}; Protocol:${{matrix.protocol}}\n steps:\n - uses: actions/checkout@v6\n - name: Run tests\n uses: ./.github/actions/run-tests\n with:\n python-version: ${{ matrix.python-version }}\n parser-backend: ${{ matrix.parser-backend }}\n redis-version: ${{ matrix.redis-version }}\n protocol: ${{ matrix.protocol }}\n\n python-compatibility-tests:\n runs-on: ubuntu-latest\n needs: [ redis_version ]\n timeout-minutes: 60\n strategy:\n max-parallel: 30\n fail-fast: false\n matrix:\n redis-version: [ '${{ needs.redis_version.outputs.CURRENT }}' ]\n python-version: ['3.11', '3.12', '3.13']\n parser-backend: [ 'plain' ]\n event-loop: [ 'asyncio' ]\n protocol: ['2', '3']\n env:\n ACTIONS_ALLOW_UNSECURE_COMMANDS: true\n name: Redis ${{ matrix.redis-version }}; Python ${{ matrix.python-version }}; RESP Parser:${{matrix.parser-backend}}; EL:${{matrix.event-loop}}; Protocol:${{matrix.protocol}}\n steps:\n - uses: actions/checkout@v6\n - name: Run tests\n uses: ./.github/actions/run-tests\n with:\n python-version: ${{ matrix.python-version }}\n parser-backend: ${{ matrix.parser-backend }}\n redis-version: ${{ matrix.redis-version }}\n protocol: ${{ matrix.protocol }}\n\n pypy-compatibility-tests:\n runs-on: ubuntu-latest\n needs: [ redis_version ]\n # in pipeline ofter these pypy jobs hang, so their timeout is set\n # to just 50 minutes - close to the max time the jobs are running at the moment\n # adding more tests will make them even slower, and we will need to adjust the timeout\n timeout-minutes: 50\n strategy:\n max-parallel: 30\n fail-fast: false\n matrix:\n redis-version: [ '${{ needs.redis_version.outputs.CURRENT }}' ]\n python-version: ['pypy-3.10', 'pypy-3.11']\n parser-backend: [ 'plain' ]\n event-loop: [ 'asyncio' ]\n protocol: ['2', '3']\n env:\n ACTIONS_ALLOW_UNSECURE_COMMANDS: true\n name: PyPy Redis ${{ matrix.redis-version }}; ${{ matrix.python-version }}; RESP Parser:${{matrix.parser-backend}}; EL:${{matrix.event-loop}}; Protocol:${{matrix.protocol}}\n steps:\n - uses: actions/checkout@v6\n - name: Run tests\n uses: ./.github/actions/run-tests\n with:\n python-version: ${{ matrix.python-version }}\n parser-backend: ${{ matrix.parser-backend }}\n redis-version: ${{ matrix.redis-version }}\n protocol: ${{ matrix.protocol }}\n\n hiredis-tests:\n runs-on: ubuntu-latest\n needs: [redis_version]\n timeout-minutes: 60\n strategy:\n max-parallel: 30\n fail-fast: false\n matrix:\n redis-version: [ '${{ needs.redis_version.outputs.CURRENT }}' ]\n python-version: [ '3.10', '3.14']\n parser-backend: [ 'hiredis' ]\n hiredis-version: [ '>=3.2.0', '<3.0.0' ]\n event-loop: [ 'asyncio' ]\n protocol: ['2', '3']\n env:\n ACTIONS_ALLOW_UNSECURE_COMMANDS: true\n name: Redis ${{ matrix.redis-version }}; Python ${{ matrix.python-version }}; RESP Parser:${{matrix.parser-backend}} (${{ matrix.hiredis-version }}); EL:${{matrix.event-loop}}; Protocol:${{matrix.protocol}}\n steps:\n - uses: actions/checkout@v6\n - name: Run tests\n uses: ./.github/actions/run-tests\n with:\n python-version: ${{ matrix.python-version }}\n parser-backend: ${{ matrix.parser-backend }}\n redis-version: ${{ matrix.redis-version }}\n hiredis-version: ${{ matrix.hiredis-version }}\n protocol: ${{ matrix.protocol }}\n\n uvloop-tests:\n runs-on: ubuntu-latest\n needs: [redis_version]\n timeout-minutes: 60\n strategy:\n max-parallel: 30\n fail-fast: false\n matrix:\n redis-version: [ '${{ needs.redis_version.outputs.CURRENT }}' ]\n python-version: [ '3.10', '3.14' ]\n parser-backend: [ 'plain' ]\n event-loop: [ 'uvloop' ]\n protocol: ['2', '3']\n env:\n ACTIONS_ALLOW_UNSECURE_COMMANDS: true\n name: Redis ${{ matrix.redis-version }}; Python ${{ matrix.python-version }}; RESP Parser:${{matrix.parser-backend}}; EL:${{matrix.event-loop}}; Protocol:${{matrix.protocol}}\n steps:\n - uses: actions/checkout@v6\n - name: Run tests\n uses: ./.github/actions/run-tests\n with:\n python-version: ${{ matrix.python-version }}\n parser-backend: ${{ matrix.parser-backend }}\n redis-version: ${{ matrix.redis-version }}\n event-loop: ${{ matrix.event-loop }}\n protocol: ${{ matrix.protocol }}\n\n build-and-test-package:\n name: Validate building and installing the package\n runs-on: ubuntu-latest\n needs: [redis_version]\n strategy:\n fail-fast: false\n matrix:\n extension: ['tar.gz', 'whl']\n steps:\n - uses: actions/checkout@v6\n - uses: actions/setup-python@v6\n with:\n python-version: \"3.10\"\n - name: Run installed unit tests\n env:\n CLIENT_LIBS_TEST_IMAGE_TAG: ${{ env.CURRENT_REDIS_VERSION }}\n CLIENT_LIBS_TEST_STACK_IMAGE_TAG: ${{ env.CURRENT_CLIENT_LIBS_TEST_STACK_IMAGE_TAG }}\n run: |\n bash .github/workflows/install_and_test.sh ${{ matrix.extension }}\n\n install-package-from-commit:\n name: Install package from commit hash\n runs-on: ubuntu-latest\n strategy:\n fail-fast: false\n matrix:\n python-version: ['3.10', '3.11', '3.12', '3.13', '3.14', 'pypy-3.10', 'pypy-3.11']\n steps:\n - uses: actions/checkout@v6\n - uses: actions/setup-python@v6\n with:\n python-version: ${{ matrix.python-version }}\n cache: 'pip'\n - name: install from pip\n run: |\n pip install --quiet git+${GITHUB_SERVER_URL}/${GITHUB_REPOSITORY}.git@${GITHUB_SHA}\n" }, { "path": ".github/workflows/pypi-publish.yaml", "content": "name: Publish tag to Pypi\n\non:\n release:\n types: [published]\n workflow_dispatch:\n\npermissions:\n contents: read # to fetch code (actions/checkout)\n\njobs:\n\n build_and_package:\n runs-on: ubuntu-latest\n steps:\n - uses: actions/checkout@v6\n - name: install python\n uses: actions/setup-python@v6\n with:\n python-version: \"3.10\"\n - run: pip install build twine\n\n - name: Build package\n run: python -m build .\n\n - name: Basic package test prior to upload\n run: |\n twine check dist/*\n\n - name: Publish to Pypi\n uses: pypa/gh-action-pypi-publish@release/v1\n with:\n user: __token__\n password: ${{ secrets.PYPI_API_TOKEN }}\n" }, { "path": ".github/workflows/release-drafter.yml", "content": "name: Release Drafter\n\non:\n push:\n # branches to consider in the event; optional, defaults to all\n branches:\n - master\n\npermissions: {}\njobs:\n update_release_draft:\n permissions:\n pull-requests: write # to add label to PR (release-drafter/release-drafter)\n contents: write # to create a github release (release-drafter/release-drafter)\n\n runs-on: ubuntu-latest\n steps:\n # Drafts your next Release notes as Pull Requests are merged into \"master\"\n - uses: release-drafter/release-drafter@v6\n with:\n # (Optional) specify config name to use, relative to .github/. Default: release-drafter.yml\n config-name: release-drafter-config.yml\n env:\n GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}\n" }, { "path": ".github/workflows/spellcheck.yml", "content": "name: spellcheck\non:\n pull_request:\njobs:\n check-spelling:\n runs-on: ubuntu-latest\n steps:\n - name: Checkout\n uses: actions/checkout@v6\n - name: Check Spelling\n uses: rojopolis/spellcheck-github-actions@0.58.0\n with:\n config_path: .github/spellcheck-settings.yml\n task_name: Markdown\n" }, { "path": ".github/workflows/stale-issues.yml", "content": "name: \"Stale Issue Management\"\non:\n schedule:\n # Run daily at midnight UTC\n - cron: \"0 0 * * *\"\n workflow_dispatch: # Allow manual triggering\n\nenv:\n # Default stale policy timeframes\n DAYS_BEFORE_STALE: 365\n DAYS_BEFORE_CLOSE: 30\n\n # Accelerated timeline for needs-information issues\n NEEDS_INFO_DAYS_BEFORE_STALE: 30\n NEEDS_INFO_DAYS_BEFORE_CLOSE: 7\n\njobs:\n stale:\n runs-on: ubuntu-latest\n steps:\n # First step: Handle regular issues (excluding needs-information)\n - name: Mark regular issues as stale\n uses: actions/stale@v10\n with:\n repo-token: ${{ secrets.GITHUB_TOKEN }}\n\n # Default stale policy\n days-before-stale: ${{ env.DAYS_BEFORE_STALE }}\n days-before-close: ${{ env.DAYS_BEFORE_CLOSE }}\n\n # Explicit stale label configuration\n stale-issue-label: \"stale\"\n stale-pr-label: \"stale\"\n\n stale-issue-message: |\n This issue has been automatically marked as stale due to inactivity.\n It will be closed in 30 days if no further activity occurs.\n If you believe this issue is still relevant, please add a comment to keep it open.\n\n close-issue-message: |\n This issue has been automatically closed due to inactivity.\n If you believe this issue is still relevant, please reopen it or create a new issue with updated information.\n\n # Exclude needs-information issues from this step\n exempt-issue-labels: 'no-stale,needs-information'\n\n # Remove stale label when issue/PR becomes active again\n remove-stale-when-updated: true\n\n # Apply to pull requests with same timeline\n days-before-pr-stale: ${{ env.DAYS_BEFORE_STALE }}\n days-before-pr-close: ${{ env.DAYS_BEFORE_CLOSE }}\n\n stale-pr-message: |\n This pull request has been automatically marked as stale due to inactivity.\n It will be closed in 30 days if no further activity occurs.\n\n close-pr-message: |\n This pull request has been automatically closed due to inactivity.\n If you would like to continue this work, please reopen the PR or create a new one.\n\n # Only exclude no-stale PRs (needs-information PRs follow standard timeline)\n exempt-pr-labels: 'no-stale'\n\n # Second step: Handle needs-information issues with accelerated timeline\n - name: Mark needs-information issues as stale\n uses: actions/stale@v10\n with:\n repo-token: ${{ secrets.GITHUB_TOKEN }}\n\n # Accelerated timeline for needs-information\n days-before-stale: ${{ env.NEEDS_INFO_DAYS_BEFORE_STALE }}\n days-before-close: ${{ env.NEEDS_INFO_DAYS_BEFORE_CLOSE }}\n\n # Explicit stale label configuration\n stale-issue-label: \"stale\"\n\n # Only target ISSUES with needs-information label (not PRs)\n only-issue-labels: 'needs-information'\n\n stale-issue-message: |\n This issue has been marked as stale because it requires additional information\n that has not been provided for 30 days. It will be closed in 7 days if the\n requested information is not provided.\n\n close-issue-message: |\n This issue has been closed because the requested information was not provided within the specified timeframe.\n If you can provide the missing information, please reopen this issue or create a new one.\n\n # Disable PR processing for this step\n days-before-pr-stale: -1\n days-before-pr-close: -1\n\n # Remove stale label when issue becomes active again\n remove-stale-when-updated: true\n" }, { "path": ".gitignore", "content": "*.pyc\nredis.egg-info\nbuild/\ndist/\ndump.rdb\n_build\nvagrant/.vagrant\n.python-version\n.cache\n.eggs\n.idea\n.vscode\n.coverage\nenv\nvenv\ncoverage.xml\n.venv*\n*.xml\n.coverage*\nprof\nprofile_output*\ndocker/stunnel/keys\n/dockers/*/node-*/*\n/dockers/*/tls/*\n/dockers/standalone/\n/dockers/cluster/\n/dockers/replica/\n/dockers/sentinel/\n/dockers/redis-stack/\n/experimenting/\n" }, { "path": ".mypy.ini", "content": "[mypy]\n#, docs/examples, tests\nfiles = redis\ncheck_untyped_defs = True\nfollow_imports_for_stubs asyncio.= True\n#disallow_any_decorated = True\ndisallow_subclassing_any = True\n#disallow_untyped_calls = True\ndisallow_untyped_decorators = True\n#disallow_untyped_defs = True\nimplicit_reexport = False\nno_implicit_optional = True\nshow_error_codes = True\nstrict_equality = True\nwarn_incomplete_stub = True\nwarn_redundant_casts = True\nwarn_unreachable = True\nwarn_unused_ignores = True\ndisallow_any_unimported = True\n#warn_return_any = True\n\n[mypy-redis.asyncio.lock]\n# TODO: Remove once locks has been rewritten\nignore_errors = True\n" }, { "path": ".readthedocs.yml", "content": "version: 2\n\npython:\n install:\n - requirements: docs/requirements.txt\n - method: pip\n path: .\n\nbuild:\n os: ubuntu-20.04\n tools:\n python: \"3.10\"\n\nsphinx:\n configuration: docs/conf.py\n" }, { "path": "CHANGES", "content": "This file contains only the changes history before redis-py version 4.0.0\nAfter redis-py version 4.0.0 all changes can be found and tracked in the Release notes pubished in GitHub\n\n* 3.5.3 (June 1, 2020)\n * Restore try/except clauses to __del__ methods. These will be removed\n in 4.0 when more explicit resource management if enforced. #1339\n * Update the master_address when Sentinels promote a new master. #847\n * Update SentinelConnectionPool to not forcefully disconnect other in-use\n connections which can negatively affect threaded applications. #1345\n* 3.5.2 (May 14, 2020)\n * Tune the locking in ConnectionPool.get_connection so that the lock is\n not held while waiting for the socket to establish and validate the\n TCP connection.\n* 3.5.1 (May 9, 2020)\n * Fix for HSET argument validation to allow any non-None key. Thanks\n @AleksMat, #1337, #1341\n* 3.5.0 (April 29, 2020)\n * Removed exception trapping from __del__ methods. redis-py objects that\n hold various resources implement __del__ cleanup methods to release\n those resources when the object goes out of scope. This provides a\n fallback for when these objects aren't explicitly closed by user code.\n Prior to this change any errors encountered in closing these resources\n would be hidden from the user. Thanks @jdufresne. #1281\n * Expanded support for connection strings specifying a username connecting\n to pre-v6 servers. #1274\n * Optimized Lock's blocking_timeout and sleep. If the lock cannot be\n acquired and the sleep value would cause the loop to sleep beyond\n blocking_timeout, fail immediately. Thanks @clslgrnc. #1263\n * Added support for passing Python memoryviews to Redis command args that\n expect strings or bytes. The memoryview instance is sent directly to\n the socket such that there are zero copies made of the underlying data\n during command packing. Thanks @Cody-G. #1265, #1285\n * HSET command now can accept multiple pairs. HMSET has been marked as\n deprecated now. Thanks to @laixintao #1271\n * Don't manually DISCARD when encountering an ExecAbortError.\n Thanks @nickgaya, #1300/#1301\n * Reset the watched state of pipelines after calling exec. This saves\n a roundtrip to the server by not having to call UNWATCH within\n Pipeline.reset(). Thanks @nickgaya, #1299/#1302\n * Added the KEEPTTL option for the SET command. Thanks\n @laixintao #1304/#1280\n * Added the MEMORY STATS command. #1268\n * Lock.extend() now has a new option, `replace_ttl`. When False (the\n default), Lock.extend() adds the `additional_time` to the lock's existing\n TTL. When replace_ttl=True, the lock's existing TTL is replaced with\n the value of `additional_time`.\n * Add testing and support for PyPy.\n* 3.4.1\n * Move the username argument in the Redis and Connection classes to the\n end of the argument list. This helps those poor souls that specify all\n their connection options as non-keyword arguments. #1276\n * Prior to ACL support, redis-py ignored the username component of\n Connection URLs. With ACL support, usernames are no longer ignored and\n are used to authenticate against an ACL rule. Some cloud vendors with\n managed Redis instances (like Heroku) provide connection URLs with a\n username component pre-ACL that is not intended to be used. Sending that\n username to Redis servers < 6.0.0 results in an error. Attempt to detect\n this condition and retry the AUTH command with only the password such\n that authentication continues to work for these users. #1274\n * Removed the __eq__ hooks to Redis and ConnectionPool that were added\n in 3.4.0. This ended up being a bad idea as two separate connection\n pools be considered equal yet manage a completely separate set of\n connections.\n* 3.4.0\n * Allow empty pipelines to be executed if there are WATCHed keys.\n This is a convenient way to test if any of the watched keys changed\n without actually running any other commands. Thanks @brianmaissy.\n #1233, #1234\n * Removed support for end of life Python 3.4.\n * Added support for all ACL commands in Redis 6. Thanks @IAmATeaPot418\n for helping.\n * Pipeline instances now always evaluate to True. Prior to this change,\n pipeline instances relied on __len__ for boolean evaluation which\n meant that pipelines with no commands on the stack would be considered\n False. #994\n * Client instances and Connection pools now support a 'client_name'\n argument. If supplied, all connections created will call CLIENT SETNAME\n as soon as the connection is opened. Thanks to @Habbie for supplying\n the basis of this change. #802\n * Added the 'ssl_check_hostname' argument to specify whether SSL\n connections should require the server hostname to match the hostname\n specified in the SSL cert. By default 'ssl_check_hostname' is False\n for backwards compatibility. #1196\n * Slightly optimized command packing. Thanks @Deneby67. #1255\n * Added support for the TYPE argument to SCAN. Thanks @netocp. #1220\n * Better thread and fork safety in ConnectionPool and\n BlockingConnectionPool. Added better locking to synchronize critical\n sections rather than relying on CPython-specific implementation details\n relating to atomic operations. Adjusted how the pools identify and\n deal with a fork. Added a ChildDeadlockedError exception that is\n raised by child processes in the very unlikely chance that a deadlock\n is encountered. Thanks @gmbnomis, @mdellweg, @yht804421715. #1270,\n #1138, #1178, #906, #1262\n * Added __eq__ hooks to the Redis and ConnectionPool classes.\n Thanks @brainix. #1240\n* 3.3.11\n * Further fix for the SSLError -> TimeoutError mapping to work\n on obscure releases of Python 2.7.\n* 3.3.10\n * Fixed a potential error handling bug for the SSLError -> TimeoutError\n mapping introduced in 3.3.9. Thanks @zbristow. #1224\n* 3.3.9\n * Mapped Python 2.7 SSLError to TimeoutError where appropriate. Timeouts\n should now consistently raise TimeoutErrors on Python 2.7 for both\n unsecured and secured connections. Thanks @zbristow. #1222\n* 3.3.8\n * Fixed MONITOR parsing to properly parse IPv6 client addresses, unix\n socket connections and commands issued from Lua. Thanks @kukey. #1201\n* 3.3.7\n * Fixed a regression introduced in 3.3.0 where socket.error exceptions\n (or subclasses) could potentially be raised instead of\n redis.exceptions.ConnectionError. #1202\n* 3.3.6\n * Fixed a regression in 3.3.5 that caused PubSub.get_message() to raise\n a socket.timeout exception when passing a timeout value. #1200\n* 3.3.5\n * Fix an issue where socket.timeout errors could be handled by the wrong\n exception handler in Python 2.7.\n* 3.3.4\n * More specifically identify nonblocking read errors for both SSL and\n non-SSL connections. 3.3.1, 3.3.2 and 3.3.3 on Python 2.7 could\n potentially mask a ConnectionError. #1197\n* 3.3.3\n * The SSL module in Python < 2.7.9 handles non-blocking sockets\n differently than 2.7.9+. This patch accommodates older versions. #1197\n* 3.3.2\n * Further fixed a regression introduced in 3.3.0 involving SSL and\n non-blocking sockets. #1197\n* 3.3.1\n * Fixed a regression introduced in 3.3.0 involving SSL and non-blocking\n sockets. #1197\n* 3.3.0\n * Resolve a race condition with the PubSubWorkerThread. #1150\n * Cleanup socket read error messages. Thanks Vic Yu. #1159\n * Cleanup the Connection's selector correctly. Thanks Bruce Merry. #1153\n * Added a Monitor object to make working with MONITOR output easy.\n Thanks Roey Prat #1033\n * Internal cleanup: Removed the legacy Token class which was necessary\n with older version of Python that are no longer supported. #1066\n * Response callbacks are now case insensitive. This allows users that\n call Redis.execute_command() directly to pass lower-case command\n names and still get reasonable responses. #1168\n * Added support for hiredis-py 1.0.0 encoding error support. This should\n make the PythonParser and the HiredisParser behave identically\n when encountering encoding errors. Thanks Brian Candler. #1161/#1162\n * All authentication errors now properly raise AuthenticationError.\n AuthenticationError is now a subclass of ConnectionError, which will\n cause the connection to be disconnected and cleaned up appropriately.\n #923\n * Add READONLY and READWRITE commands. Thanks @theodesp. #1114\n * Remove selectors in favor of nonblocking sockets. Selectors had\n issues in some environments including eventlet and gevent. This should\n resolve those issues with no other side effects.\n * Fixed an issue with XCLAIM and previously claimed but not removed\n messages. Thanks @thomdask. #1192/#1191\n * Allow for single connection client instances. These instances\n are not thread safe but offer other benefits including a subtle\n performance increase.\n * Added extensive health checks that keep the connections lively.\n Passing the \"health_check_interval=N\" option to the Redis client class\n or to a ConnectionPool ensures that a round trip PING/PONG is successful\n before any command if the underlying connection has been idle for more\n than N seconds. ConnectionErrors and TimeoutErrors are automatically\n retried once for health checks.\n * Changed the PubSubWorkerThread to use a threading.Event object rather\n than a boolean to control the thread's life cycle. Thanks Timothy\n Rule. #1194/#1195.\n * Fixed a bug in Pipeline error handling that would incorrectly retry\n ConnectionErrors.\n* 3.2.1\n * Fix SentinelConnectionPool to work in multiprocess/forked environments.\n* 3.2.0\n * Added support for `select.poll` to test whether data can be read\n on a socket. This should allow for significantly more connections to\n be used with pubsub. Fixes #486/#1115\n * Attempt to guarantee that the ConnectionPool hands out healthy\n connections. Healthy connections are those that have an established\n socket connection to the Redis server, are ready to accept a command\n and have no data available to read. Fixes #1127/#886\n * Use the socket.IPPROTO_TCP constant instead of socket.SOL_TCP.\n IPPROTO_TCP is available on more interpreters (Jython for instance).\n Thanks @Junnplus. #1130\n * Fixed a regression introduced in 3.0 that mishandles exceptions not\n derived from the base Exception class. KeyboardInterrupt and\n gevent.timeout notable. Thanks Christian Fersch. #1128/#1129\n * Significant improvements to handing connections with forked processes.\n Parent and child processes no longer trample on each others' connections.\n Thanks to Jay Rolette for the patch and highlighting this issue.\n #504/#732/#784/#863\n * PythonParser no longer closes the associated connection's socket. The\n connection itself will close the socket. #1108/#1085\n* 3.1.0\n * Connection URLs must have one of the following schemes:\n redis://, rediss://, unix://. Thanks @jdupl123. #961/#969\n * Fixed an issue with retry_on_timeout logic that caused some TimeoutErrors\n to be retried. Thanks Aaron Yang. #1022/#1023\n * Added support for SNI for SSL. Thanks @oridistor and Roey Prat. #1087\n * Fixed ConnectionPool repr for pools with no connections. Thanks\n Cody Scott. #1043/#995\n * Fixed GEOHASH to return a None value when specifying a place that\n doesn't exist on the server. Thanks @guybe7. #1126\n * Fixed XREADGROUP to return an empty dictionary for messages that\n have been deleted but still exist in the unacknowledged queue. Thanks\n @xeizmendi. #1116\n * Added an owned method to Lock objects. owned returns a boolean\n indicating whether the current lock instance still owns the lock.\n Thanks Dave Johansen. #1112\n * Allow lock.acquire() to accept an optional token argument. If\n provided, the token argument is used as the unique value used to claim\n the lock. Thankd Dave Johansen. #1112\n * Added a reacquire method to Lock objects. reacquire attempts to renew\n the lock such that the timeout is extended to the same value that the\n lock was initially acquired with. Thanks Ihor Kalnytskyi. #1014\n * Stream names found within XREAD and XREADGROUP responses now properly\n respect the decode_responses flag.\n * XPENDING_RANGE now requires the user the specify the min, max and\n count arguments. Newer versions of Redis prevent count from being\n infinite so it's left to the user to specify these values explicitly.\n * ZADD now returns None when xx=True and incr=True and an element\n is specified that doesn't exist in the sorted set. This matches\n what the server returns in this case. #1084\n * Added client_kill_filter that accepts various filters to identify\n and kill clients. Thanks Theofanis Despoudis. #1098\n * Fixed a race condition that occurred when unsubscribing and\n resubscribing to the same channel or pattern in rapid succession.\n Thanks Marcin Raczyński. #764\n * Added a LockNotOwnedError that is raised when trying to extend or\n release a lock that is no longer owned. This is a subclass of LockError\n so previous code should continue to work as expected. Thanks Joshua\n Harlow. #1095\n * Fixed a bug in GEORADIUS that forced decoding of places without\n respecting the decode_responses option. Thanks Bo Bayles. #1082\n* 3.0.1\n * Fixed regression with UnixDomainSocketConnection caused by 3.0.0.\n Thanks Jyrki Muukkonen\n * Fixed an issue with the new asynchronous flag on flushdb and flushall.\n Thanks rogeryen\n * Updated Lock.locked() method to indicate whether *any* process has\n acquired the lock, not just the current one. This is in line with\n the behavior of threading.Lock. Thanks Alan Justino da Silva\n* 3.0.0\n BACKWARDS INCOMPATIBLE CHANGES\n * When using a Lock as a context manager and the lock fails to be acquired\n a LockError is now raised. This prevents the code block inside the\n context manager from being executed if the lock could not be acquired.\n * Renamed LuaLock to Lock.\n * Removed the pipeline based Lock implementation in favor of the LuaLock\n implementation.\n * Only bytes, strings and numbers (ints, longs and floats) are acceptable\n for keys and values. Previously redis-py attempted to cast other types\n to str() and store the result. This caused must confusion and frustration\n when passing boolean values (cast to 'True' and 'False') or None values\n (cast to 'None'). It is now the user's responsibility to cast all\n key names and values to bytes, strings or numbers before passing the\n value to redis-py.\n * The StrictRedis class has been renamed to Redis. StrictRedis will\n continue to exist as an alias of Redis for the foreseeable future.\n * The legacy Redis client class has been removed. It caused much confusion\n to users.\n * ZINCRBY arguments 'value' and 'amount' have swapped order to match the\n the Redis server. The new argument order is: keyname, amount, value.\n * MGET no longer raises an error if zero keys are passed in. Instead an\n empty list is returned.\n * MSET and MSETNX now require all keys/values to be specified in a single\n dictionary argument named mapping. This was changed to allow for future\n options to these commands in the future.\n * ZADD now requires all element names/scores be specified in a single\n dictionary argument named mapping. This was required to allow the NX,\n XX, CH and INCR options to be specified.\n * ssl_cert_reqs now has a default value of 'required' by default. This\n should make connecting to a remote Redis server over SSL more secure.\n Thanks u2mejc\n * Removed support for EOL Python 2.6 and 3.3. Thanks jdufresne\n OTHER CHANGES\n * Added missing DECRBY command. Thanks derek-dchu\n * CLUSTER INFO and CLUSTER NODES responses are now properly decoded to\n strings.\n * Added a 'locked()' method to Lock objects. This method returns True\n if the lock has been acquired and owned by the current process,\n otherwise False.\n * EXISTS now supports multiple keys. It's return value is now the number\n of keys in the list that exist.\n * Ensure all commands can accept key names as bytes. This fixes issues\n with BLPOP, BRPOP and SORT.\n * All errors resulting from bad user input are raised as DataError\n exceptions. DataError is a subclass of RedisError so this should be\n transparent to anyone previously catching these.\n * Added support for NX, XX, CH and INCR options to ZADD\n * Added support for the MIGRATE command\n * Added support for the MEMORY USAGE and MEMORY PURGE commands. Thanks\n Itamar Haber\n * Added support for the 'asynchronous' argument to FLUSHDB and FLUSHALL\n commands. Thanks Itamar Haber\n * Added support for the BITFIELD command. Thanks Charles Leifer and\n Itamar Haber\n * Improved performance on pipeline requests with large chunks of data.\n Thanks tzickel\n * Fixed test suite to not fail if another client is connected to the\n server the tests are running against.\n * Added support for SWAPDB. Thanks Itamar Haber\n * Added support for all STREAM commands. Thanks Roey Prat and Itamar Haber\n * SHUTDOWN now accepts the 'save' and 'nosave' arguments. Thanks\n dwilliams-kenzan\n * Added support for ZPOPMAX, ZPOPMIN, BZPOPMAX, BZPOPMIN. Thanks\n Itamar Haber\n * Added support for the 'type' argument in CLIENT LIST. Thanks Roey Prat\n * Added support for CLIENT PAUSE. Thanks Roey Prat\n * Added support for CLIENT ID and CLIENT UNBLOCK. Thanks Itamar Haber\n * GEODIST now returns a None value when referencing a place that does\n not exist. Thanks qingping209\n * Added a ping() method to pubsub objects. Thanks krishan-carbon\n * Fixed a bug with keys in the INFO dict that contained ':' symbols.\n Thanks mzalimeni\n * Fixed the select system call retry compatibility with Python 2.x.\n Thanks lddubeau\n * max_connections is now a valid querystring argument for creating\n connection pools from URLs. Thanks mmaslowskicc\n * Added the UNLINK command. Thanks yozel\n * Added socket_type option to Connection for configurability.\n Thanks garlicnation\n * Lock.do_acquire now atomically sets acquires the lock and sets the\n expire value via set(nx=True, px=timeout). Thanks 23doors\n * Added 'count' argument to SPOP. Thanks AlirezaSadeghi\n * Fixed an issue parsing client_list responses that contained an '='.\n Thanks swilly22\n* 2.10.6\n * Various performance improvements. Thanks cjsimpson\n * Fixed a bug with SRANDMEMBER where the behavior for `number=0` did\n not match the spec. Thanks Alex Wang\n * Added HSTRLEN command. Thanks Alexander Putilin\n * Added the TOUCH command. Thanks Anis Jonischkeit\n * Remove unnecessary calls to the server when registering Lua scripts.\n Thanks Ben Greenberg\n * SET's EX and PX arguments now allow values of zero. Thanks huangqiyin\n * Added PUBSUB {CHANNELS, NUMPAT, NUMSUB} commands. Thanks Angus Pearson\n * PubSub connections that encounter `InterruptedError`s now\n retry automatically. Thanks Carlton Gibson and Seth M. Larson\n * LPUSH and RPUSH commands run on PyPy now correctly returns the number\n of items of the list. Thanks Jeong YunWon\n * Added support to automatically retry socket EINTR errors. Thanks\n Thomas Steinacher\n * PubSubWorker threads started with `run_in_thread` are now daemonized\n so the thread shuts down when the running process goes away. Thanks\n Keith Ainsworth\n * Added support for GEO commands. Thanks Pau Freixes, Alex DeBrie and\n Abraham Toriz\n * Made client construction from URLs smarter. Thanks Tim Savage\n * Added support for CLUSTER * commands. Thanks Andy Huang\n * The RESTORE command now accepts an optional `replace` boolean.\n Thanks Yoshinari Takaoka\n * Attempt to connect to a new Sentinel if a TimeoutError occurs. Thanks\n Bo Lopker\n * Fixed a bug in the client's `__getitem__` where a KeyError would be\n raised if the value returned by the server is an empty string.\n Thanks Javier Candeira.\n * Socket timeouts when connecting to a server are now properly raised\n as TimeoutErrors.\n* 2.10.5\n * Allow URL encoded parameters in Redis URLs. Characters like a \"/\" can\n now be URL encoded and redis-py will correctly decode them. Thanks\n Paul Keene.\n * Added support for the WAIT command. Thanks \n * Better shutdown support for the PubSub Worker Thread. It now properly\n cleans up the connection, unsubscribes from any channels and patterns\n previously subscribed to and consumes any waiting messages on the socket.\n * Added the ability to sleep for a brief period in the event of a\n WatchError occurring. Thanks Joshua Harlow.\n * Fixed a bug with pipeline error reporting when dealing with characters\n in error messages that could not be encoded to the connection's\n character set. Thanks Hendrik Muhs.\n * Fixed a bug in Sentinel connections that would inadvertently connect\n to the master when the connection pool resets. Thanks\n \n * Better timeout support in Pubsub get_message. Thanks Andy Isaacson.\n * Fixed a bug with the HiredisParser that would cause the parser to\n get stuck in an endless loop if a specific number of bytes were\n delivered from the socket. This fix also increases performance of\n parsing large responses from the Redis server.\n * Added support for ZREVRANGEBYLEX.\n * ConnectionErrors are now raised if Redis refuses a connection due to\n the maxclients limit being exceeded. Thanks Roman Karpovich.\n * max_connections can now be set when instantiating client instances.\n Thanks Ohad Perry.\n* 2.10.4\n (skipped due to a PyPI snafu)\n* 2.10.3\n * Fixed a bug with the bytearray support introduced in 2.10.2. Thanks\n Josh Owen.\n* 2.10.2\n * Added support for Hiredis's new bytearray support. Thanks\n \n * POSSIBLE BACKWARDS INCOMPATIBLE CHANGE: Fixed a possible race condition\n when multiple threads share the same Lock instance with a timeout. Lock\n tokens are now stored in thread local storage by default. If you have\n code that acquires a lock in one thread and passes that lock instance to\n another thread to release it, you need to disable thread local storage.\n Refer to the doc strings on the Lock class about the thread_local\n argument information.\n * Fixed a regression in from_url where \"charset\" and \"errors\" weren't\n valid options. \"encoding\" and \"encoding_errors\" are still accepted\n and preferred.\n * The \"charset\" and \"errors\" options have been deprecated. Passing\n either to StrictRedis.__init__ or from_url will still work but will\n also emit a DeprecationWarning. Instead use the \"encoding\" and\n \"encoding_errors\" options.\n * Fixed a compatibility bug with Python 3 when the server closes a\n connection.\n * Added BITPOS command. Thanks .\n * Fixed a bug when attempting to send large values to Redis in a Pipeline.\n* 2.10.1\n * Fixed a bug where Sentinel connections to a server that's no longer a\n master and receives a READONLY error will disconnect and reconnect to\n the master.\n* 2.10.0\n * Discontinued support for Python 2.5. Upgrade. You'll be happier.\n * The HiRedis parser will now properly raise ConnectionErrors.\n * Completely refactored PubSub support. Fixes all known PubSub bugs and\n adds a bunch of new features. Docs can be found in the README under the\n new \"Publish / Subscribe\" section.\n * Added the new HyperLogLog commands (PFADD, PFCOUNT, PFMERGE). Thanks\n Pepijn de Vos and Vincent Ohprecio.\n * Updated TTL and PTTL commands with Redis 2.8+ semantics. Thanks Markus\n Kaiserswerth.\n * *SCAN commands now return a long (int on Python3) cursor value rather\n than the string representation. This might be slightly backwards\nincompatible in code using*SCAN commands loops such as\n \"while cursor != '0':\".\n * Added extra *SCAN commands that return iterators instead of the normal\n [cursor, data] type. Use scan_iter, hscan_iter, sscan_iter, and\n zscan_iter for iterators. Thanks Mathieu Longtin.\n * Added support for SLOWLOG commands. Thanks Rick van Hattem.\n * Added lexicographical commands ZRANGEBYLEX, ZREMRANGEBYLEX, and ZLEXCOUNT\n for sorted sets.\n * Connection objects now support an optional argument, socket_read_size,\n indicating how much data to read during each socket.recv() call. After\n benchmarking, increased the default size to 64k, which dramatically\n improves performance when fetching large values, such as many results\n in a pipeline or a large (>1MB) string value.\n * Improved the pack_command and send_packed_command functions to increase\n performance when sending large (>1MB) values.\n * Sentinel Connections to master servers now detect when a READONLY error\n is encountered and disconnect themselves and all other active connections\n to the same master so that the new master can be discovered.\n * Fixed Sentinel state parsing on Python 3.\n * Added support for SENTINEL MONITOR, SENTINEL REMOVE, and SENTINEL SET\n commands. Thanks Greg Murphy.\n * INFO output that doesn't follow the \"key:value\" format will now be\n appended to a key named \"__raw__\" in the INFO dictionary. Thanks Pedro\n Larroy.\n * The \"vagrant\" directory contains a complete vagrant environment for\n redis-py developers. The environment runs a Redis master, a Redis slave,\n and 3 Sentinels. Future iterations of the test suite will incorporate\n more integration style tests, ensuring things like failover happen\n correctly.\n * It's now possible to create connection pool instances from a URL.\n StrictRedis.from_url() now uses this feature to create a connection pool\n instance and use that when creating a new client instance. Thanks\n \n * When creating client instances or connection pool instances from an URL,\n it's now possible to pass additional options to the connection pool with\n querystring arguments.\n * Fixed a bug where some encodings (like utf-16) were unusable on Python 3\n as command names and literals would get encoded.\n * Added an SSLConnection class that allows for secure connections through\n stunnel or other means. Construct an SSL connection with the ssl=True\n option on client classes, using the rediss:// scheme from an URL, or\n by passing the SSLConnection class to a connection pool's\n connection_class argument. Thanks .\n * Added a socket_connect_timeout option to control how long to wait while\n establishing a TCP connection before timing out. This lets the client\n fail fast when attempting to connect to a downed server while keeping\n a more lenient timeout for all other socket operations.\n * Added TCP Keep-alive support by passing use the socket_keepalive=True\n option. Finer grain control can be achieved using the\n socket_keepalive_options option which expects a dictionary with any of\n the keys (socket.TCP_KEEPIDLE, socket.TCP_KEEPCNT, socket.TCP_KEEPINTVL)\n and integers for values. Thanks Yossi Gottlieb.\n * Added a `retry_on_timeout` option that controls how socket.timeout errors\n are handled. By default it is set to False and will cause the client to\n raise a TimeoutError anytime a socket.timeout is encountered. If\n `retry_on_timeout` is set to True, the client will retry a command that\n timed out once like other `socket.error`s.\n * Completely refactored the Lock system. There is now a LuaLock class\n that's used when the Redis server is capable of running Lua scripts along\n with a fallback class for Redis servers < 2.6. The new locks fix several\n subtle race consider that the old lock could face. In additional, a\n new method, \"extend\" is available on lock instances that all a lock\n owner to extend the amount of time they have the lock for. Thanks to\n Eli Finkelshteyn and for contributions.\n* 2.9.1\n * IPv6 support. Thanks \n* 2.9.0\n * Performance improvement for packing commands when using the PythonParser.\n Thanks Guillaume Viot.\n * Executing an empty pipeline transaction no longer sends MULTI/EXEC to\n the server. Thanks EliFinkelshteyn.\n * Errors when authenticating (incorrect password) and selecting a database\n now close the socket.\n * Full Sentinel support thanks to Vitja Makarov. Thanks!\n * Better repr support for client and connection pool instances. Thanks\n Mark Roberts.\n * Error messages that the server sends to the client are now included\n in the client error message. Thanks Sangjin Lim.\n * Added the SCAN, SSCAN, HSCAN, and ZSCAN commands. Thanks Jingchao Hu.\n * ResponseErrors generated by pipeline execution provide addition context\n including the position of the command in the pipeline and the actual\n command text generated the error.\n * ConnectionPools now play nicer in threaded environments that fork. Thanks\n Christian Joergensen.\n* 2.8.0\n * redis-py should play better with gevent when a gevent Timeout is raised.\n Thanks leifkb.\n * Added SENTINEL command. Thanks Anna Janackova.\n * Fixed a bug where pipelines could potentially corrupt a connection\n if the MULTI command generated a ResponseError. Thanks EliFinkelshteyn\n for the report.\n * Connections now call socket.shutdown() prior to socket.close() to\n ensure communication ends immediately per the note at\n \n Thanks to David Martin for pointing this out.\n * Lock checks are now based on floats rather than ints. Thanks\n Vitja Makarov.\n* 2.7.6\n * Added CONFIG RESETSTAT command. Thanks Yossi Gottlieb.\n * Fixed a bug introduced in 2.7.3 that caused issues with script objects\n and pipelines. Thanks Carpentier Pierre-Francois.\n * Converted redis-py's test suite to use the awesome py.test library.\n * Fixed a bug introduced in 2.7.5 that prevented a ConnectionError from\n being raised when the Redis server is LOADING data.\n * Added a BusyLoadingError exception that's raised when the Redis server\n is starting up and not accepting commands yet. BusyLoadingError\n subclasses ConnectionError, which this state previously returned.\n Thanks Yossi Gottlieb.\n* 2.7.5\n * DEL, HDEL and ZREM commands now return the numbers of keys deleted\n instead of just True/False.\n * from_url now supports URIs with a port number. Thanks Aaron Westendorf.\n* 2.7.4\n * Added missing INCRBY method. Thanks Krzysztof Dorosz.\n * SET now accepts the EX, PX, NX and XX options from Redis 2.6.12. These\n options will generate errors if these options are used when connected\n to a Redis server < 2.6.12. Thanks George Yoshida.\n* 2.7.3\n * Fixed a bug with BRPOPLPUSH and lists with empty strings.\n * All empty except: clauses have been replaced to only catch Exception\n subclasses. This prevents a KeyboardInterrupt from triggering exception\n handlers. Thanks Lucian Branescu Mihaila.\n * All exceptions that are the result of redis server errors now share a\n command Exception subclass, ServerError. Thanks Matt Robenolt.\n * Prevent DISCARD from being called if MULTI wasn't also called. Thanks\n Pete Aykroyd.\n * SREM now returns an integer indicating the number of items removed from\n the set. Thanks .\n * Fixed a bug with BGSAVE and BGREWRITEAOF response callbacks with Python3.\n Thanks Nathan Wan.\n * Added CLIENT GETNAME and CLIENT SETNAME commands.\n Thanks .\n * It's now possible to use len() on a pipeline instance to determine the\n number of commands that will be executed. Thanks Jon Parise.\n * Fixed a bug in INFO's parse routine with floating point numbers. Thanks\n Ali Onur Uyar.\n * Fixed a bug with BITCOUNT to allow `start` and `end` to both be zero.\n Thanks Tim Bart.\n * The transaction() method now accepts a boolean keyword argument,\n value_from_callable. By default, or if False is passes, the transaction()\n method will return the value of the pipelines execution. Otherwise, it\n will return whatever func() returns.\n * Python3 compatibility fix ensuring we're not already bytes(). Thanks\n Salimane Adjao Moustapha.\n * Added PSETEX. Thanks YAMAMOTO Takashi.\n * Added a BlockingConnectionPool to limit the number of connections that\n can be created. Thanks James Arthur.\n * SORT now accepts a `groups` option that if specified, will return\n tuples of n-length, where n is the number of keys specified in the GET\n argument. This allows for convenient row-based iteration. Thanks\n Ionuț Arțăriși.\n* 2.7.2\n * Parse errors are now *always* raised on multi/exec pipelines, regardless\n of the `raise_on_error` flag. See\n \n for more info.\n* 2.7.1\n * Packaged tests with source code\n* 2.7.0\n * Added BITOP and BITCOUNT commands. Thanks Mark Tozzi.\n * Added the TIME command. Thanks Jason Knight.\n * Added support for LUA scripting. Thanks to Angus Peart, Drew Smathers,\n Issac Kelly, Louis-Philippe Perron, Sean Bleier, Jeffrey Kaditz, and\n Dvir Volk for various patches and contributions to this feature.\n * Changed the default error handling in pipelines. By default, the first\n error in a pipeline will now be raised. A new parameter to the\n pipeline's execute, `raise_on_error`, can be set to False to keep the\n old behavior of embeedding the exception instances in the result.\n * Fixed a bug with pipelines where parse errors won't corrupt the\n socket.\n * Added the optional `number` argument to SRANDMEMBER for use with\n Redis 2.6+ servers.\n * Added PEXPIRE/PEXPIREAT/PTTL commands. Thanks Luper Rouch.\n * Added INCRBYFLOAT/HINCRBYFLOAT commands. Thanks Nikita Uvarov.\n * High precision floating point values won't lose their precision when\n being sent to the Redis server. Thanks Jason Oster and Oleg Pudeyev.\n * Added CLIENT LIST/CLIENT KILL commands\n* 2.6.2\n * `from_url` is now available as a classmethod on client classes. Thanks\n Jon Parise for the patch.\n * Fixed several encoding errors resulting from the Python 3.x support.\n* 2.6.1\n * Python 3.x support! Big thanks to Alex Grönholm.\n * Fixed a bug in the PythonParser's read_response that could hide an error\n from the client (#251).\n* 2.6.0\n * Changed (p)subscribe and (p)unsubscribe to no longer return messages\n indicating the channel was subscribed/unsubscribed to. These messages\n are available in the listen() loop instead. This is to prevent the\n following scenario:\n * Client A is subscribed to \"foo\"\n * Client B publishes message to \"foo\"\n * Client A subscribes to channel \"bar\" at the same time.\n Prior to this change, the subscribe() call would return the published\n messages on \"foo\" rather than the subscription confirmation to \"bar\".\n * Added support for GETRANGE, thanks Jean-Philippe Caruana\n * A new setting \"decode_responses\" specifies whether return values from\n Redis commands get decoded automatically using the client's charset\n value. Thanks to Frankie Dintino for the patch.\n* 2.4.13\n * redis.from_url() can take an URL representing a Redis connection string\n and return a client object. Thanks Kenneth Reitz for the patch.\n* 2.4.12\n * ConnectionPool is now fork-safe. Thanks Josiah Carson for the patch.\n* 2.4.11\n * AuthenticationError will now be correctly raised if an invalid password\n is supplied.\n * If Hiredis is unavailable, the HiredisParser will raise a RedisError\n if selected manually.\n * Made the INFO command more tolerant of Redis changes formatting. Fix\n for #217.\n* 2.4.10\n * Buffer reads from socket in the PythonParser. Fix for a Windows-specific\n bug (#205).\n * Added the OBJECT and DEBUG OBJECT commands.\n * Added __del__ methods for classes that hold on to resources that need to\n be cleaned up. This should prevent resource leakage when these objects\n leave scope due to misuse or unhandled exceptions. Thanks David Wolever\n for the suggestion.\n * Added the ECHO command for completeness.\n * Fixed a bug where attempting to subscribe to a PubSub channel of a Redis\n server that's down would blow out the stack. Fixes #179 and #195. Thanks\n Ovidiu Predescu for the test case.\n * StrictRedis's TTL command now returns a -1 when querying a key with no\n expiration. The Redis class continues to return None.\n * ZADD and SADD now return integer values indicating the number of items\n added. Thanks Homer Strong.\n * Renamed the base client class to StrictRedis, replacing ZADD and LREM in\n favor of their official argument order. The Redis class is now a subclass\n of StrictRedis, implementing the legacy redis-py implementations of ZADD\n and LREM. Docs have been updated to suggesting the use of StrictRedis.\n * SETEX in StrictRedis is now compliant with official Redis SETEX command.\n the name, value, time implementation moved to \"Redis\" for backwards\n compatibility.\n* 2.4.9\n * Removed socket retry logic in Connection. This is the responsibility of\n the caller to determine if the command is safe and can be retried. Thanks\n David Wolver.\n * Added some extra guards around various types of exceptions being raised\n when sending or parsing data. Thanks David Wolver and Denis Bilenko.\n* 2.4.8\n * Imported with_statement from __future__ for Python 2.5 compatibility.\n* 2.4.7\n * Fixed a bug where some connections were not getting released back to the\n connection pool after pipeline execution.\n * Pipelines can now be used as context managers. This is the preferred way\n of use to ensure that connections get cleaned up properly. Thanks\n David Wolever.\n * Added a convenience method called transaction() on the base Redis class.\n This method eliminates much of the boilerplate used when using pipelines\n to watch Redis keys. See the documentation for details on usage.\n* 2.4.6\n * Variadic arguments for SADD, SREM, ZREN, HDEL, LPUSH, and RPUSH. Thanks\n Raphaël Vinot.\n * (CRITICAL) Fixed an error in the Hiredis parser that occasionally caused\n the socket connection to become corrupted and unusable. This became\n noticeable once connection pools started to be used.\n * ZRANGE, ZREVRANGE, ZRANGEBYSCORE, and ZREVRANGEBYSCORE now take an\n additional optional argument, score_cast_func, which is a callable used\n to cast the score value in the return type. The default is float.\n * Removed the PUBLISH method from the PubSub class. Connections that are\n [P]SUBSCRIBEd cannot issue PUBLISH commands, so it doesn't make sense\n to have it here.\n * Pipelines now contain WATCH and UNWATCH. Calling WATCH or UNWATCH from\n the base client class will result in a deprecation warning. After\n WATCHing one or more keys, the pipeline will be placed in immediate\n execution mode until UNWATCH or MULTI are called. Refer to the new\n pipeline docs in the README for more information. Thanks to David Wolever\n and Randall Leeds for greatly helping with this.\n* 2.4.5\n * The PythonParser now works better when reading zero length strings.\n* 2.4.4\n * Fixed a typo introduced in 2.4.3\n* 2.4.3\n * Fixed a bug in the UnixDomainSocketConnection caused when trying to\n form an error message after a socket error.\n* 2.4.2\n * Fixed a bug in pipeline that caused an exception while trying to\n reconnect after a connection timeout.\n* 2.4.1\n * Fixed a bug in the PythonParser if disconnect is called before connect.\n* 2.4.0\n * WARNING: 2.4 contains several backwards incompatible changes.\n * Completely refactored Connection objects. Moved much of the Redis\n protocol packing for requests here, and eliminated the nasty dependencies\n it had on the client to do AUTH and SELECT commands on connect.\n * Connection objects now have a parser attribute. Parsers are responsible\n for reading data Redis sends. Two parsers ship with redis-py: a\n PythonParser and the HiRedis parser. redis-py will automatically use the\n HiRedis parser if you have the Python hiredis module installed, otherwise\n it will fall back to the PythonParser. You can force or the other, or even\n an external one by passing the `parser_class` argument to ConnectionPool.\n * Added a UnixDomainSocketConnection for users wanting to talk to the Redis\n instance running on a local machine only. You can use this connection\n by passing it to the `connection_class` argument of the ConnectionPool.\n * Connections no longer derive from threading.local. See threading.local\n note below.\n * ConnectionPool has been completely refactored. The ConnectionPool now\n maintains a list of connections. The redis-py client only hangs on to\n a ConnectionPool instance, calling get_connection() anytime it needs to\n send a command. When get_connection() is called, the command name and\n any keys involved in the command are passed as arguments. Subclasses of\n ConnectionPool could use this information to identify the shard the keys\n belong to and return a connection to it. ConnectionPool also implements\n disconnect() to force all connections in the pool to disconnect from\n the Redis server.\n * redis-py no longer support the SELECT command. You can still connect to\n a specific database by specifying it when instantiating a client instance\n or by creating a connection pool. If you need to talk to multiple\n databases within your application, you should use a separate client\n instance for each database you want to talk to.\n * Completely refactored Publish/Subscribe support. The subscribe and listen\n commands are no longer available on the redis-py Client class. Instead,\n the `pubsub` method returns an instance of the PubSub class which contains\n all publish/subscribe support. Note, you can still PUBLISH from the\n redis-py client class if you desire.\n * Removed support for all previously deprecated commands or options.\n * redis-py no longer uses threading.local in any way. Since the Client\n class no longer holds on to a connection, it's no longer needed. You can\n now pass client instances between threads, and commands run on those\n threads will retrieve an available connection from the pool, use it and\n release it. It should now be trivial to use redis-py with eventlet or\n greenlet.\n * ZADD now accepts pairs of value=score keyword arguments. This should help\n resolve the long standing #72. The older value and score arguments have\n been deprecated in favor of the keyword argument style.\n * Client instances now get their own copy of RESPONSE_CALLBACKS. The new\n set_response_callback method adds a user defined callback to the instance.\n * Support Jython, fixing #97. Thanks to Adam Vandenberg for the patch.\n * Using __getitem__ now properly raises a KeyError when the key is not\n found. Thanks Ionuț Arțăriși for the patch.\n * Newer Redis versions return a LOADING message for some commands while\n the database is loading from disk during server start. This could cause\n problems with SELECT. We now force a socket disconnection prior to\n raising a ResponseError so subsequent connections have to reconnect and\n re-select the appropriate database. Thanks to Benjamin Anderson for\n finding this and fixing.\n* 2.2.4\n * WARNING: Potential backwards incompatible change - Changed order of\n parameters of ZREVRANGEBYSCORE to match those of the actual Redis command.\n This is only backwards-incompatible if you were passing max and min via\n keyword args. If passing by normal args, nothing in user code should have\n to change. Thanks Stéphane Angel for the fix.\n * Fixed INFO to properly parse the Redis data correctly for both 2.2.x and\n 2.3+. Thanks Stéphane Angel for the fix.\n * Lock objects now store their timeout value as a float. This allows floats\n to be used as timeout values. No changes to existing code required.\n * WATCH now supports multiple keys. Thanks Rich Schumacher.\n * Broke out some code that was Python 2.4 incompatible. redis-py should\n now be usable on 2.4, but this hasn't actually been tested. Thanks\n Dan Colish for the patch.\n * Optimized some code using izip and islice. Should have a pretty good\n speed up on larger data sets. Thanks Dan Colish.\n * Better error handling when submitting an empty mapping to HMSET. Thanks\n Dan Colish.\n * Subscription status is now reset after every (re)connection.\n* 2.2.3\n * Added support for Hiredis. To use, simply \"pip install hiredis\" or\n \"easy_install hiredis\". Thanks for Pieter Noordhuis for the hiredis-py\n bindings and the patch to redis-py.\n * The connection class is chosen based on whether hiredis is installed\n or not. To force the use of the PythonConnection, simply create\n your own ConnectionPool instance with the connection_class argument\n assigned to to PythonConnection class.\n * Added missing command ZREVRANGEBYSCORE. Thanks Jay Baird for the patch.\n * The INFO command should be parsed correctly on 2.2.x server versions\n and is backwards compatible with older versions. Thanks Brett Hoerner.\n* 2.2.2\n * Fixed a bug in ZREVRANK where retrieving the rank of a value not in\n the zset would raise an error.\n * Fixed a bug in Connection.send where the errno import was getting\n overwritten by a local variable.\n * Fixed a bug in SLAVEOF when promoting an existing slave to a master.\n * Reverted change of download URL back to redis-VERSION.tar.gz. 2.2.1's\n change of this actually broke Pypi for Pip installs. Sorry!\n* 2.2.1\n * Changed archive name to redis-py-VERSION.tar.gz to not conflict\n with the Redis server archive.\n* 2.2.0\n * Implemented SLAVEOF\n * Implemented CONFIG as config_get and config_set\n * Implemented GETBIT/SETBIT\n * Implemented BRPOPLPUSH\n * Implemented STRLEN\n * Implemented PERSIST\n * Implemented SETRANGE\n * Changed type annotation of the `num` parameter in `zrange` from `int` to `Optional[int]" }, { "path": "CONTRIBUTING.md", "content": "# Contributing\n\n## Introduction\n\nWe appreciate your interest in considering contributing to redis-py.\nCommunity contributions mean a lot to us.\n\n## Contributions we need\n\nYou may already know how you'd like to contribute, whether it's a fix for a bug you\nencountered, or a new feature your team wants to use.\n\nIf you don't know where to start, consider improving\ndocumentation, bug triaging, and writing tutorials are all examples of\nhelpful contributions that mean less work for you.\n\n## Your First Contribution\n\nUnsure where to begin contributing? You can start by looking through\n[help-wanted\nissues](https://github.com/andymccurdy/redis-py/issues?q=is%3Aopen+is%3Aissue+label%3ahelp-wanted).\n\nNever contributed to open source before? Here are a couple of friendly\ntutorials:\n\n- \n- \n\n## Getting Started\n\nHere's how to get started with your code contribution:\n\n1. Create your own fork of redis-py\n2. Do the changes in your fork\n3. Create a virtualenv and install the development dependencies from the dev_requirements.txt file:\n ```\n python -m venv .venv\n source .venv/bin/activate\n pip install -r dev_requirements.txt\n pip install -e .[jwt]\n ```\n\n4. If you need a development environment, run `invoke devenv`. Note: this relies on docker-compose to build environments, and assumes that you have a version supporting [docker profiles](https://docs.docker.com/compose/profiles/).\n5. While developing, make sure the tests pass by running `invoke tests`\n6. If you like the change and think the project could use it, send a\n pull request\n\nTo see what else is part of the automation, run `invoke -l`\n\n## The Development Environment\n\nRunning `invoke devenv` installs the development dependencies specified\nin the dev_requirements.txt. It starts all of the dockers used by this\nproject, and leaves them running. These can be easily cleaned up with\n`invoke clean`. NOTE: it is assumed that the user running these tests,\ncan execute docker and its various commands.\n\n- A master Redis node\n- A Redis replica node\n- Three sentinel Redis nodes\n- A redis cluster\n- An stunnel docker, fronting the master Redis node\n\nThe replica node, is a replica of the master node, using the\n[leader-follower replication](https://redis.io/topics/replication)\nfeature.\n\nThe sentinels monitor the master node in a [sentinel high-availability\nconfiguration](https://redis.io/topics/sentinel).\n\n## Testing\n\nCall `invoke tests` to run all tests, or `invoke all-tests` to run linters\ntests as well. With the 'tests' and 'all-tests' targets, all Redis and\nRedisCluster tests will be run.\n\nIt is possible to run only Redis client tests (with cluster mode disabled) by\nusing `invoke standalone-tests`; similarly, RedisCluster tests can be run by using\n`invoke cluster-tests`.\n\nEach run of tests starts and stops the various dockers required. Sometimes\nthings get stuck, an `invoke clean` can help.\n\n## Linting\n\nCall `invoke linters` to run linters without also running tests. \nCall `invoke linters-fix` to run linters and automatically fix issues.\n\n## Documentation\n\nIf relevant, update the code documentation, via docstrings, or in `/docs`.\n\nYou can check how the documentation looks locally by running `invoke build-docs`\nand loading the generated HTML files in a browser.\n\nHistorically there is a mix of styles in the docstrings, but the preferred way\nof documenting code is by applying the\n[Google style](https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html).\nType hints should be added according to PEP484, and should not be repeated in\nthe docstrings.\n\n### Docker Tips\n\nFollowing are a few tips that can help you work with the Docker-based\ndevelopment environment.\n\nTo get a bash shell inside of a container:\n\n`$ docker run -it /bin/bash`\n\nContainers run a minimal Debian image that probably lacks tools you want\nto use. To install packages, first get a bash session (see previous tip)\nand then run:\n\n`$ apt update && apt install `\n\nYou can see the logging output of a containers like this:\n\n`$ docker logs -f `\n\n### Troubleshooting\n\nIf you get any errors when running `make dev` or `make test`, make sure\nthat you are using supported versions of Docker.\n\nPlease try at least versions of Docker.\n\n- Docker 19.03.12\n\n## How to Report a Bug\n\n### Security Vulnerabilities\n\n**NOTE**: If you find a security vulnerability, do NOT open an issue.\nEmail [Redis Open Source ()](mailto:oss@redis.com) instead.\n\nIn order to determine whether you are dealing with a security issue, ask\nyourself these two questions:\n\n- Can I access something that's not mine, or something I shouldn't\n have access to?\n- Can I disable something for other people?\n\nIf the answer to either of those two questions are *yes*, then you're\nprobably dealing with a security issue. Note that even if you answer\n*no* to both questions, you may still be dealing with a security\nissue, so if you're unsure, just email [us](mailto:oss@redis.com).\n\n### Everything Else\n\nWhen filing an issue, make sure to answer these five questions:\n\n1. What version of redis-py are you using?\n2. What version of redis are you using?\n3. What did you do?\n4. What did you expect to see?\n5. What did you see instead?\n\n## Suggest a feature or enhancement\n\nIf you'd like to contribute a new feature, make sure you check our\nissue list to see if someone has already proposed it. Work may already\nbe underway on the feature you want or we may have rejected a\nfeature like it already.\n\nIf you don't see anything, open a new issue that describes the feature\nyou would like and how it should work.\n\n## Code review process\n\nThe core team regularly looks at pull requests. We will provide\nfeedback as soon as possible. After receiving our feedback, please respond\nwithin two weeks. After that time, we may close your PR if it isn't\nshowing any activity.\n" }, { "path": "LICENSE", "content": "MIT License\n\nCopyright (c) 2022-2023, Redis, inc.\n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof this software and associated documentation files (the \"Software\"), to deal\nin the Software without restriction, including without limitation the rights\nto use, copy, modify, merge, publish, distribute, sublicense, and/or sell\ncopies of the Software, and to permit persons to whom the Software is\nfurnished to do so, subject to the following conditions:\n\nThe above copyright notice and this permission notice shall be included in all\ncopies or substantial portions of the Software.\n\nTHE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\nSOFTWARE.\n" }, { "path": "README.md", "content": "# redis-py\n\nThe Python interface to the Redis key-value store.\n\n[![CI](https://github.com/redis/redis-py/workflows/CI/badge.svg?branch=master)](https://github.com/redis/redis-py/actions?query=workflow%3ACI+branch%3Amaster)\n[![docs](https://readthedocs.org/projects/redis/badge/?version=stable&style=flat)](https://redis.readthedocs.io/en/stable/)\n[![MIT licensed](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/redis/redis-py/blob/master/LICENSE)\n[![pypi](https://badge.fury.io/py/redis.svg)](https://pypi.org/project/redis/)\n[![pre-release](https://img.shields.io/github/v/release/redis/redis-py?include_prereleases&label=latest-prerelease)](https://github.com/redis/redis-py/releases)\n[![codecov](https://codecov.io/gh/redis/redis-py/branch/master/graph/badge.svg?token=yenl5fzxxr)](https://codecov.io/gh/redis/redis-py)\n\n[Installation](#installation) | [Usage](#usage) | [Advanced Topics](#advanced-topics) | [Contributing](https://github.com/redis/redis-py/blob/master/CONTRIBUTING.md)\n\n---------------------------------------------\n\n**Note:** redis-py 5.0 is the last version of redis-py that supports Python 3.7, as it has reached [end of life](https://devguide.python.org/versions/). redis-py 5.1 supports Python 3.8+.
\n**Note:** redis-py 6.1.0 is the last version of redis-py that supports Python 3.8, as it has reached [end of life](https://devguide.python.org/versions/). redis-py 6.2.0 supports Python 3.9+.\n\n---------------------------------------------\n\n## How do I Redis?\n\n[Learn for free at Redis University](https://redis.io/learn/university)\n\n[Try the Redis Cloud](https://redis.io/try-free/)\n\n[Dive in developer tutorials](https://redis.io/learn)\n\n[Join the Redis community](https://redis.io/community/)\n\n[Work at Redis](https://redis.io/careers/)\n\n## Installation\n\nStart a redis via docker (for Redis versions >= 8.0):\n\n``` bash\ndocker run -p 6379:6379 -it redis:latest\n```\n\nStart a redis via docker (for Redis versions < 8.0):\n\n``` bash\ndocker run -p 6379:6379 -it redis/redis-stack:latest\n```\nTo install redis-py, simply:\n\n``` bash\n$ pip install redis\n```\n\nFor faster performance, install redis with hiredis support, this provides a compiled response parser, and *for most cases* requires zero code changes.\nBy default, if hiredis >= 1.0 is available, redis-py will attempt to use it for response parsing.\n\n``` bash\n$ pip install \"redis[hiredis]\"\n```\n\nLooking for a high-level library to handle object mapping? See [redis-om-python](https://github.com/redis/redis-om-python)!\n\n## Supported Redis Versions\n\nThe most recent version of this library supports Redis version [7.2](https://github.com/redis/redis/blob/7.2/00-RELEASENOTES), [7.4](https://github.com/redis/redis/blob/7.4/00-RELEASENOTES), [8.0](https://github.com/redis/redis/blob/8.0/00-RELEASENOTES) and [8.2](https://github.com/redis/redis/blob/8.2/00-RELEASENOTES).\n\nThe table below highlights version compatibility of the most-recent library versions and redis versions.\n\n| Library version | Supported redis versions |\n|-----------------|-------------------|\n| 3.5.3 | <= 6.2 Family of releases |\n| >= 4.5.0 | Version 5.0 to 7.0 |\n| >= 5.0.0 | Version 5.0 to 7.4 |\n| >= 6.0.0 | Version 7.2 to current |\n\n\n## Usage\n\n### Basic Example\n\n``` python\n>>> import redis\n>>> r = redis.Redis(host='localhost', port=6379, db=0)\n>>> r.set('foo', 'bar')\nTrue\n>>> r.get('foo')\nb'bar'\n```\n\nThe above code connects to localhost on port 6379, sets a value in Redis, and retrieves it. All responses are returned as bytes in Python, to receive decoded strings, set *decode_responses=True*. For this, and more connection options, see [these examples](https://redis.readthedocs.io/en/stable/examples.html).\n\n\n#### RESP3 Support\nTo enable support for RESP3, ensure you have at least version 5.0 of the client, and change your connection object to include *protocol=3*\n\n``` python\n>>> import redis\n>>> r = redis.Redis(host='localhost', port=6379, db=0, protocol=3)\n```\n\n### Connection Pools\n\nBy default, redis-py uses a connection pool to manage connections. Each instance of a Redis class receives its own connection pool. You can however define your own [redis.ConnectionPool](https://redis.readthedocs.io/en/stable/connections.html#connection-pools).\n\n``` python\n>>> pool = redis.ConnectionPool(host='localhost', port=6379, db=0)\n>>> r = redis.Redis(connection_pool=pool)\n```\n\nAlternatively, you might want to look at [Async connections](https://redis.readthedocs.io/en/stable/examples/asyncio_examples.html), or [Cluster connections](https://redis.readthedocs.io/en/stable/connections.html#cluster-client), or even [Async Cluster connections](https://redis.readthedocs.io/en/stable/connections.html#async-cluster-client).\n\n### Redis Commands\n\nThere is built-in support for all of the [out-of-the-box Redis commands](https://redis.io/commands). They are exposed using the raw Redis command names (`HSET`, `HGETALL`, etc.) except where a word (i.e. del) is reserved by the language. The complete set of commands can be found [here](https://github.com/redis/redis-py/tree/master/redis/commands), or [the documentation](https://redis.readthedocs.io/en/stable/commands.html).\n\n## Advanced Topics\n\nThe [official Redis command documentation](https://redis.io/commands)\ndoes a great job of explaining each command in detail. redis-py attempts\nto adhere to the official command syntax. There are a few exceptions:\n\n- **MULTI/EXEC**: These are implemented as part of the Pipeline class.\n The pipeline is wrapped with the MULTI and EXEC statements by\n default when it is executed, which can be disabled by specifying\n transaction=False. See more about Pipelines below.\n\n- **SUBSCRIBE/LISTEN**: Similar to pipelines, PubSub is implemented as\n a separate class as it places the underlying connection in a state\n where it can\\'t execute non-pubsub commands. Calling the pubsub\n method from the Redis client will return a PubSub instance where you\n can subscribe to channels and listen for messages. You can only call\n PUBLISH from the Redis client (see [this comment on issue\n #151](https://github.com/redis/redis-py/issues/151#issuecomment-1545015)\n for details).\n\nFor more details, please see the documentation on [advanced topics page](https://redis.readthedocs.io/en/stable/advanced_features.html).\n\n### Pipelines\n\nThe following is a basic example of a [Redis pipeline](https://redis.io/docs/manual/pipelining/), a method to optimize round-trip calls, by batching Redis commands, and receiving their results as a list.\n\n\n``` python\n>>> pipe = r.pipeline()\n>>> pipe.set('foo', 5)\n>>> pipe.set('bar', 18.5)\n>>> pipe.set('blee', \"hello world!\")\n>>> pipe.execute()\n[True, True, True]\n```\n\n### PubSub\n\nThe following example shows how to utilize [Redis Pub/Sub](https://redis.io/docs/manual/pubsub/) to subscribe to specific channels.\n\n``` python\n>>> r = redis.Redis(...)\n>>> p = r.pubsub()\n>>> p.subscribe('my-first-channel', 'my-second-channel', ...)\n>>> p.get_message()\n{'pattern': None, 'type': 'subscribe', 'channel': b'my-second-channel', 'data': 1}\n```\n\n### Redis’ search and query capabilities default dialect\n\nRelease 6.0.0 introduces a client-side default dialect for Redis’ search and query capabilities.\nBy default, the client now overrides the server-side dialect with version 2, automatically appending *DIALECT 2* to commands like *FT.AGGREGATE* and *FT.SEARCH*.\n\n**Important**: Be aware that the query dialect may impact the results returned. If needed, you can revert to a different dialect version by configuring the client accordingly.\n\n``` python\n>>> from redis.commands.search.field import TextField\n>>> from redis.commands.search.query import Query\n>>> from redis.commands.search.index_definition import IndexDefinition\n>>> import redis\n\n>>> r = redis.Redis(host='localhost', port=6379, db=0)\n>>> r.ft().create_index(\n>>> (TextField(\"name\"), TextField(\"lastname\")),\n>>> definition=IndexDefinition(prefix=[\"test:\"]),\n>>> )\n\n>>> r.hset(\"test:1\", \"name\", \"James\")\n>>> r.hset(\"test:1\", \"lastname\", \"Brown\")\n\n>>> # Query with default DIALECT 2\n>>> query = \"@name: James Brown\"\n>>> q = Query(query)\n>>> res = r.ft().search(q)\n\n>>> # Query with explicit DIALECT 1\n>>> query = \"@name: James Brown\"\n>>> q = Query(query).dialect(1)\n>>> res = r.ft().search(q)\n```\n\nYou can find further details in the [query dialect documentation](https://redis.io/docs/latest/develop/interact/search-and-query/advanced-concepts/dialects/).\n\n### Multi-database client (Active-Active)\n\nThe multi-database client allows your application to connect to multiple Redis databases, which are typically replicas of each other. It is designed to work with Redis Software and Redis Cloud Active-Active setups. The client continuously monitors database health, detects failures, and automatically fails over to the next healthy database using a configurable strategy. When the original database becomes healthy again, the client can automatically switch back to it.
\nThis is useful when:\n\n1. You have more than one Redis deployment. This might include two independent Redis servers or two or more Redis databases replicated across multiple [active-active Redis Enterprise](https://redis.io/docs/latest/operate/rs/databases/active-active/) clusters.\n2. You want your application to connect to one deployment at a time and to fail over to the next available deployment if the first deployment becomes unavailable.\n\nFor the complete failover configuration options and examples, see the [Multi-database client docs](https://redis.readthedocs.io/en/latest/multi_database.html).\n\n---------------------------------------------\n\n### Author\n\nredis-py is developed and maintained by [Redis Inc](https://redis.io). It can be found [here](\nhttps://github.com/redis/redis-py), or downloaded from [pypi](https://pypi.org/project/redis/).\n\nSpecial thanks to:\n\n- Andy McCurdy () the original author of redis-py.\n- Ludovico Magnocavallo, author of the original Python Redis client,\n from which some of the socket code is still used.\n- Alexander Solovyov for ideas on the generic response callback\n system.\n- Paul Hubbard for initial packaging support.\n\n[![Redis](https://github.com/redis/redis-py/blob/master/docs/_static/logo-redis.svg)](https://redis.io)\n" }, { "path": "benchmarks/__init__.py", "content": "" }, { "path": "benchmarks/base.py", "content": "import functools\nimport itertools\nimport sys\nimport timeit\n\nimport redis\n\n\nclass Benchmark:\n ARGUMENTS = ()\n\n def __init__(self):\n self._client = None\n\n def get_client(self, **kwargs):\n # eventually make this more robust and take optional args from\n # argparse\n if self._client is None or kwargs:\n defaults = {\"db\": 9}\n defaults.update(kwargs)\n pool = redis.ConnectionPool(**kwargs)\n self._client = redis.Redis(connection_pool=pool)\n return self._client\n\n def setup(self, **kwargs):\n pass\n\n def run(self, **kwargs):\n pass\n\n def run_benchmark(self):\n group_names = [group[\"name\"] for group in self.ARGUMENTS]\n group_values = [group[\"values\"] for group in self.ARGUMENTS]\n for value_set in itertools.product(*group_values):\n pairs = list(zip(group_names, value_set))\n arg_string = \", \".join(f\"{p[0]}={p[1]}\" for p in pairs)\n sys.stdout.write(f\"Benchmark: {arg_string}... \")\n sys.stdout.flush()\n kwargs = dict(pairs)\n setup = functools.partial(self.setup, **kwargs)\n run = functools.partial(self.run, **kwargs)\n t = timeit.timeit(stmt=run, setup=setup, number=1000)\n sys.stdout.write(f\"{t:f}\\n\")\n sys.stdout.flush()\n" }, { "path": "benchmarks/basic_operations.py", "content": "import time\nfrom argparse import ArgumentParser\nfrom functools import wraps\n\nimport redis\n\n\ndef parse_args():\n parser = ArgumentParser()\n parser.add_argument(\n \"-n\", type=int, help=\"Total number of requests (default 100000)\", default=100000\n )\n parser.add_argument(\n \"-P\",\n type=int,\n help=(\"Pipeline requests. Default 1 (no pipeline).\"),\n default=1,\n )\n parser.add_argument(\n \"-s\",\n type=int,\n help=\"Data size of SET/GET value in bytes (default 2)\",\n default=2,\n )\n\n args = parser.parse_args()\n return args\n\n\ndef run():\n args = parse_args()\n r = redis.Redis()\n r.flushall()\n set_str(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n set_int(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n get_str(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n get_int(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n incr(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n lpush(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n lrange_300(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n lpop(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n hmset(conn=r, num=args.n, pipeline_size=args.P, data_size=args.s)\n\n\ndef timer(func):\n @wraps(func)\n def wrapper(*args, **kwargs):\n start = time.monotonic()\n ret = func(*args, **kwargs)\n duration = time.monotonic() - start\n if \"num\" in kwargs:\n count = kwargs[\"num\"]\n else:\n count = args[1]\n print(f\"{func.__name__} - {count} Requests\")\n print(f\"Duration = {duration}\")\n print(f\"Rate = {count / duration}\")\n print()\n return ret\n\n return wrapper\n\n\n@timer\ndef set_str(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n set_data = \"a\".ljust(data_size, \"0\")\n for i in range(num):\n conn.set(f\"set_str:{i}\", set_data)\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef set_int(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n set_data = 10 ** (data_size - 1)\n for i in range(num):\n conn.set(f\"set_int:{i}\", set_data)\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef get_str(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n for i in range(num):\n conn.get(f\"set_str:{i}\")\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef get_int(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n for i in range(num):\n conn.get(f\"set_int:{i}\")\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef incr(conn, num, pipeline_size, *args, **kwargs):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n for i in range(num):\n conn.incr(\"incr_key\")\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef lpush(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n set_data = 10 ** (data_size - 1)\n for i in range(num):\n conn.lpush(\"lpush_key\", set_data)\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef lrange_300(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n for i in range(num):\n conn.lrange(\"lpush_key\", i, i + 300)\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef lpop(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n for i in range(num):\n conn.lpop(\"lpush_key\")\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n if pipeline_size > 1:\n conn.execute()\n\n\n@timer\ndef hmset(conn, num, pipeline_size, data_size):\n if pipeline_size > 1:\n conn = conn.pipeline()\n\n set_data = {\"str_value\": \"string\", \"int_value\": 123456, \"float_value\": 123456.0}\n for i in range(num):\n conn.hmset(\"hmset_key\", set_data)\n if pipeline_size > 1 and i % pipeline_size == 0:\n conn.execute()\n\n if pipeline_size > 1:\n conn.execute()\n\n\nif __name__ == \"__main__\":\n run()\n" }, { "path": "benchmarks/cluster_async.py", "content": "import asyncio\nimport functools\nimport time\n\nimport aioredis_cluster\nimport aredis\nimport uvloop\n\nimport redis.asyncio as redispy\n\n\ndef timer(func):\n @functools.wraps(func)\n async def wrapper(*args, **kwargs):\n tic = time.perf_counter()\n await func(*args, **kwargs)\n toc = time.perf_counter()\n return f\"{toc - tic:.4f}\"\n\n return wrapper\n\n\n@timer\nasync def set_str(client, gather, data):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(\n asyncio.create_task(client.set(f\"bench:str_{i}\", data))\n for i in range(100)\n )\n )\n else:\n for i in range(count):\n await client.set(f\"bench:str_{i}\", data)\n\n\n@timer\nasync def set_int(client, gather, data):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(\n asyncio.create_task(client.set(f\"bench:int_{i}\", data))\n for i in range(100)\n )\n )\n else:\n for i in range(count):\n await client.set(f\"bench:int_{i}\", data)\n\n\n@timer\nasync def get_str(client, gather):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(asyncio.create_task(client.get(f\"bench:str_{i}\")) for i in range(100))\n )\n else:\n for i in range(count):\n await client.get(f\"bench:str_{i}\")\n\n\n@timer\nasync def get_int(client, gather):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(asyncio.create_task(client.get(f\"bench:int_{i}\")) for i in range(100))\n )\n else:\n for i in range(count):\n await client.get(f\"bench:int_{i}\")\n\n\n@timer\nasync def hset(client, gather, data):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(\n asyncio.create_task(client.hset(\"bench:hset\", str(i), data))\n for i in range(100)\n )\n )\n else:\n for i in range(count):\n await client.hset(\"bench:hset\", str(i), data)\n\n\n@timer\nasync def hget(client, gather):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(\n asyncio.create_task(client.hget(\"bench:hset\", str(i)))\n for i in range(100)\n )\n )\n else:\n for i in range(count):\n await client.hget(\"bench:hset\", str(i))\n\n\n@timer\nasync def incr(client, gather):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(asyncio.create_task(client.incr(\"bench:incr\")) for i in range(100))\n )\n else:\n for i in range(count):\n await client.incr(\"bench:incr\")\n\n\n@timer\nasync def lpush(client, gather, data):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(\n asyncio.create_task(client.lpush(\"bench:lpush\", data))\n for i in range(100)\n )\n )\n else:\n for i in range(count):\n await client.lpush(\"bench:lpush\", data)\n\n\n@timer\nasync def lrange_300(client, gather):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(\n asyncio.create_task(client.lrange(\"bench:lpush\", i, i + 300))\n for i in range(100)\n )\n )\n else:\n for i in range(count):\n await client.lrange(\"bench:lpush\", i, i + 300)\n\n\n@timer\nasync def lpop(client, gather):\n if gather:\n for _ in range(count // 100):\n await asyncio.gather(\n *(asyncio.create_task(client.lpop(\"bench:lpush\")) for i in range(100))\n )\n else:\n for i in range(count):\n await client.lpop(\"bench:lpush\")\n\n\n@timer\nasync def warmup(client):\n await asyncio.gather(\n *(asyncio.create_task(client.exists(f\"bench:warmup_{i}\")) for i in range(100))\n )\n\n\n@timer\nasync def run(client, gather):\n data_str = \"a\" * size\n data_int = int(\"1\" * size)\n\n if gather is False:\n for ret in await asyncio.gather(\n asyncio.create_task(set_str(client, gather, data_str)),\n asyncio.create_task(set_int(client, gather, data_int)),\n asyncio.create_task(hset(client, gather, data_str)),\n asyncio.create_task(incr(client, gather)),\n asyncio.create_task(lpush(client, gather, data_int)),\n ):\n print(ret)\n for ret in await asyncio.gather(\n asyncio.create_task(get_str(client, gather)),\n asyncio.create_task(get_int(client, gather)),\n asyncio.create_task(hget(client, gather)),\n asyncio.create_task(lrange_300(client, gather)),\n asyncio.create_task(lpop(client, gather)),\n ):\n print(ret)\n else:\n print(await set_str(client, gather, data_str))\n print(await set_int(client, gather, data_int))\n print(await hset(client, gather, data_str))\n print(await incr(client, gather))\n print(await lpush(client, gather, data_int))\n\n print(await get_str(client, gather))\n print(await get_int(client, gather))\n print(await hget(client, gather))\n print(await lrange_300(client, gather))\n print(await lpop(client, gather))\n\n\nasync def main(loop, gather=None):\n arc = aredis.StrictRedisCluster(\n host=host,\n port=port,\n password=password,\n max_connections=2**31,\n max_connections_per_node=2**31,\n readonly=False,\n reinitialize_steps=count,\n skip_full_coverage_check=True,\n decode_responses=False,\n max_idle_time=count,\n idle_check_interval=count,\n )\n print(f\"{loop} {gather} {await warmup(arc)} aredis\")\n print(await run(arc, gather=gather))\n arc.connection_pool.disconnect()\n\n aiorc = await aioredis_cluster.create_redis_cluster(\n [(host, port)],\n password=password,\n state_reload_interval=count,\n idle_connection_timeout=count,\n pool_maxsize=2**31,\n )\n print(f\"{loop} {gather} {await warmup(aiorc)} aioredis-cluster\")\n print(await run(aiorc, gather=gather))\n aiorc.close()\n await aiorc.wait_closed()\n\n async with redispy.RedisCluster(\n host=host,\n port=port,\n password=password,\n reinitialize_steps=count,\n read_from_replicas=False,\n decode_responses=False,\n max_connections=2**31,\n ) as rca:\n print(f\"{loop} {gather} {await warmup(rca)} redispy\")\n print(await run(rca, gather=gather))\n\n\nif __name__ == \"__main__\":\n host = \"localhost\"\n port = 16379\n password = None\n\n count = 10000\n size = 256\n\n asyncio.run(main(\"asyncio\"))\n asyncio.run(main(\"asyncio\", gather=False))\n asyncio.run(main(\"asyncio\", gather=True))\n\n uvloop.install()\n\n asyncio.run(main(\"uvloop\"))\n asyncio.run(main(\"uvloop\", gather=False))\n asyncio.run(main(\"uvloop\", gather=True))\n" }, { "path": "benchmarks/cluster_async_pipeline.py", "content": "import asyncio\nimport functools\nimport time\n\nimport aioredis_cluster\nimport aredis\nimport uvloop\n\nimport redis.asyncio as redispy\n\n\ndef timer(func):\n @functools.wraps(func)\n async def wrapper(*args, **kwargs):\n tic = time.perf_counter()\n await func(*args, **kwargs)\n toc = time.perf_counter()\n return f\"{toc - tic:.4f}\"\n\n return wrapper\n\n\n@timer\nasync def warmup(client):\n await asyncio.gather(\n *(asyncio.create_task(client.exists(f\"bench:warmup_{i}\")) for i in range(100))\n )\n\n\n@timer\nasync def run(client):\n data_str = \"a\" * size\n data_int = int(\"1\" * size)\n\n for i in range(count):\n with client.pipeline() as pipe:\n await (\n pipe.set(f\"bench:str_{i}\", data_str)\n .set(f\"bench:int_{i}\", data_int)\n .get(f\"bench:str_{i}\")\n .get(f\"bench:int_{i}\")\n .hset(\"bench:hset\", str(i), data_str)\n .hget(\"bench:hset\", str(i))\n .incr(\"bench:incr\")\n .lpush(\"bench:lpush\", data_int)\n .lrange(\"bench:lpush\", 0, 300)\n .lpop(\"bench:lpush\")\n .execute()\n )\n\n\nasync def main(loop):\n arc = aredis.StrictRedisCluster(\n host=host,\n port=port,\n password=password,\n max_connections=2**31,\n max_connections_per_node=2**31,\n readonly=False,\n reinitialize_steps=count,\n skip_full_coverage_check=True,\n decode_responses=False,\n max_idle_time=count,\n idle_check_interval=count,\n )\n print(f\"{loop} {await warmup(arc)} aredis\")\n print(await run(arc))\n arc.connection_pool.disconnect()\n\n aiorc = await aioredis_cluster.create_redis_cluster(\n [(host, port)],\n password=password,\n state_reload_interval=count,\n idle_connection_timeout=count,\n pool_maxsize=2**31,\n )\n print(f\"{loop} {await warmup(aiorc)} aioredis-cluster\")\n print(await run(aiorc))\n aiorc.close()\n await aiorc.wait_closed()\n\n async with redispy.RedisCluster(\n host=host,\n port=port,\n password=password,\n reinitialize_steps=count,\n read_from_replicas=False,\n decode_responses=False,\n max_connections=2**31,\n ) as rca:\n print(f\"{loop} {await warmup(rca)} redispy\")\n print(await run(rca))\n\n\nif __name__ == \"__main__\":\n host = \"localhost\"\n port = 16379\n password = None\n\n count = 10000\n size = 256\n\n asyncio.run(main(\"asyncio\"))\n\n uvloop.install()\n\n asyncio.run(main(\"uvloop\"))\n" }, { "path": "benchmarks/command_packer_benchmark.py", "content": "from base import Benchmark\n\nfrom redis.connection import SYM_CRLF, SYM_DOLLAR, SYM_EMPTY, SYM_STAR, Connection\n\n\nclass StringJoiningConnection(Connection):\n def send_packed_command(self, command, check_health=True):\n \"Send an already packed command to the Redis server\"\n if not self._sock:\n self.connect()\n try:\n self._sock.sendall(command)\n except OSError as e:\n self.disconnect()\n if len(e.args) == 1:\n _errno, errmsg = \"UNKNOWN\", e.args[0]\n else:\n _errno, errmsg = e.args\n raise ConnectionError(f\"Error {_errno} while writing to socket. {errmsg}.\")\n except Exception:\n self.disconnect()\n raise\n\n def pack_command(self, *args):\n \"Pack a series of arguments into a value Redis command\"\n args_output = SYM_EMPTY.join(\n [\n SYM_EMPTY.join(\n (SYM_DOLLAR, str(len(k)).encode(), SYM_CRLF, k, SYM_CRLF)\n )\n for k in map(self.encoder.encode, args)\n ]\n )\n output = SYM_EMPTY.join(\n (SYM_STAR, str(len(args)).encode(), SYM_CRLF, args_output)\n )\n return output\n\n\nclass ListJoiningConnection(Connection):\n def send_packed_command(self, command, check_health=True):\n if not self._sock:\n self.connect()\n try:\n if isinstance(command, str):\n command = [command]\n for item in command:\n self._sock.sendall(item)\n except OSError as e:\n self.disconnect()\n if len(e.args) == 1:\n _errno, errmsg = \"UNKNOWN\", e.args[0]\n else:\n _errno, errmsg = e.args\n raise ConnectionError(f\"Error {_errno} while writing to socket. {errmsg}.\")\n except Exception:\n self.disconnect()\n raise\n\n def pack_command(self, *args):\n output = []\n buff = SYM_EMPTY.join((SYM_STAR, str(len(args)).encode(), SYM_CRLF))\n\n for k in map(self.encoder.encode, args):\n if len(buff) > 6000 or len(k) > 6000:\n buff = SYM_EMPTY.join(\n (buff, SYM_DOLLAR, str(len(k)).encode(), SYM_CRLF)\n )\n output.append(buff)\n output.append(k)\n buff = SYM_CRLF\n else:\n buff = SYM_EMPTY.join(\n (buff, SYM_DOLLAR, str(len(k)).encode(), SYM_CRLF, k, SYM_CRLF)\n )\n output.append(buff)\n return output\n\n\nclass CommandPackerBenchmark(Benchmark):\n ARGUMENTS = (\n {\n \"name\": \"connection_class\",\n \"values\": [StringJoiningConnection, ListJoiningConnection],\n },\n {\n \"name\": \"value_size\",\n \"values\": [10, 100, 1000, 10000, 100000, 1000000, 10000000, 100000000],\n },\n )\n\n def setup(self, connection_class, value_size):\n self.get_client(connection_class=connection_class)\n\n def run(self, connection_class, value_size):\n r = self.get_client()\n x = \"a\" * value_size\n r.set(\"benchmark\", x)\n\n\nif __name__ == \"__main__\":\n CommandPackerBenchmark().run_benchmark()\n" }, { "path": "benchmarks/otel_benchmark.py", "content": "#!/usr/bin/env python3\n\"\"\"\nOpenTelemetry Benchmark for redis-py.\n\nThis module provides benchmarking infrastructure to measure the performance impact\nof OpenTelemetry instrumentation on redis-py operations.\n\nThe benchmark uses a comprehensive load generator that exercises all Redis operations\n(commands, pubsub, streaming, CSC, connections) in each iteration. This ensures\nconsistent test conditions when comparing different OTel configurations.\n\nRun one scenario at a time (sync client - default):\n python -m benchmarks.otel_benchmark --scenario baseline --baseline-tag v5.2.1\n python -m benchmarks.otel_benchmark --scenario otel_disabled\n python -m benchmarks.otel_benchmark --scenario otel_noop\n python -m benchmarks.otel_benchmark --scenario otel_inmemory\n python -m benchmarks.otel_benchmark --scenario otel_enabled_http\n python -m benchmarks.otel_benchmark --scenario otel_enabled_grpc\n\nRun with async client:\n python -m benchmarks.otel_benchmark --scenario baseline --baseline-tag v5.2.1 --async\n python -m benchmarks.otel_benchmark --scenario otel_disabled --async\n python -m benchmarks.otel_benchmark --scenario otel_enabled_http --async\n\nSpecify which OTel metric groups to enable:\n python -m benchmarks.otel_benchmark --scenario otel_enabled_http --metric-groups command,pubsub\n python -m benchmarks.otel_benchmark --scenario otel_enabled_http --metric-groups all\n\"\"\"\n\nimport argparse\nimport asyncio\nimport json\nimport os\nimport statistics\nimport subprocess\nimport sys\nimport time\nfrom dataclasses import dataclass, field, asdict\nfrom pathlib import Path\nfrom tempfile import TemporaryDirectory\nfrom typing import Dict, List, Optional, Any\n\ntry:\n import psutil\n PSUTIL_AVAILABLE = True\nexcept ImportError:\n PSUTIL_AVAILABLE = False\n\n\n@dataclass\nclass BenchmarkResult:\n \"\"\"Results from a single benchmark run.\"\"\"\n scenario: str\n duration_seconds: float\n total_operations: int\n operations_per_second: float\n avg_latency_ms: float\n p50_latency_ms: float\n p95_latency_ms: float\n p99_latency_ms: float\n min_latency_ms: float\n max_latency_ms: float\n errors: int = 0\n first_error: Optional[str] = None\n # Resource usage metrics\n cpu_percent_avg: Optional[float] = None\n cpu_percent_max: Optional[float] = None\n memory_mb_avg: Optional[float] = None\n memory_mb_max: Optional[float] = None\n metadata: Dict = field(default_factory=dict)\n\n\n@dataclass\nclass LoadGeneratorConfig:\n \"\"\"Configuration for the load generator.\"\"\"\n duration_seconds: float = 30.0\n value_size_bytes: int = 100\n key_prefix: str = \"otel_bench\"\n warmup_seconds: float = 5.0\n redis_host: str = \"localhost\"\n redis_port: int = 6379\n\n\nclass ComprehensiveLoadGenerator:\n \"\"\"\n Comprehensive load generator that exercises all Redis operations.\n\n Each iteration performs:\n - Command operations (SET/GET) - triggers COMMAND metrics\n - PubSub operations (PUBLISH) - triggers PUBSUB metrics\n - Streaming operations (XADD/XREAD) - triggers STREAMING metrics\n - Connection pool operations - triggers CONNECTION metrics\n\n CSC (Client-Side Caching) requires RESP3 protocol and is tested separately\n when available.\n\n This ensures consistent test conditions across all OTel configurations.\n \"\"\"\n\n def __init__(self, config: LoadGeneratorConfig, redis_module: Any = None):\n \"\"\"\n Initialize the comprehensive load generator.\n\n Args:\n config: Load generator configuration\n redis_module: Optional redis module to use (for baseline testing with\n a different redis-py version). If None, imports redis normally.\n \"\"\"\n self.config = config\n self.redis_module = redis_module\n self.latencies: List[float] = []\n self.errors: int = 0\n self.first_error: Optional[str] = None\n self._value = \"x\" * config.value_size_bytes\n self._key_counter = 0\n self._message_counter = 0\n\n # Resources (initialized in setup)\n self.client = None\n self.pubsub_publisher = None\n self.pubsub = None\n self.stream_name = f\"{config.key_prefix}:stream\"\n self.pubsub_channel = f\"{config.key_prefix}:channel\"\n self.consumer_group = \"benchmark_group\"\n self.consumer_name = \"benchmark_consumer\"\n\n # CSC client (optional, requires RESP3)\n self.csc_client = None\n\n # Resource tracking\n self.cpu_samples: List[float] = []\n self.memory_samples: List[float] = []\n self._process = psutil.Process() if PSUTIL_AVAILABLE else None\n\n def _get_redis_module(self) -> Any:\n \"\"\"Get the redis module to use.\"\"\"\n if self.redis_module is not None:\n return self.redis_module\n import redis\n return redis\n\n def _get_key(self) -> str:\n \"\"\"Generate a key for the current operation.\"\"\"\n key = f\"{self.config.key_prefix}:{self._key_counter % 1000}\"\n self._key_counter += 1\n return key\n\n def setup(self) -> None:\n \"\"\"Set up all Redis resources.\"\"\"\n redis = self._get_redis_module()\n\n # Main client for commands\n self.client = redis.Redis(\n host=self.config.redis_host,\n port=self.config.redis_port,\n decode_responses=True\n )\n\n # PubSub publisher (separate connection)\n self.pubsub_publisher = redis.Redis(\n host=self.config.redis_host,\n port=self.config.redis_port,\n decode_responses=True\n )\n\n # PubSub subscriber\n subscriber = redis.Redis(\n host=self.config.redis_host,\n port=self.config.redis_port,\n decode_responses=True\n )\n self.pubsub = subscriber.pubsub()\n self.pubsub.subscribe(self.pubsub_channel)\n # Consume subscription confirmation\n self.pubsub.get_message(timeout=1.0)\n\n # Create stream and consumer group\n try:\n self.client.xgroup_create(\n self.stream_name,\n self.consumer_group,\n id=\"0\",\n mkstream=True\n )\n except Exception:\n # Group may already exist\n pass\n\n # Try to set up CSC client (requires RESP3, may not be available)\n try:\n from redis.cache import CacheConfig\n self.csc_client = redis.Redis(\n host=self.config.redis_host,\n port=self.config.redis_port,\n decode_responses=True,\n protocol=3,\n cache_config=CacheConfig(max_size=1000)\n )\n # Test that CSC works\n self.csc_client.ping()\n except Exception:\n # CSC not available (old redis-py version or server doesn't support RESP3)\n self.csc_client = None\n\n def teardown(self) -> None:\n \"\"\"Clean up all Redis resources.\"\"\"\n if self.pubsub:\n try:\n self.pubsub.unsubscribe()\n self.pubsub.close()\n except Exception:\n pass\n\n if self.pubsub_publisher:\n try:\n self.pubsub_publisher.close()\n except Exception:\n pass\n\n if self.client:\n try:\n self.client.delete(self.stream_name)\n except Exception:\n pass\n try:\n self.client.close()\n except Exception:\n pass\n\n if self.csc_client:\n try:\n self.csc_client.close()\n except Exception:\n pass\n\n def _run_operation(self) -> float:\n \"\"\"\n Run a comprehensive operation cycle and return latency in ms.\n\n Each cycle includes:\n 1. SET + GET (command metrics)\n 2. PUBLISH + get_message (pubsub metrics)\n 3. XADD + XREADGROUP + XACK (streaming metrics)\n 4. CSC GET operations if available (csc metrics)\n \"\"\"\n key = self._get_key()\n start = time.perf_counter()\n\n try:\n # 1. Command operations (SET/GET)\n self.client.set(key, self._value)\n self.client.get(key)\n\n # 2. PubSub operations\n self._message_counter += 1\n self.pubsub_publisher.publish(self.pubsub_channel, f\"msg:{self._message_counter}\")\n self.pubsub.get_message(timeout=0.001) # Non-blocking receive\n\n # 3. Streaming operations\n entry_id = self.client.xadd(\n self.stream_name,\n {\"data\": self._value[:50]}, # Smaller payload for streams\n maxlen=1000\n )\n messages = self.client.xreadgroup(\n self.consumer_group,\n self.consumer_name,\n {self.stream_name: \">\"},\n count=1,\n block=1 # 1ms timeout\n )\n if messages:\n for stream_name, entries in messages:\n for eid, _ in entries:\n self.client.xack(self.stream_name, self.consumer_group, eid)\n\n # 4. CSC operations (if available)\n if self.csc_client:\n csc_key = f\"{self.config.key_prefix}:csc:{self._key_counter % 100}\"\n self.csc_client.set(csc_key, self._value)\n self.csc_client.get(csc_key) # Cache miss\n self.csc_client.get(csc_key) # Cache hit\n\n except Exception as e:\n if self.first_error is None:\n self.first_error = str(e)\n self.errors += 1\n\n end = time.perf_counter()\n return (end - start) * 1000 # Convert to milliseconds\n\n def warmup(self) -> None:\n \"\"\"Run warmup operations to stabilize connections.\"\"\"\n print(f\" Warming up for {self.config.warmup_seconds}s...\")\n end_time = time.monotonic() + self.config.warmup_seconds\n while time.monotonic() < end_time:\n self._run_operation()\n self.latencies.clear()\n self.errors = 0\n self.first_error = None\n self._key_counter = 0\n self._message_counter = 0\n\n def _sample_resources(self) -> None:\n \"\"\"Sample current CPU and memory usage.\"\"\"\n if self._process is None:\n return\n try:\n # CPU percent since last call (non-blocking)\n cpu = self._process.cpu_percent(interval=None)\n if cpu > 0: # Skip first sample which is always 0\n self.cpu_samples.append(cpu)\n # Memory in MB\n mem_info = self._process.memory_info()\n self.memory_samples.append(mem_info.rss / (1024 * 1024))\n except Exception:\n pass # Ignore errors in resource sampling\n\n def run(self) -> BenchmarkResult:\n \"\"\"Run the load generator for the configured duration.\"\"\"\n self.latencies = []\n self.errors = 0\n self.first_error = None\n self.cpu_samples = []\n self.memory_samples = []\n\n # Initialize CPU percent tracking\n if self._process:\n try:\n self._process.cpu_percent(interval=None)\n except Exception:\n pass\n\n print(f\" Running load for {self.config.duration_seconds}s...\")\n start_time = time.monotonic()\n end_time = start_time + self.config.duration_seconds\n last_sample_time = start_time\n sample_interval = 0.5 # Sample resources every 500ms\n\n while time.monotonic() < end_time:\n latency = self._run_operation()\n if latency > 0:\n self.latencies.append(latency)\n\n # Sample resources periodically\n current_time = time.monotonic()\n if current_time - last_sample_time >= sample_interval:\n self._sample_resources()\n last_sample_time = current_time\n\n # Final resource sample\n self._sample_resources()\n\n actual_duration = time.monotonic() - start_time\n return self._calculate_results(actual_duration)\n\n def _calculate_results(self, duration: float) -> BenchmarkResult:\n \"\"\"Calculate benchmark results from collected latencies.\"\"\"\n if not self.latencies:\n return BenchmarkResult(\n scenario=\"unknown\", duration_seconds=duration, total_operations=0,\n operations_per_second=0, avg_latency_ms=0, p50_latency_ms=0,\n p95_latency_ms=0, p99_latency_ms=0, min_latency_ms=0,\n max_latency_ms=0, errors=self.errors, first_error=self.first_error,\n )\n\n sorted_latencies = sorted(self.latencies)\n # Count operations per cycle:\n # - 2 command ops (SET + GET)\n # - 2 pubsub ops (PUBLISH + get_message)\n # - 3 stream ops (XADD + XREADGROUP + XACK)\n # - 3 CSC ops if available (SET + 2x GET)\n ops_per_cycle = 7 + (3 if self.csc_client else 0)\n total_ops = len(self.latencies) * ops_per_cycle\n\n # Calculate resource usage stats\n cpu_avg = statistics.mean(self.cpu_samples) if self.cpu_samples else None\n cpu_max = max(self.cpu_samples) if self.cpu_samples else None\n mem_avg = statistics.mean(self.memory_samples) if self.memory_samples else None\n mem_max = max(self.memory_samples) if self.memory_samples else None\n\n return BenchmarkResult(\n scenario=\"unknown\",\n duration_seconds=duration,\n total_operations=total_ops,\n operations_per_second=total_ops / duration,\n avg_latency_ms=statistics.mean(self.latencies),\n p50_latency_ms=sorted_latencies[len(sorted_latencies) // 2],\n p95_latency_ms=sorted_latencies[int(len(sorted_latencies) * 0.95)],\n p99_latency_ms=sorted_latencies[int(len(sorted_latencies) * 0.99)],\n min_latency_ms=min(self.latencies),\n max_latency_ms=max(self.latencies),\n errors=self.errors,\n first_error=self.first_error,\n cpu_percent_avg=cpu_avg,\n cpu_percent_max=cpu_max,\n memory_mb_avg=mem_avg,\n memory_mb_max=mem_max,\n )\n\n\nclass AsyncComprehensiveLoadGenerator:\n \"\"\"\n Async comprehensive load generator that exercises all Redis operations concurrently.\n\n Each iteration performs operations concurrently where possible:\n - Command operations (SET/GET) - triggers COMMAND metrics\n - PubSub operations (PUBLISH) - triggers PUBSUB metrics\n - Streaming operations (XADD/XREAD) - triggers STREAMING metrics\n - Connection pool operations - triggers CONNECTION metrics\n\n This ensures consistent test conditions across all OTel configurations\n while maximizing throughput through concurrent execution.\n \"\"\"\n\n def __init__(self, config: LoadGeneratorConfig, redis_module: Any = None):\n \"\"\"\n Initialize the async comprehensive load generator.\n\n Args:\n config: Load generator configuration\n redis_module: Optional redis module to use (for baseline testing with\n a different redis-py version). If None, imports redis normally.\n \"\"\"\n self.config = config\n self.redis_module = redis_module\n self.latencies: List[float] = []\n self.errors: int = 0\n self.first_error: Optional[str] = None\n self._value = \"x\" * config.value_size_bytes\n self._key_counter = 0\n self._message_counter = 0\n\n # Resources (initialized in setup)\n self.client = None\n self.pubsub_publisher = None\n self.pubsub = None\n self.stream_name = f\"{config.key_prefix}:stream\"\n self.pubsub_channel = f\"{config.key_prefix}:channel\"\n self.consumer_group = \"benchmark_group\"\n self.consumer_name = \"benchmark_consumer\"\n\n # Resource tracking\n self.cpu_samples: List[float] = []\n self.memory_samples: List[float] = []\n self._process = psutil.Process() if PSUTIL_AVAILABLE else None\n\n def _get_redis_module(self) -> Any:\n \"\"\"Get the redis module to use.\"\"\"\n if self.redis_module is not None:\n return self.redis_module\n import redis\n return redis\n\n def _get_key(self) -> str:\n \"\"\"Generate a key for the current operation.\"\"\"\n key = f\"{self.config.key_prefix}:{self._key_counter % 1000}\"\n self._key_counter += 1\n return key\n\n async def setup(self) -> None:\n \"\"\"Set up all Redis resources.\"\"\"\n redis_mod = self._get_redis_module()\n Redis = redis_mod.asyncio.Redis\n\n # Main client for commands\n self.client = Redis(\n host=self.config.redis_host,\n port=self.config.redis_port,\n decode_responses=True\n )\n\n # PubSub publisher (separate connection)\n self.pubsub_publisher = Redis(\n host=self.config.redis_host,\n port=self.config.redis_port,\n decode_responses=True\n )\n\n # PubSub subscriber\n subscriber = Redis(\n host=self.config.redis_host,\n port=self.config.redis_port,\n decode_responses=True\n )\n self.pubsub = subscriber.pubsub()\n await self.pubsub.subscribe(self.pubsub_channel)\n # Consume subscription confirmation\n await self.pubsub.get_message(timeout=1.0)\n\n # Create stream and consumer group\n try:\n await self.client.xgroup_create(\n self.stream_name,\n self.consumer_group,\n id=\"0\",\n mkstream=True\n )\n except Exception:\n # Group may already exist\n pass\n\n async def teardown(self) -> None:\n \"\"\"Clean up all Redis resources.\"\"\"\n if self.pubsub:\n try:\n await self.pubsub.unsubscribe()\n await self.pubsub.aclose()\n except Exception:\n pass\n\n if self.pubsub_publisher:\n try:\n await self.pubsub_publisher.aclose()\n except Exception:\n pass\n\n if self.client:\n try:\n await self.client.delete(self.stream_name)\n except Exception:\n pass\n try:\n await self.client.aclose()\n except Exception:\n pass\n\n async def _run_operation(self) -> float:\n \"\"\"\n Run a comprehensive operation cycle concurrently and return latency in ms.\n\n Each cycle includes concurrent execution of:\n 1. SET + GET (command metrics)\n 2. PUBLISH + get_message (pubsub metrics)\n 3. XADD + XREADGROUP + XACK (streaming metrics)\n\n Note: CSC (Client-Side Caching) is not included in async benchmark\n as CSC metrics export is not supported for async client.\n \"\"\"\n key = self._get_key()\n start = time.perf_counter()\n\n try:\n self._message_counter += 1\n\n # Group 1: Independent write operations (run concurrently)\n write_tasks = [\n self.client.set(key, self._value),\n self.pubsub_publisher.publish(self.pubsub_channel, f\"msg:{self._message_counter}\"),\n self.client.xadd(\n self.stream_name,\n {\"data\": self._value[:50]},\n maxlen=1000\n ),\n ]\n\n await asyncio.gather(*write_tasks)\n\n # Group 2: Independent read operations (run concurrently)\n read_tasks = [\n self.client.get(key),\n self.pubsub.get_message(timeout=0.001),\n ]\n\n await asyncio.gather(*read_tasks)\n\n # Group 3: Stream read (depends on xadd completing first)\n messages = await self.client.xreadgroup(\n self.consumer_group,\n self.consumer_name,\n {self.stream_name: \">\"},\n count=1,\n block=1\n )\n\n # Acknowledge messages concurrently\n if messages:\n ack_tasks = [\n self.client.xack(self.stream_name, self.consumer_group, eid)\n for stream_name, entries in messages\n for eid, _ in entries\n ]\n if ack_tasks:\n await asyncio.gather(*ack_tasks)\n\n except Exception as e:\n if self.first_error is None:\n self.first_error = str(e)\n self.errors += 1\n\n end = time.perf_counter()\n return (end - start) * 1000 # Convert to milliseconds\n\n async def warmup(self) -> None:\n \"\"\"Run warmup operations to stabilize connections.\"\"\"\n print(f\" Warming up for {self.config.warmup_seconds}s...\")\n end_time = time.monotonic() + self.config.warmup_seconds\n while time.monotonic() < end_time:\n await self._run_operation()\n self.latencies.clear()\n self.errors = 0\n self.first_error = None\n self._key_counter = 0\n self._message_counter = 0\n\n def _sample_resources(self) -> None:\n \"\"\"Sample current CPU and memory usage.\"\"\"\n if self._process is None:\n return\n try:\n cpu = self._process.cpu_percent(interval=None)\n if cpu > 0:\n self.cpu_samples.append(cpu)\n mem_info = self._process.memory_info()\n self.memory_samples.append(mem_info.rss / (1024 * 1024))\n except Exception:\n pass\n\n async def run(self) -> BenchmarkResult:\n \"\"\"Run the load generator for the configured duration.\"\"\"\n self.latencies = []\n self.errors = 0\n self.first_error = None\n self.cpu_samples = []\n self.memory_samples = []\n\n # Initialize CPU percent tracking\n if self._process:\n try:\n self._process.cpu_percent(interval=None)\n except Exception:\n pass\n\n print(f\" Running load for {self.config.duration_seconds}s...\")\n start_time = time.monotonic()\n end_time = start_time + self.config.duration_seconds\n last_sample_time = start_time\n sample_interval = 0.5\n\n while time.monotonic() < end_time:\n latency = await self._run_operation()\n if latency > 0:\n self.latencies.append(latency)\n\n # Sample resources periodically\n current_time = time.monotonic()\n if current_time - last_sample_time >= sample_interval:\n self._sample_resources()\n last_sample_time = current_time\n\n # Final resource sample\n self._sample_resources()\n\n actual_duration = time.monotonic() - start_time\n return self._calculate_results(actual_duration)\n\n def _calculate_results(self, duration: float) -> BenchmarkResult:\n \"\"\"Calculate benchmark results from collected latencies.\"\"\"\n if not self.latencies:\n return BenchmarkResult(\n scenario=\"unknown\", duration_seconds=duration, total_operations=0,\n operations_per_second=0, avg_latency_ms=0, p50_latency_ms=0,\n p95_latency_ms=0, p99_latency_ms=0, min_latency_ms=0,\n max_latency_ms=0, errors=self.errors, first_error=self.first_error,\n )\n\n sorted_latencies = sorted(self.latencies)\n # Count operations per cycle:\n # - 2 command ops (SET + GET)\n # - 2 pubsub ops (PUBLISH + get_message)\n # - 3 stream ops (XADD + XREADGROUP + XACK)\n # Note: CSC is not included in async benchmark (not supported)\n ops_per_cycle = 7\n total_ops = len(self.latencies) * ops_per_cycle\n\n # Calculate resource usage stats\n cpu_avg = statistics.mean(self.cpu_samples) if self.cpu_samples else None\n cpu_max = max(self.cpu_samples) if self.cpu_samples else None\n mem_avg = statistics.mean(self.memory_samples) if self.memory_samples else None\n mem_max = max(self.memory_samples) if self.memory_samples else None\n\n return BenchmarkResult(\n scenario=\"unknown\",\n duration_seconds=duration,\n total_operations=total_ops,\n operations_per_second=total_ops / duration,\n avg_latency_ms=statistics.mean(self.latencies),\n p50_latency_ms=sorted_latencies[len(sorted_latencies) // 2],\n p95_latency_ms=sorted_latencies[int(len(sorted_latencies) * 0.95)],\n p99_latency_ms=sorted_latencies[int(len(sorted_latencies) * 0.99)],\n min_latency_ms=min(self.latencies),\n max_latency_ms=max(self.latencies),\n errors=self.errors,\n first_error=self.first_error,\n cpu_percent_avg=cpu_avg,\n cpu_percent_max=cpu_max,\n memory_mb_avg=mem_avg,\n memory_mb_max=mem_max,\n )\n\n\ndef print_result(result: BenchmarkResult, iterations: int = 1) -> None:\n \"\"\"Print a single benchmark result.\"\"\"\n print(\"\\n\" + \"=\" * 60)\n print(f\"BENCHMARK RESULT: {result.scenario}\")\n print(\"=\" * 60)\n if iterations > 1:\n print(f\" Iterations: {iterations} (averaged)\")\n print(f\" Duration: {result.duration_seconds:.2f}s\")\n print(f\" Operations: {result.total_operations:,}\")\n print(f\" Ops/sec: {result.operations_per_second:,.0f}\")\n print(f\" Avg latency: {result.avg_latency_ms:.3f}ms\")\n print(f\" P50 latency: {result.p50_latency_ms:.3f}ms\")\n print(f\" P95 latency: {result.p95_latency_ms:.3f}ms\")\n print(f\" P99 latency: {result.p99_latency_ms:.3f}ms\")\n print(f\" Min latency: {result.min_latency_ms:.3f}ms\")\n print(f\" Max latency: {result.max_latency_ms:.3f}ms\")\n print(f\" Errors: {result.errors}\")\n if result.first_error:\n print(f\" First error: {result.first_error}\")\n # Resource usage\n if result.cpu_percent_avg is not None:\n print(f\" CPU avg: {result.cpu_percent_avg:.1f}%\")\n if result.cpu_percent_max is not None:\n print(f\" CPU max: {result.cpu_percent_max:.1f}%\")\n if result.memory_mb_avg is not None:\n print(f\" Memory avg: {result.memory_mb_avg:.1f} MB\")\n if result.memory_mb_max is not None:\n print(f\" Memory max: {result.memory_mb_max:.1f} MB\")\n if result.metadata.get(\"description\"):\n print(f\" Description: {result.metadata['description']}\")\n print(\"=\" * 60)\n\n\ndef average_results(results: List[BenchmarkResult]) -> BenchmarkResult:\n \"\"\"Average multiple benchmark results into a single result.\"\"\"\n if not results:\n raise ValueError(\"Cannot average empty results list\")\n if len(results) == 1:\n return results[0]\n\n n = len(results)\n # Find first error from any iteration\n first_error = None\n for r in results:\n if r.first_error:\n first_error = r.first_error\n break\n\n # Average CPU and memory (only from results that have them)\n cpu_avgs = [r.cpu_percent_avg for r in results if r.cpu_percent_avg is not None]\n cpu_maxs = [r.cpu_percent_max for r in results if r.cpu_percent_max is not None]\n mem_avgs = [r.memory_mb_avg for r in results if r.memory_mb_avg is not None]\n mem_maxs = [r.memory_mb_max for r in results if r.memory_mb_max is not None]\n\n return BenchmarkResult(\n scenario=results[0].scenario,\n duration_seconds=sum(r.duration_seconds for r in results) / n,\n total_operations=int(sum(r.total_operations for r in results) / n),\n operations_per_second=sum(r.operations_per_second for r in results) / n,\n avg_latency_ms=sum(r.avg_latency_ms for r in results) / n,\n p50_latency_ms=sum(r.p50_latency_ms for r in results) / n,\n p95_latency_ms=sum(r.p95_latency_ms for r in results) / n,\n p99_latency_ms=sum(r.p99_latency_ms for r in results) / n,\n min_latency_ms=min(r.min_latency_ms for r in results),\n max_latency_ms=max(r.max_latency_ms for r in results),\n errors=sum(r.errors for r in results),\n first_error=first_error,\n cpu_percent_avg=sum(cpu_avgs) / len(cpu_avgs) if cpu_avgs else None,\n cpu_percent_max=max(cpu_maxs) if cpu_maxs else None,\n memory_mb_avg=sum(mem_avgs) / len(mem_avgs) if mem_avgs else None,\n memory_mb_max=max(mem_maxs) if mem_maxs else None,\n metadata=results[0].metadata,\n )\n\n\ndef parse_args() -> argparse.Namespace:\n \"\"\"Parse command line arguments.\"\"\"\n parser = argparse.ArgumentParser(\n description=\"Benchmark OTel instrumentation overhead in redis-py. Run one scenario at a time.\"\n )\n parser.add_argument(\n \"--scenario\", type=str, required=True,\n choices=[\"baseline\", \"otel_disabled\", \"otel_noop\", \"otel_inmemory\", \"otel_enabled_http\", \"otel_enabled_grpc\"],\n help=\"Scenario to run (required)\"\n )\n parser.add_argument(\n \"--baseline-tag\", type=str, default=None,\n help=\"Git tag to use for baseline scenario (required when --scenario baseline)\"\n )\n parser.add_argument(\n \"--duration\", type=float, default=30.0,\n help=\"Duration of benchmark in seconds (default: 30)\"\n )\n parser.add_argument(\n \"--warmup\", type=float, default=5.0,\n help=\"Warmup duration in seconds (default: 5)\"\n )\n parser.add_argument(\n \"--value-size\", type=int, default=100,\n help=\"Size of values in bytes (default: 100)\"\n )\n parser.add_argument(\n \"--host\", type=str, default=\"localhost\",\n help=\"Redis host (default: localhost)\"\n )\n parser.add_argument(\n \"--port\", type=int, default=6379,\n help=\"Redis port (default: 6379)\"\n )\n parser.add_argument(\n \"--json\", action=\"store_true\",\n help=\"Output result as JSON\"\n )\n parser.add_argument(\n \"--iterations\", type=int, default=5,\n help=\"Number of iterations to run (default: 5). Final result is averaged.\"\n )\n parser.add_argument(\n \"--metric-groups\", type=str, default=None,\n help=(\n \"Comma-separated list of metric groups to enable. \"\n \"Options: command, pubsub, streaming, csc, connection_basic, connection_advanced, resiliency, all. \"\n \"Default: resiliency,connection_basic. \"\n \"Example: --metric-groups command,pubsub,connection_basic\"\n )\n )\n parser.add_argument(\n \"--export-interval\", type=int, default=10000,\n help=\"OTel metric export interval in milliseconds (default: 10000)\"\n )\n parser.add_argument(\n \"--async\", dest=\"use_async\", action=\"store_true\",\n help=\"Use async Redis client instead of sync client\"\n )\n return parser.parse_args()\n\n\ndef _clear_redis_modules() -> None:\n \"\"\"Remove all redis.* modules from sys.modules to allow fresh import.\"\"\"\n to_remove = [key for key in sys.modules if key == \"redis\" or key.startswith(\"redis.\")]\n for key in to_remove:\n del sys.modules[key]\n\n\ndef run_baseline_scenario(tag: str, config: LoadGeneratorConfig) -> Optional[BenchmarkResult]:\n \"\"\"\n Run benchmark against a baseline git tag using ComprehensiveLoadGenerator.\n\n This clones the repo at the specified tag, manipulates sys.path to import\n the old redis-py version, and runs the benchmark using the same comprehensive\n load generator for consistent comparison with other scenarios.\n\n Returns:\n BenchmarkResult or None on failure\n \"\"\"\n repo_root = Path(__file__).parent.parent\n\n with TemporaryDirectory() as tmpdir:\n print(f\" Cloning repository at tag {tag}...\")\n try:\n subprocess.run(\n [\"git\", \"clone\", \"--depth\", \"1\", \"--branch\", tag, str(repo_root), tmpdir],\n check=True, capture_output=True, text=True\n )\n except subprocess.CalledProcessError as e:\n print(f\" ERROR: Failed to clone tag {tag}: {e.stderr}\")\n return None\n\n # Save original sys.path and modules state\n original_path = sys.path.copy()\n\n try:\n # Clear any existing redis imports and prepend cloned directory\n _clear_redis_modules()\n sys.path.insert(0, tmpdir)\n\n # Import redis from the cloned directory\n import redis as baseline_redis\n\n print(f\" Using redis from: {baseline_redis.__file__}\")\n\n # Create comprehensive generator with baseline redis module\n generator = ComprehensiveLoadGenerator(config, redis_module=baseline_redis)\n generator.setup()\n try:\n generator.warmup()\n result = generator.run()\n result.scenario = \"baseline\"\n result.metadata[\"tag\"] = tag\n result.metadata[\"description\"] = \"Baseline without OTel code\"\n return result\n finally:\n generator.teardown()\n\n finally:\n # Restore original sys.path and clear redis modules again\n sys.path[:] = original_path\n _clear_redis_modules()\n\n\nasync def run_baseline_scenario_async(tag: str, config: LoadGeneratorConfig) -> Optional[BenchmarkResult]:\n \"\"\"\n Run async benchmark against a baseline git tag using AsyncComprehensiveLoadGenerator.\n\n This clones the repo at the specified tag, manipulates sys.path to import\n the old redis-py version, and runs the benchmark using the async comprehensive\n load generator for consistent comparison with other scenarios.\n\n Returns:\n BenchmarkResult or None on failure\n \"\"\"\n repo_root = Path(__file__).parent.parent\n\n with TemporaryDirectory() as tmpdir:\n print(f\" Cloning repository at tag {tag}...\")\n try:\n subprocess.run(\n [\"git\", \"clone\", \"--depth\", \"1\", \"--branch\", tag, str(repo_root), tmpdir],\n check=True, capture_output=True, text=True\n )\n except subprocess.CalledProcessError as e:\n print(f\" ERROR: Failed to clone tag {tag}: {e.stderr}\")\n return None\n\n # Save original sys.path and modules state\n original_path = sys.path.copy()\n\n try:\n # Clear any existing redis imports and prepend cloned directory\n _clear_redis_modules()\n sys.path.insert(0, tmpdir)\n\n # Import redis from the cloned directory\n import redis as baseline_redis\n\n print(f\" Using redis from: {baseline_redis.__file__}\")\n\n # Create async comprehensive generator with baseline redis module\n generator = AsyncComprehensiveLoadGenerator(config, redis_module=baseline_redis)\n await generator.setup()\n try:\n await generator.warmup()\n result = await generator.run()\n result.scenario = \"baseline\"\n result.metadata[\"tag\"] = tag\n result.metadata[\"description\"] = \"Baseline without OTel code (async)\"\n result.metadata[\"client_type\"] = \"async\"\n return result\n finally:\n await generator.teardown()\n\n finally:\n # Restore original sys.path and clear redis modules again\n sys.path[:] = original_path\n _clear_redis_modules()\n\n\ndef _get_metric_groups_for_benchmark(metric_group_names: Optional[List[str]]) -> Optional[List[Any]]:\n \"\"\"\n Convert metric group names to MetricGroup enum values.\n\n Args:\n metric_group_names: List of metric group names (command, pubsub, streaming, csc,\n connection_basic, connection_advanced, resiliency, all)\n\n Returns:\n List of MetricGroup enum values, or None for defaults\n \"\"\"\n if not metric_group_names:\n return None\n\n from redis.observability.config import MetricGroup\n\n name_to_group = {\n \"command\": MetricGroup.COMMAND,\n \"pubsub\": MetricGroup.PUBSUB,\n \"streaming\": MetricGroup.STREAMING,\n \"csc\": MetricGroup.CSC,\n \"connection_basic\": MetricGroup.CONNECTION_BASIC,\n \"connection_advanced\": MetricGroup.CONNECTION_ADVANCED,\n \"resiliency\": MetricGroup.RESILIENCY,\n }\n\n # Handle \"all\" - return all metric groups\n if \"all\" in metric_group_names:\n return list(name_to_group.values())\n\n groups = []\n for name in metric_group_names:\n if name in name_to_group:\n groups.append(name_to_group[name])\n\n return groups if groups else None\n\n\ndef setup_scenario(\n scenario: str,\n metric_group_names: Optional[List[str]] = None,\n export_interval_millis: int = 10000\n) -> str:\n \"\"\"\n Set up OTel for a scenario. Returns the description.\n This should only be called once per process.\n\n Args:\n scenario: The scenario name\n metric_group_names: List of metric group names to enable\n export_interval_millis: Export interval in milliseconds for PeriodicExportingMetricReader\n \"\"\"\n if scenario == \"otel_disabled\":\n return \"OTel not initialized\"\n\n # Determine which metric groups to enable\n metric_groups = _get_metric_groups_for_benchmark(metric_group_names)\n groups_desc = \"\"\n if metric_group_names:\n groups_desc = f\" [groups: {', '.join(metric_group_names)}]\"\n\n if scenario == \"otel_noop\":\n from opentelemetry import metrics\n from opentelemetry.metrics import NoOpMeterProvider\n metrics.set_meter_provider(NoOpMeterProvider())\n\n from redis.observability.providers import get_observability_instance\n from redis.observability.config import OTelConfig\n otel = get_observability_instance()\n if metric_groups:\n otel.init(OTelConfig(metric_groups=metric_groups))\n else:\n otel.init(OTelConfig())\n return f\"OTel with NoOpMeterProvider{groups_desc}\"\n\n elif scenario == \"otel_inmemory\":\n from opentelemetry import metrics\n from opentelemetry.sdk.metrics import MeterProvider\n from opentelemetry.sdk.metrics.export import InMemoryMetricReader\n\n reader = InMemoryMetricReader()\n provider = MeterProvider(metric_readers=[reader])\n metrics.set_meter_provider(provider)\n\n from redis.observability.providers import get_observability_instance\n from redis.observability.config import OTelConfig\n otel = get_observability_instance()\n if metric_groups:\n otel.init(OTelConfig(metric_groups=metric_groups))\n else:\n otel.init(OTelConfig())\n return f\"OTel with InMemoryMetricReader{groups_desc}\"\n\n elif scenario == \"otel_enabled_http\":\n from opentelemetry import metrics\n from opentelemetry.sdk.metrics import MeterProvider\n from opentelemetry.sdk.metrics.export import PeriodicExportingMetricReader\n from opentelemetry.exporter.otlp.proto.http.metric_exporter import OTLPMetricExporter\n\n # HTTP exporter - host configurable via OTEL_COLLECTOR_HOST env var,\n # default is localhost (port 4318)\n host = os.environ.get(\"OTEL_COLLECTOR_HOST\", \"localhost\")\n endpoint = f\"http://{host}:4318/v1/metrics\"\n exporter = OTLPMetricExporter(endpoint=endpoint)\n reader = PeriodicExportingMetricReader(exporter, export_interval_millis=export_interval_millis)\n provider = MeterProvider(metric_readers=[reader])\n metrics.set_meter_provider(provider)\n\n from redis.observability.providers import get_observability_instance\n from redis.observability.config import OTelConfig\n otel = get_observability_instance()\n if metric_groups:\n otel.init(OTelConfig(metric_groups=metric_groups))\n else:\n otel.init(OTelConfig())\n return f\"OTel with PeriodicExportingMetricReader (HTTP) -> {endpoint} [export: {export_interval_millis}ms]{groups_desc}\"\n\n elif scenario == \"otel_enabled_grpc\":\n from opentelemetry import metrics\n from opentelemetry.sdk.metrics import MeterProvider\n from opentelemetry.sdk.metrics.export import PeriodicExportingMetricReader\n from opentelemetry.exporter.otlp.proto.grpc.metric_exporter import OTLPMetricExporter\n\n # gRPC exporter - host configurable via OTEL_COLLECTOR_HOST env var,\n # default is localhost (port 4317)\n host = os.environ.get(\"OTEL_COLLECTOR_HOST\", \"localhost\")\n endpoint = f\"{host}:4317\"\n exporter = OTLPMetricExporter(endpoint=endpoint, insecure=True)\n reader = PeriodicExportingMetricReader(exporter, export_interval_millis=export_interval_millis)\n provider = MeterProvider(metric_readers=[reader])\n metrics.set_meter_provider(provider)\n\n from redis.observability.providers import get_observability_instance\n from redis.observability.config import OTelConfig\n otel = get_observability_instance()\n if metric_groups:\n otel.init(OTelConfig(metric_groups=metric_groups))\n else:\n otel.init(OTelConfig())\n return f\"OTel with PeriodicExportingMetricReader (gRPC) -> {endpoint} [export: {export_interval_millis}ms]{groups_desc}\"\n\n else:\n raise ValueError(f\"Unknown scenario: {scenario}\")\n\n\ndef run_iteration(scenario: str, config: LoadGeneratorConfig, description: str) -> BenchmarkResult:\n \"\"\"\n Run a single benchmark iteration using the ComprehensiveLoadGenerator.\n\n Args:\n scenario: The scenario name\n config: Load generator configuration\n description: Description of the scenario\n\n Returns:\n BenchmarkResult from the comprehensive load generator\n \"\"\"\n generator = ComprehensiveLoadGenerator(config)\n generator.setup()\n try:\n generator.warmup()\n result = generator.run()\n result.scenario = scenario\n result.metadata[\"description\"] = description\n return result\n finally:\n generator.teardown()\n\n\nasync def run_iteration_async(scenario: str, config: LoadGeneratorConfig, description: str) -> BenchmarkResult:\n \"\"\"\n Run a single async benchmark iteration using the AsyncComprehensiveLoadGenerator.\n\n Args:\n scenario: The scenario name\n config: Load generator configuration\n description: Description of the scenario\n\n Returns:\n BenchmarkResult from the async comprehensive load generator\n \"\"\"\n generator = AsyncComprehensiveLoadGenerator(config)\n await generator.setup()\n try:\n await generator.warmup()\n result = await generator.run()\n result.scenario = scenario\n result.metadata[\"description\"] = description\n result.metadata[\"client_type\"] = \"async\"\n return result\n finally:\n await generator.teardown()\n\n\ndef main() -> int:\n \"\"\"Main entry point for the benchmark.\"\"\"\n args = parse_args()\n\n # Validate baseline scenario requires --baseline-tag\n if args.scenario == \"baseline\" and not args.baseline_tag:\n print(\"ERROR: --baseline-tag is required when --scenario baseline\")\n return 1\n\n # Parse metric groups\n metric_group_names: Optional[List[str]] = None\n if args.metric_groups:\n metric_group_names = [g.strip().lower() for g in args.metric_groups.split(\",\")]\n # Validate metric group names\n valid_groups = {\"command\", \"pubsub\", \"streaming\", \"csc\", \"connection_basic\", \"connection_advanced\", \"resiliency\", \"all\"}\n invalid_groups = [g for g in metric_group_names if g not in valid_groups]\n if invalid_groups:\n print(f\"ERROR: Invalid metric groups: {', '.join(invalid_groups)}\")\n print(f\"Valid options: {', '.join(sorted(valid_groups))}\")\n return 1\n\n client_type = \"async\" if args.use_async else \"sync\"\n print(\"=\" * 60)\n print(f\"OTel Benchmark: {args.scenario} ({client_type} client)\")\n if args.baseline_tag:\n print(f\"Baseline tag: {args.baseline_tag}\")\n if metric_group_names:\n print(f\"Metric groups: {', '.join(metric_group_names)}\")\n print(\"=\" * 60)\n print(f\"Client type: {client_type}\")\n print(f\"Duration: {args.duration}s per iteration\")\n print(f\"Warmup: {args.warmup}s per iteration\")\n print(f\"Iterations: {args.iterations}\")\n print(f\"Value size: {args.value_size} bytes\")\n print(f\"Redis: {args.host}:{args.port}\")\n\n config = LoadGeneratorConfig(\n duration_seconds=args.duration,\n warmup_seconds=args.warmup,\n value_size_bytes=args.value_size,\n redis_host=args.host,\n redis_port=args.port,\n )\n\n # Set up OTel once (for non-baseline scenarios)\n description = \"\"\n if args.scenario != \"baseline\":\n print(\"\\nSetting up OTel...\")\n description = setup_scenario(\n args.scenario,\n metric_group_names=metric_group_names,\n export_interval_millis=args.export_interval\n )\n print(f\" {description}\")\n\n # Run benchmark iterations\n if args.use_async:\n return asyncio.run(_run_async_benchmark(args, config, description))\n else:\n return _run_sync_benchmark(args, config, description)\n\n\ndef _run_sync_benchmark(args: argparse.Namespace, config: LoadGeneratorConfig, description: str) -> int:\n \"\"\"Run sync benchmark iterations.\"\"\"\n results: List[BenchmarkResult] = []\n for i in range(args.iterations):\n print(f\"\\n--- Iteration {i + 1}/{args.iterations} ---\")\n\n if args.scenario == \"baseline\":\n result = run_baseline_scenario(tag=args.baseline_tag, config=config)\n if result is None:\n print(\"ERROR: Baseline benchmark failed\")\n return 1\n else:\n result = run_iteration(args.scenario, config, description)\n\n results.append(result)\n print(f\" Ops/sec: {result.operations_per_second:,.0f}\")\n\n # Average results across all iterations\n final_result = average_results(results)\n\n # Output results\n if args.json:\n print(json.dumps(asdict(final_result), indent=2))\n else:\n print_result(final_result, iterations=args.iterations)\n\n return 0\n\n\nasync def _run_async_benchmark(args: argparse.Namespace, config: LoadGeneratorConfig, description: str) -> int:\n \"\"\"Run async benchmark iterations.\"\"\"\n results: List[BenchmarkResult] = []\n for i in range(args.iterations):\n print(f\"\\n--- Iteration {i + 1}/{args.iterations} ---\")\n\n if args.scenario == \"baseline\":\n result = await run_baseline_scenario_async(tag=args.baseline_tag, config=config)\n if result is None:\n print(\"ERROR: Baseline benchmark failed\")\n return 1\n else:\n result = await run_iteration_async(args.scenario, config, description)\n\n results.append(result)\n print(f\" Ops/sec: {result.operations_per_second:,.0f}\")\n\n # Average results across all iterations\n final_result = average_results(results)\n\n # Output results\n if args.json:\n print(json.dumps(asdict(final_result), indent=2))\n else:\n print_result(final_result, iterations=args.iterations)\n\n return 0\n\n\nif __name__ == \"__main__\":\n sys.exit(main())\n" }, { "path": "benchmarks/socket_read_size.py", "content": "from base import Benchmark\n\nfrom redis.connection import PythonParser, _HiredisParser\n\n\nclass SocketReadBenchmark(Benchmark):\n ARGUMENTS = (\n {\"name\": \"parser\", \"values\": [PythonParser, _HiredisParser]},\n {\n \"name\": \"value_size\",\n \"values\": [10, 100, 1000, 10000, 100000, 1000000, 10000000, 100000000],\n },\n {\"name\": \"read_size\", \"values\": [4096, 8192, 16384, 32768, 65536, 131072]},\n )\n\n def setup(self, value_size, read_size, parser):\n r = self.get_client(parser_class=parser, socket_read_size=read_size)\n r.set(\"benchmark\", \"a\" * value_size)\n\n def run(self, value_size, read_size, parser):\n r = self.get_client()\n r.get(\"benchmark\")\n\n\nif __name__ == \"__main__\":\n SocketReadBenchmark().run_benchmark()\n" }, { "path": "codecov.yml", "content": "ignore:\n - \"benchmarks/**\"\n - \"tasks.py\"\n\ncodecov:\n require_ci_to_pass: yes\n\ncoverage:\n precision: 2\n round: down\n range: \"80...100\"\n status:\n patch: off # off for now as it yells about everything\n" }, { "path": "dev_requirements.txt", "content": "build\nbuild==1.2.2.post1 ; platform_python_implementation == \"PyPy\"\nclick==8.0.4\ninvoke==2.2.0\npackaging>=20.4\npackaging==24.2 ; platform_python_implementation == \"PyPy\"\n\npytest\npytest==8.3.4 ; platform_python_implementation == \"PyPy\"\npytest-asyncio>=0.24.0\npytest-asyncio==1.1.0 ; platform_python_implementation == \"PyPy\"\npytest-cov\ncoverage<7.11.1\npytest-cov==6.0.0 ; platform_python_implementation == \"PyPy\"\ncoverage==7.6.12 ; platform_python_implementation == \"PyPy\"\npytest-profiling==1.8.1\npytest-timeout\npytest-timeout==2.3.1 ; platform_python_implementation == \"PyPy\"\n\nruff==0.9.6\nujson>=4.2.0\nuvloop<=0.21.0; platform_python_implementation == \"CPython\" and python_version < \"3.14\"\nuvloop>=0.22; platform_python_implementation == \"CPython\" and python_version >= \"3.14\"\nvulture>=2.3.0\n\nnumpy>=1.24.0 ; platform_python_implementation == \"CPython\"\nnumpy>=1.24.0,<2.0 ; platform_python_implementation == \"PyPy\"\n\nredis-entraid==1.0.0\npybreaker>=1.4.0\n\nxxhash==3.6.0\n\nopentelemetry-api>=1.30.0\nopentelemetry-sdk>=1.30.0\nopentelemetry-exporter-otlp>=1.30.0\n" }, { "path": "docker-compose.yml", "content": "---\n# image tag 8.0-RC2-pre is the one matching the 8.0 GA release\nx-client-libs-stack-image: &client-libs-stack-image\n image: \"redislabs/client-libs-test:${CLIENT_LIBS_TEST_STACK_IMAGE_TAG:-8.4.0}\"\n\nx-client-libs-image: &client-libs-image\n image: \"redislabs/client-libs-test:${CLIENT_LIBS_TEST_IMAGE_TAG:-8.4.0}\"\n\nnetworks:\n redis-net:\n driver: bridge\nservices:\n\n redis:\n <<: *client-libs-image\n container_name: redis-standalone\n environment:\n - TLS_ENABLED=yes\n - TLS_CLIENT_CNS=test_user\n - REDIS_CLUSTER=no\n - PORT=6379\n - TLS_PORT=6666\n command: ${REDIS_EXTRA_ARGS:---enable-debug-command yes --enable-module-command yes --tls-auth-clients optional --save \"\"}\n ports:\n - 6379:6379\n - 6666:6666 # TLS port\n volumes:\n - \"./dockers/standalone:/redis/work\"\n profiles:\n - standalone\n - sentinel\n - replica\n - all-stack\n - all\n\n replica:\n <<: *client-libs-image\n container_name: redis-replica\n depends_on:\n - redis\n environment:\n - TLS_ENABLED=no\n - REDIS_CLUSTER=no\n - PORT=6380\n command: ${REDIS_EXTRA_ARGS:---enable-debug-command yes --replicaof redis 6379 --protected-mode no --save \"\"}\n ports:\n - 6380:6380\n volumes:\n - \"./dockers/replica:/redis/work\"\n profiles:\n - replica\n - all-stack\n - all\n\n cluster:\n <<: *client-libs-image\n container_name: redis-cluster\n environment:\n - REDIS_CLUSTER=yes\n - NODES=6\n - REPLICAS=1\n - TLS_ENABLED=yes\n - TLS_CLIENT_CNS=test_user\n - PORT=16379\n - TLS_PORT=27379\n command: ${REDIS_EXTRA_ARGS:---enable-debug-command yes --enable-module-command yes --tls-auth-clients optional --save \"\" --tls-cluster yes}\n ports:\n - \"16379-16384:16379-16384\"\n - \"27379-27384:27379-27384\"\n volumes:\n - \"./dockers/cluster:/redis/work\"\n profiles:\n - cluster\n - all-stack\n - all\n\n sentinel:\n <<: *client-libs-image\n container_name: redis-sentinel\n depends_on:\n - redis\n environment:\n - REDIS_CLUSTER=no\n - NODES=3\n - PORT=26379\n command: ${REDIS_EXTRA_ARGS:---sentinel}\n ports:\n - 26379:26379\n - 26380:26380\n - 26381:26381\n volumes:\n - \"./dockers/sentinel.conf:/redis/config-default/redis.conf\"\n - \"./dockers/sentinel:/redis/work\"\n profiles:\n - sentinel\n - all-stack\n - all\n\n redis-stack:\n <<: *client-libs-stack-image\n container_name: redis-stack\n environment:\n - REDIS_CLUSTER=no\n - PORT=6379\n command: ${REDIS_EXTRA_ARGS:---enable-debug-command yes --enable-module-command yes --save \"\"}\n ports:\n - 6479:6379\n volumes:\n - \"./dockers/redis-stack:/redis/work\"\n profiles:\n - standalone\n - all-stack\n - all\n\n redis-proxied:\n <<: *client-libs-image\n container_name: redis-proxied\n ports:\n - \"3000:3000\"\n networks:\n - redis-net\n healthcheck:\n test: [\"CMD\", \"redis-cli\", \"-p\", \"3000\", \"PING\"]\n interval: 10s\n timeout: 3s\n retries: 5\n\n resp-proxy:\n image: redislabs/client-resp-proxy\n container_name: resp-proxy\n environment:\n LISTEN_HOST: \"0.0.0.0\"\n LISTEN_PORT: \"15379,15380,15381\"\n TARGET_HOST: \"redis-proxied\"\n TARGET_PORT: \"3000\"\n API_PORT: \"4000\"\n ENABLE_LOGGING: true\n SIMULATE_CLUSTER: true\n DEFAULT_INTERCEPTORS: \"cluster,hitless,logger\"\n\n ports:\n - \"15379:15379\"\n - \"15380:15380\"\n - \"15381:15381\"\n - \"4000:4000\"\n depends_on:\n - redis-proxied\n networks:\n - redis-net\n healthcheck:\n test: [\"CMD\", \"sh\", \"-c\", \"wget -qO- http://localhost:4000/stats || exit 1\"]\n interval: 10s\n timeout: 3s\n retries: 5\n" }, { "path": "dockers/sentinel.conf", "content": "sentinel resolve-hostnames yes\nsentinel monitor redis-py-test redis 6379 2\n# Be much more tolerant to transient stalls (index builds, GC, I/O)\nsentinel down-after-milliseconds redis-py-test 60000\n# Avoid rapid repeated failover attempts\nsentinel failover-timeout redis-py-test 180000\n# Keep it conservative: sync one replica at a time\nsentinel parallel-syncs redis-py-test 1" }, { "path": "docs/Makefile", "content": "# Makefile for Sphinx documentation\n#\n\n# You can set these variables from the command line.\nSPHINXOPTS =\nSPHINXBUILD = sphinx-build\nPAPER =\nBUILDDIR = _build\n\n# Internal variables.\nPAPEROPT_a4 = -D latex_paper_size=a4\nPAPEROPT_letter = -D latex_paper_size=letter\nALLSPHINXOPTS = -d $(BUILDDIR)/doctrees $(PAPEROPT_$(PAPER)) $(SPHINXOPTS) .\n# the i18n builder cannot share the environment and doctrees with the others\nI18NSPHINXOPTS = $(PAPEROPT_$(PAPER)) $(SPHINXOPTS) .\n\n.PHONY: help clean html dirhtml singlehtml pickle json htmlhelp qthelp devhelp epub latex latexpdf text man changes linkcheck doctest gettext\n\nhelp:\n\t@echo \"Please use \\`make ' where is one of\"\n\t@echo \" html to make standalone HTML files\"\n\t@echo \" dirhtml to make HTML files named index.html in directories\"\n\t@echo \" singlehtml to make a single large HTML file\"\n\t@echo \" pickle to make pickle files\"\n\t@echo \" json to make JSON files\"\n\t@echo \" htmlhelp to make HTML files and a HTML help project\"\n\t@echo \" qthelp to make HTML files and a qthelp project\"\n\t@echo \" devhelp to make HTML files and a Devhelp project\"\n\t@echo \" epub to make an epub\"\n\t@echo \" latex to make LaTeX files, you can set PAPER=a4 or PAPER=letter\"\n\t@echo \" latexpdf to make LaTeX files and run them through pdflatex\"\n\t@echo \" text to make text files\"\n\t@echo \" man to make manual pages\"\n\t@echo \" texinfo to make Texinfo files\"\n\t@echo \" info to make Texinfo files and run them through makeinfo\"\n\t@echo \" gettext to make PO message catalogs\"\n\t@echo \" changes to make an overview of all changed/added/deprecated items\"\n\t@echo \" linkcheck to check all external links for integrity\"\n\t@echo \" doctest to run all doctests embedded in the documentation (if enabled)\"\n\nclean:\n\t-rm -rf $(BUILDDIR)/*\n\nhtml:\n\t$(SPHINXBUILD) -b html $(ALLSPHINXOPTS) $(BUILDDIR)/html\n\t@echo\n\t@echo \"Build finished. The HTML pages are in $(BUILDDIR)/html.\"\n\ndirhtml:\n\t$(SPHINXBUILD) -b dirhtml $(ALLSPHINXOPTS) $(BUILDDIR)/dirhtml\n\t@echo\n\t@echo \"Build finished. The HTML pages are in $(BUILDDIR)/dirhtml.\"\n\nsinglehtml:\n\t$(SPHINXBUILD) -b singlehtml $(ALLSPHINXOPTS) $(BUILDDIR)/singlehtml\n\t@echo\n\t@echo \"Build finished. The HTML page is in $(BUILDDIR)/singlehtml.\"\n\npickle:\n\t$(SPHINXBUILD) -b pickle $(ALLSPHINXOPTS) $(BUILDDIR)/pickle\n\t@echo\n\t@echo \"Build finished; now you can process the pickle files.\"\n\njson:\n\t$(SPHINXBUILD) -b json $(ALLSPHINXOPTS) $(BUILDDIR)/json\n\t@echo\n\t@echo \"Build finished; now you can process the JSON files.\"\n\nhtmlhelp:\n\t$(SPHINXBUILD) -b htmlhelp $(ALLSPHINXOPTS) $(BUILDDIR)/htmlhelp\n\t@echo\n\t@echo \"Build finished; now you can run HTML Help Workshop with the\" \\\n\t \".hhp project file in $(BUILDDIR)/htmlhelp.\"\n\nqthelp:\n\t$(SPHINXBUILD) -b qthelp $(ALLSPHINXOPTS) $(BUILDDIR)/qthelp\n\t@echo\n\t@echo \"Build finished; now you can run \"qcollectiongenerator\" with the\" \\\n\t \".qhcp project file in $(BUILDDIR)/qthelp, like this:\"\n\t@echo \"# qcollectiongenerator $(BUILDDIR)/qthelp/redis-py.qhcp\"\n\t@echo \"To view the help file:\"\n\t@echo \"# assistant -collectionFile $(BUILDDIR)/qthelp/redis-py.qhc\"\n\ndevhelp:\n\t$(SPHINXBUILD) -b devhelp $(ALLSPHINXOPTS) $(BUILDDIR)/devhelp\n\t@echo\n\t@echo \"Build finished.\"\n\t@echo \"To view the help file:\"\n\t@echo \"# mkdir -p $$HOME/.local/share/devhelp/redis-py\"\n\t@echo \"# ln -s $(BUILDDIR)/devhelp $$HOME/.local/share/devhelp/redis-py\"\n\t@echo \"# devhelp\"\n\nepub:\n\t$(SPHINXBUILD) -b epub $(ALLSPHINXOPTS) $(BUILDDIR)/epub\n\t@echo\n\t@echo \"Build finished. The epub file is in $(BUILDDIR)/epub.\"\n\nlatex:\n\t$(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex\n\t@echo\n\t@echo \"Build finished; the LaTeX files are in $(BUILDDIR)/latex.\"\n\t@echo \"Run \\`make' in that directory to run these through (pdf)latex\" \\\n\t \"(use \\`make latexpdf' here to do that automatically).\"\n\nlatexpdf:\n\t$(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex\n\t@echo \"Running LaTeX files through pdflatex...\"\n\t$(MAKE) -C $(BUILDDIR)/latex all-pdf\n\t@echo \"pdflatex finished; the PDF files are in $(BUILDDIR)/latex.\"\n\ntext:\n\t$(SPHINXBUILD) -b text $(ALLSPHINXOPTS) $(BUILDDIR)/text\n\t@echo\n\t@echo \"Build finished. The text files are in $(BUILDDIR)/text.\"\n\nman:\n\t$(SPHINXBUILD) -b man $(ALLSPHINXOPTS) $(BUILDDIR)/man\n\t@echo\n\t@echo \"Build finished. The manual pages are in $(BUILDDIR)/man.\"\n\ntexinfo:\n\t$(SPHINXBUILD) -b texinfo $(ALLSPHINXOPTS) $(BUILDDIR)/texinfo\n\t@echo\n\t@echo \"Build finished. The Texinfo files are in $(BUILDDIR)/texinfo.\"\n\t@echo \"Run \\`make' in that directory to run these through makeinfo\" \\\n\t \"(use \\`make info' here to do that automatically).\"\n\ninfo:\n\t$(SPHINXBUILD) -b texinfo $(ALLSPHINXOPTS) $(BUILDDIR)/texinfo\n\t@echo \"Running Texinfo files through makeinfo...\"\n\tmake -C $(BUILDDIR)/texinfo info\n\t@echo \"makeinfo finished; the Info files are in $(BUILDDIR)/texinfo.\"\n\ngettext:\n\t$(SPHINXBUILD) -b gettext $(I18NSPHINXOPTS) $(BUILDDIR)/locale\n\t@echo\n\t@echo \"Build finished. The message catalogs are in $(BUILDDIR)/locale.\"\n\nchanges:\n\t$(SPHINXBUILD) -b changes $(ALLSPHINXOPTS) $(BUILDDIR)/changes\n\t@echo\n\t@echo \"The overview file is in $(BUILDDIR)/changes.\"\n\nlinkcheck:\n\t$(SPHINXBUILD) -b linkcheck $(ALLSPHINXOPTS) $(BUILDDIR)/linkcheck\n\t@echo\n\t@echo \"Link check complete; look for any errors in the above output \" \\\n\t \"or in $(BUILDDIR)/linkcheck/output.txt.\"\n\ndoctest:\n\t$(SPHINXBUILD) -b doctest $(ALLSPHINXOPTS) $(BUILDDIR)/doctest\n\t@echo \"Testing of doctests in the sources finished, look at the \" \\\n\t \"results in $(BUILDDIR)/doctest/output.txt.\"\n" }, { "path": "docs/_static/.keep", "content": "" }, { "path": "docs/_templates/.keep", "content": "" }, { "path": "docs/advanced_features.rst", "content": "Advanced Features\n=================\n\nA note about threading\n----------------------\n\nRedis client instances can safely be shared between threads. Internally,\nconnection instances are only retrieved from the connection pool during\ncommand execution, and returned to the pool directly after. Command\nexecution never modifies state on the client instance.\n\nHowever, there is one caveat: the Redis SELECT command. The SELECT\ncommand allows you to switch the database currently in use by the\nconnection. That database remains selected until another is selected or\nuntil the connection is closed. This creates an issue in that\nconnections could be returned to the pool that are connected to a\ndifferent database.\n\nAs a result, redis-py does not implement the SELECT command on client\ninstances. If you use multiple Redis databases within the same\napplication, you should create a separate client instance (and possibly\na separate connection pool) for each database.\n\nIt is not safe to pass PubSub or Pipeline objects between threads.\n\nPipelines\n---------\n\nDefault pipelines\n~~~~~~~~~~~~~~~~~\n\nPipelines are a subclass of the base Redis class that provide support\nfor buffering multiple commands to the server in a single request. They\ncan be used to dramatically increase the performance of groups of\ncommands by reducing the number of back-and-forth TCP packets between\nthe client and server.\n\nPipelines are quite simple to use:\n\n.. code:: python\n\n >>> r = redis.Redis(...)\n >>> r.set('bing', 'baz')\n >>> # Use the pipeline() method to create a pipeline instance\n >>> pipe = r.pipeline()\n >>> # The following SET commands are buffered\n >>> pipe.set('foo', 'bar')\n >>> pipe.get('bing')\n >>> # the EXECUTE call sends all buffered commands to the server, returning\n >>> # a list of responses, one for each command.\n >>> pipe.execute()\n [True, b'baz']\n\nFor ease of use, all commands being buffered into the pipeline return\nthe pipeline object itself. Therefore calls can be chained like:\n\n.. code:: python\n\n >>> pipe.set('foo', 'bar').sadd('faz', 'baz').incr('auto_number').execute()\n [True, True, 6]\n\nIn addition, pipelines can also ensure the buffered commands are\nexecuted atomically as a group. This happens by default. If you want to\ndisable the atomic nature of a pipeline but still want to buffer\ncommands, you can turn off transactions.\n\n.. code:: python\n\n >>> pipe = r.pipeline(transaction=False)\n\nA common issue occurs when requiring atomic transactions but needing to\nretrieve values in Redis prior for use within the transaction. For\ninstance, let's assume that the INCR command didn't exist and we need to\nbuild an atomic version of INCR in Python.\n\nThe completely naive implementation could GET the value, increment it in\nPython, and SET the new value back. However, this is not atomic because\nmultiple clients could be doing this at the same time, each getting the\nsame value from GET.\n\nEnter the WATCH command. WATCH provides the ability to monitor one or\nmore keys prior to starting a transaction. If any of those keys change\nprior the execution of that transaction, the entire transaction will be\ncanceled and a WatchError will be raised. To implement our own\nclient-side INCR command, we could do something like this:\n\n.. code:: python\n\n >>> with r.pipeline() as pipe:\n ... while True:\n ... try:\n ... # put a WATCH on the key that holds our sequence value\n ... pipe.watch('OUR-SEQUENCE-KEY')\n ... # after WATCHing, the pipeline is put into immediate execution\n ... # mode until we tell it to start buffering commands again.\n ... # this allows us to get the current value of our sequence\n ... current_value = pipe.get('OUR-SEQUENCE-KEY')\n ... next_value = int(current_value) + 1\n ... # now we can put the pipeline back into buffered mode with MULTI\n ... pipe.multi()\n ... pipe.set('OUR-SEQUENCE-KEY', next_value)\n ... # and finally, execute the pipeline (the set command)\n ... pipe.execute()\n ... # if a WatchError wasn't raised during execution, everything\n ... # we just did happened atomically.\n ... break\n ... except WatchError:\n ... # another client must have changed 'OUR-SEQUENCE-KEY' between\n ... # the time we started WATCHing it and the pipeline's execution.\n ... # our best bet is to just retry.\n ... continue\n\nNote that, because the Pipeline must bind to a single connection for the\nduration of a WATCH, care must be taken to ensure that the connection is\nreturned to the connection pool by calling the reset() method. If the\nPipeline is used as a context manager (as in the example above) reset()\nwill be called automatically. Of course you can do this the manual way\nby explicitly calling reset():\n\n.. code:: python\n\n >>> pipe = r.pipeline()\n >>> while True:\n ... try:\n ... pipe.watch('OUR-SEQUENCE-KEY')\n ... ...\n ... pipe.execute()\n ... break\n ... except WatchError:\n ... continue\n ... finally:\n ... pipe.reset()\n\nA convenience method named \"transaction\" exists for handling all the\nboilerplate of handling and retrying watch errors. It takes a callable\nthat should expect a single parameter, a pipeline object, and any number\nof keys to be WATCHed. Our client-side INCR command above can be written\nlike this, which is much easier to read:\n\n.. code:: python\n\n >>> def client_side_incr(pipe):\n ... current_value = pipe.get('OUR-SEQUENCE-KEY')\n ... next_value = int(current_value) + 1\n ... pipe.multi()\n ... pipe.set('OUR-SEQUENCE-KEY', next_value)\n >>>\n >>> r.transaction(client_side_incr, 'OUR-SEQUENCE-KEY')\n [True]\n\nBe sure to call pipe.multi() in the callable passed to Redis.transaction\nprior to any write commands.\n\nPipelines in clusters\n~~~~~~~~~~~~~~~~~~~~~\n\nClusterPipeline is a subclass of RedisCluster that provides support for\nRedis pipelines in cluster mode. When calling the execute() command, all\nthe commands are grouped by the node on which they will be executed, and\nare then executed by the respective nodes in parallel. The pipeline\ninstance will wait for all the nodes to respond before returning the\nresult to the caller. Command responses are returned as a list sorted in\nthe same order in which they were sent. Pipelines can be used to\ndramatically increase the throughput of Redis Cluster by significantly\nreducing the number of network round trips between the client and\nthe server.\n\n.. code:: python\n\n >>> rc = RedisCluster()\n >>> with rc.pipeline() as pipe:\n ... pipe.set('foo', 'value1')\n ... pipe.set('bar', 'value2')\n ... pipe.get('foo')\n ... pipe.get('bar')\n ... print(pipe.execute())\n [True, True, b'value1', b'value2']\n ... pipe.set('foo1', 'bar1').get('foo1').execute()\n [True, b'bar1']\n\nPlease note:\n\n- RedisCluster pipelines currently only support key-based commands.\n- The pipeline gets its ‘load_balancing_strategy’ value from the\n cluster’s parameter. Thus, if read from replications is enabled in\n the cluster instance, the pipeline will also direct read commands to\n replicas.\n\n\nTransactions in clusters\n~~~~~~~~~~~~~~~~~~~~~~~~\n\nTransactions are supported in cluster-mode with one caveat: all keys of\nall commands issued on a transaction pipeline must reside on the\nsame slot. This is similar to the limitation of multikey commands in\ncluster. The reason behind this is that the Redis engine does not offer\na mechanism to block or exchange key data across nodes on the fly. A\nclient may add some logic to abstract engine limitations when running\non a cluster, such as the pipeline behavior explained on the previous\nblock, but there is no simple way that a client can enforce atomicity\nacross nodes on a distributed system.\n\nThe compromise of limiting the transaction pipeline to same-slot keys\nis exactly that: a compromise. While this behavior is different from\nnon-transactional cluster pipelines, it simplifies migration of clients\nfrom standalone to cluster under some circumstances. Note that application\ncode that issues multi/exec commands on a standalone client without\nembedding them within a pipeline would eventually get ‘AttributeError’.\nWith this approach, if the application uses ‘client.pipeline(transaction=True)’,\nthen switching the client with a cluster-aware instance would simplify\ncode changes (to some extent). This may be true for application code that\nmakes use of hash keys, since its transactions may already be\nmapping all commands to the same slot.\n\nAn alternative is some kind of two-step commit solution, where a slot\nvalidation is run before the actual commands are run. This could work\nwith controlled node maintenance but does not cover single node failures.\n\nGiven the cluster limitations for transactions, by default pipeline isn't in\ntransactional mode. To enable transactional context set:\n\n.. code:: python\n\n >>> p = rc.pipeline(transaction=True)\n\nAfter entering the transactional context you can add commands to a transactional\ncontext, by one of the following ways:\n\n.. code:: python\n\n >>> p = rc.pipeline(transaction=True) # Chaining commands\n >>> p.set(\"key\", \"value\")\n >>> p.get(\"key\")\n >>> response = p.execute()\n\nOr\n\n.. code:: python\n\n >>> with rc.pipeline(transaction=True) as pipe: # Using context manager\n ... pipe.set(\"key\", \"value\")\n ... pipe.get(\"key\")\n ... response = pipe.execute()\n\nAs you see there's no need to explicitly send `MULTI/EXEC` commands to control context start/end\n`ClusterPipeline` will take care of it.\n\nTo ensure that different keys will be mapped to a same hash slot on the server side\nprepend your keys with the same hash tag, the technique that allows you to control\nkeys distribution.\nMore information `here `_\n\n.. code:: python\n\n >>> with rc.pipeline(transaction=True) as pipe:\n ... pipe.set(\"{tag}foo\", \"bar\")\n ... pipe.set(\"{tag}bar\", \"foo\")\n ... pipe.get(\"{tag}foo\")\n ... pipe.get(\"{tag}bar\")\n ... response = pipe.execute()\n\nCAS Transactions\n~~~~~~~~~~~~~~~~~~~~~~~~\n\nIf you want to apply optimistic locking for certain keys, you have to execute\n`WATCH` command in transactional context. `WATCH` command follows the same limitations\nas any other multi key command - all keys should be mapped to the same hash slot.\n\nHowever, the difference between CAS transaction and normal one is that you have to\nexplicitly call MULTI command to indicate the start of transactional context, WATCH\ncommand itself and any subsequent commands before MULTI will be immediately executed\non the server side so you can apply optimistic locking and get necessary data before\ntransaction execution.\n\n.. code:: python\n\n >>> with rc.pipeline(transaction=True) as pipe:\n ... pipe.watch(\"mykey\") # Apply locking by immediately executing command\n ... val = pipe.get(\"mykey\") # Immediately retrieves value\n ... val = val + 1 # Increment value\n ... pipe.multi() # Starting transaction context\n ... pipe.set(\"mykey\", val) # Command will be pipelined\n ... response = pipe.execute() # Returns OK or None if key was modified in the meantime\n\n\nPublish / Subscribe\n-------------------\n\nredis-py includes a PubSub object that subscribes to channels and\nlistens for new messages. Creating a PubSub object is easy.\n\n.. code:: python\n\n >>> r = redis.Redis(...)\n >>> p = r.pubsub()\n\nOnce a PubSub instance is created, channels and patterns can be\nsubscribed to.\n\n.. code:: python\n\n >>> p.subscribe('my-first-channel', 'my-second-channel', ...)\n >>> p.psubscribe('my-*', ...)\n\nThe PubSub instance is now subscribed to those channels/patterns. The\nsubscription confirmations can be seen by reading messages from the\nPubSub instance.\n\n.. code:: python\n\n >>> p.get_message()\n {'pattern': None, 'type': 'subscribe', 'channel': b'my-second-channel', 'data': 1}\n >>> p.get_message()\n {'pattern': None, 'type': 'subscribe', 'channel': b'my-first-channel', 'data': 2}\n >>> p.get_message()\n {'pattern': None, 'type': 'psubscribe', 'channel': b'my-*', 'data': 3}\n\nEvery message read from a PubSub instance will be a dictionary with the\nfollowing keys.\n\n- **type**: One of the following: 'subscribe', 'unsubscribe',\n 'psubscribe', 'punsubscribe', 'message', 'pmessage'\n- **channel**: The channel [un]subscribed to or the channel a message\n was published to\n- **pattern**: The pattern that matched a published message's channel.\n Will be None in all cases except for 'pmessage' types.\n- **data**: The message data. With [un]subscribe messages, this value\n will be the number of channels and patterns the connection is\n currently subscribed to. With [p]message messages, this value will be\n the actual published message.\n\nLet's send a message now.\n\n.. code:: python\n\n # the publish method returns the number matching channel and pattern\n # subscriptions. 'my-first-channel' matches both the 'my-first-channel'\n # subscription and the 'my-*' pattern subscription, so this message will\n # be delivered to 2 channels/patterns\n >>> r.publish('my-first-channel', 'some data')\n 2\n >>> p.get_message()\n {'channel': b'my-first-channel', 'data': b'some data', 'pattern': None, 'type': 'message'}\n >>> p.get_message()\n {'channel': b'my-first-channel', 'data': b'some data', 'pattern': b'my-*', 'type': 'pmessage'}\n\nUnsubscribing works just like subscribing. If no arguments are passed to\n[p]unsubscribe, all channels or patterns will be unsubscribed from.\n\n.. code:: python\n\n >>> p.unsubscribe()\n >>> p.punsubscribe('my-*')\n >>> p.get_message()\n {'channel': b'my-second-channel', 'data': 2, 'pattern': None, 'type': 'unsubscribe'}\n >>> p.get_message()\n {'channel': b'my-first-channel', 'data': 1, 'pattern': None, 'type': 'unsubscribe'}\n >>> p.get_message()\n {'channel': b'my-*', 'data': 0, 'pattern': None, 'type': 'punsubscribe'}\n\nredis-py also allows you to register callback functions to handle\npublished messages. Message handlers take a single argument, the\nmessage, which is a dictionary just like the examples above. To\nsubscribe to a channel or pattern with a message handler, pass the\nchannel or pattern name as a keyword argument with its value being the\ncallback function.\n\nWhen a message is read on a channel or pattern with a message handler,\nthe message dictionary is created and passed to the message handler. In\nthis case, a None value is returned from get_message() since the message\nwas already handled.\n\n.. code:: python\n\n >>> def my_handler(message):\n ... print('MY HANDLER: ', message['data'])\n >>> p.subscribe(**{'my-channel': my_handler})\n # read the subscribe confirmation message\n >>> p.get_message()\n {'pattern': None, 'type': 'subscribe', 'channel': b'my-channel', 'data': 1}\n >>> r.publish('my-channel', 'awesome data')\n 1\n # for the message handler to work, we need tell the instance to read data.\n # this can be done in several ways (read more below). we'll just use\n # the familiar get_message() function for now\n >>> message = p.get_message()\n MY HANDLER: awesome data\n # note here that the my_handler callback printed the string above.\n # `message` is None because the message was handled by our handler.\n >>> print(message)\n None\n\nIf your application is not interested in the (sometimes noisy)\nsubscribe/unsubscribe confirmation messages, you can ignore them by\npassing ignore_subscribe_messages=True to r.pubsub(). This will cause\nall subscribe/unsubscribe messages to be read, but they won't bubble up\nto your application.\n\n.. code:: python\n\n >>> p = r.pubsub(ignore_subscribe_messages=True)\n >>> p.subscribe('my-channel')\n >>> p.get_message() # hides the subscribe message and returns None\n >>> r.publish('my-channel', 'my data')\n 1\n >>> p.get_message()\n {'channel': b'my-channel', 'data': b'my data', 'pattern': None, 'type': 'message'}\n\nThere are three different strategies for reading messages.\n\nThe examples above have been using pubsub.get_message(). Behind the\nscenes, get_message() uses the system's 'select' module to quickly poll\nthe connection's socket. If there's data available to be read,\nget_message() will read it, format the message and return it or pass it\nto a message handler. If there's no data to be read, get_message() will\nimmediately return None. This makes it trivial to integrate into an\nexisting event loop inside your application.\n\n.. code:: python\n\n >>> while True:\n >>> message = p.get_message()\n >>> if message:\n >>> # do something with the message\n >>> time.sleep(0.001) # be nice to the system :)\n\nOlder versions of redis-py only read messages with pubsub.listen().\nlisten() is a generator that blocks until a message is available. If\nyour application doesn't need to do anything else but receive and act on\nmessages received from redis, listen() is an easy way to get up an\nrunning.\n\n.. code:: python\n\n >>> for message in p.listen():\n ... # do something with the message\n\nThe third option runs an event loop in a separate thread.\npubsub.run_in_thread() creates a new thread and starts the event loop.\nThe thread object is returned to the caller of run_in_thread(). The\ncaller can use the thread.stop() method to shut down the event loop and\nthread. Behind the scenes, this is simply a wrapper around get_message()\nthat runs in a separate thread, essentially creating a tiny non-blocking\nevent loop for you. run_in_thread() takes an optional sleep_time\nargument. If specified, the event loop will call time.sleep() with the\nvalue in each iteration of the loop.\n\nNote: Since we're running in a separate thread, there's no way to handle\nmessages that aren't automatically handled with registered message\nhandlers. Therefore, redis-py prevents you from calling run_in_thread()\nif you're subscribed to patterns or channels that don't have message\nhandlers attached.\n\n.. code:: python\n\n >>> p.subscribe(**{'my-channel': my_handler})\n >>> thread = p.run_in_thread(sleep_time=0.001)\n # the event loop is now running in the background processing messages\n # when it's time to shut it down...\n >>> thread.stop()\n\nrun_in_thread also supports an optional exception handler, which lets\nyou catch exceptions that occur within the worker thread and handle them\nappropriately. The exception handler will take as arguments the\nexception itself, the pubsub object, and the worker thread returned by\nrun_in_thread.\n\n.. code:: python\n\n >>> p.subscribe(**{'my-channel': my_handler})\n >>> def exception_handler(ex, pubsub, thread):\n >>> print(ex)\n >>> thread.stop()\n >>> thread = p.run_in_thread(exception_handler=exception_handler)\n\nA PubSub object adheres to the same encoding semantics as the client\ninstance it was created from. Any channel or pattern that's unicode will\nbe encoded using the encoding specified on the client before being sent\nto Redis. If the client's decode_responses flag is set the False (the\ndefault), the 'channel', 'pattern' and 'data' values in message\ndictionaries will be byte strings (str on Python 2, bytes on Python 3).\nIf the client's decode_responses is True, then the 'channel', 'pattern'\nand 'data' values will be automatically decoded to unicode strings using\nthe client's encoding.\n\nPubSub objects remember what channels and patterns they are subscribed\nto. In the event of a disconnection such as a network error or timeout,\nthe PubSub object will re-subscribe to all prior channels and patterns\nwhen reconnecting. Messages that were published while the client was\ndisconnected cannot be delivered. When you're finished with a PubSub\nobject, call its .close() method to shutdown the connection.\n\n.. code:: python\n\n >>> p = r.pubsub()\n >>> ...\n >>> p.close()\n\nThe PUBSUB set of subcommands CHANNELS, NUMSUB and NUMPAT are also\nsupported:\n\n.. code:: python\n\n >>> r.pubsub_channels()\n [b'foo', b'bar']\n >>> r.pubsub_numsub('foo', 'bar')\n [(b'foo', 9001), (b'bar', 42)]\n >>> r.pubsub_numsub('baz')\n [(b'baz', 0)]\n >>> r.pubsub_numpat()\n 1204\n\nSharded pubsub\n~~~~~~~~~~~~~~\n\n`Sharded pubsub `_ is a feature introduced with Redis 7.0, and fully supported by redis-py as of 5.0. It helps scale the usage of pub/sub in cluster mode, by having the cluster shard messages to nodes that own a slot for a shard channel. Here, the cluster ensures the published shard messages are forwarded to the appropriate nodes. Clients subscribe to a channel by connecting to either the master responsible for the slot, or any of its replicas.\n\nThis makes use of the `SSUBSCRIBE `_ and `SPUBLISH `_ commands within Redis.\n\nThe following, is a simplified example:\n\n.. code:: python\n\n >>> from redis.cluster import RedisCluster, ClusterNode\n >>> r = RedisCluster(startup_nodes=[ClusterNode('localhost', 6379), ClusterNode('localhost', 6380)])\n >>> p = r.pubsub()\n >>> p.ssubscribe('foo')\n >>> # assume someone sends a message along the channel via a publish\n >>> message = p.get_sharded_message()\n\nSimilarly, the same process can be used to acquire sharded pubsub messages, that have already been sent to a specific node, by passing the node to get_sharded_message:\n\n.. code:: python\n\n >>> from redis.cluster import RedisCluster, ClusterNode\n >>> first_node = ClusterNode['localhost', 6379]\n >>> second_node = ClusterNode['localhost', 6380]\n >>> r = RedisCluster(startup_nodes=[first_node, second_node])\n >>> p = r.pubsub()\n >>> p.ssubscribe('foo')\n >>> # assume someone sends a message along the channel via a publish\n >>> message = p.get_sharded_message(target_node=second_node)\n\n\nMonitor\n~~~~~~~\n\nredis-py includes a Monitor object that streams every command processed\nby the Redis server. Use listen() on the Monitor object to block until a\ncommand is received.\n\n.. code:: python\n\n >>> r = redis.Redis(...)\n >>> with r.monitor() as m:\n >>> for command in m.listen():\n >>> print(command)\n" }, { "path": "docs/backoff.rst", "content": ".. _backoff-label:\n\nBackoff\n#############\n\n.. automodule:: redis.backoff\n :members: " }, { "path": "docs/clustering.rst", "content": "Clustering\n==========\n\nredis-py now supports cluster mode and provides a client for `Redis\nCluster `__.\n\nThe cluster client is based on Grokzen’s\n`redis-py-cluster `__, has\nadded bug fixes, and now supersedes that library. Support for these\nchanges is thanks to his contributions.\n\nTo learn more about Redis Cluster, see `Redis Cluster\nspecifications `__.\n\n`Creating clusters <#creating-clusters>`__ \\| `Specifying Target\nNodes <#specifying-target-nodes>`__ \\| `Multi-key\nCommands <#multi-key-commands>`__ \\| `Known PubSub\nLimitations <#known-pubsub-limitations>`__\n\nConnecting to cluster\n---------------------\n\nConnecting redis-py to a Redis Cluster instance(s) requires at a minimum\na single node for cluster discovery. There are multiple ways in which a\ncluster instance can be created:\n\n- Using ‘host’ and ‘port’ arguments:\n\n.. code:: python\n\n >>> from redis.cluster import RedisCluster as Redis\n >>> rc = Redis(host='localhost', port=6379)\n >>> print(rc.get_nodes())\n [[host=127.0.0.1,port=6379,name=127.0.0.1:6379,server_type=primary,redis_connection=Redis>>], [host=127.0.0.1,port=6378,name=127.0.0.1:6378,server_type=primary,redis_connection=Redis>>], [host=127.0.0.1,port=6377,name=127.0.0.1:6377,server_type=replica,redis_connection=Redis>>]]\n\n- Using the Redis URL specification:\n\n.. code:: python\n\n >>> from redis.cluster import RedisCluster as Redis\n >>> rc = Redis.from_url(\"redis://localhost:6379/0\")\n\n- Directly, via the ClusterNode class:\n\n.. code:: python\n\n >>> from redis.cluster import RedisCluster as Redis\n >>> from redis.cluster import ClusterNode\n >>> nodes = [ClusterNode('localhost', 6379), ClusterNode('localhost', 6378)]\n >>> rc = Redis(startup_nodes=nodes)\n\nWhen a RedisCluster instance is being created it first attempts to\nestablish a connection to one of the provided startup nodes. If none of\nthe startup nodes are reachable, a ‘RedisClusterException’ will be\nthrown. After a connection to the one of the cluster’s nodes is\nestablished, the RedisCluster instance will be initialized with 3\ncaches: a slots cache which maps each of the 16384 slots to the node/s\nhandling them, a nodes cache that contains ClusterNode objects (name,\nhost, port, redis connection) for all of the cluster’s nodes, and a\ncommands cache contains all the server supported commands that were\nretrieved using the Redis ‘COMMAND’ output. See *RedisCluster specific\noptions* below for more.\n\nRedisCluster instance can be directly used to execute Redis commands.\nWhen a command is being executed through the cluster instance, the\ntarget node(s) will be internally determined. When using a key-based\ncommand, the target node will be the node that holds the key’s slot.\nCluster management commands and other commands that are not key-based\nhave a parameter called ‘target_nodes’ where you can specify which nodes\nto execute the command on. In the absence of target_nodes, the command\nwill be executed on the default cluster node. As part of cluster\ninstance initialization, the cluster’s default node is randomly selected\nfrom the cluster’s primaries, and will be updated upon reinitialization.\nUsing r.get_default_node(), you can get the cluster’s default node, or\nyou can change it using the ‘set_default_node’ method.\n\nThe ‘target_nodes’ parameter is explained in the following section,\n‘Specifying Target Nodes’.\n\n.. code:: python\n\n >>> # target-nodes: the node that holds 'foo1's key slot\n >>> rc.set('foo1', 'bar1')\n >>> # target-nodes: the node that holds 'foo2's key slot\n >>> rc.set('foo2', 'bar2')\n >>> # target-nodes: the node that holds 'foo1's key slot\n >>> print(rc.get('foo1'))\n b'bar'\n >>> # target-node: default-node\n >>> print(rc.keys())\n [b'foo1']\n >>> # target-node: default-node\n >>> rc.ping()\n\nSpecifying Target Nodes\n-----------------------\n\nAs mentioned above, all non key-based RedisCluster commands accept the\nkwarg parameter ‘target_nodes’ that specifies the node/nodes that the\ncommand should be executed on. The best practice is to specify target\nnodes using RedisCluster class’s node flags: PRIMARIES, REPLICAS,\nALL_NODES, RANDOM. When a nodes flag is passed along with a command, it\nwill be internally resolved to the relevant node/s. If the nodes\ntopology of the cluster changes during the execution of a command, the\nclient will be able to resolve the nodes flag again with the new\ntopology and attempt to retry executing the command.\n\n.. code:: python\n\n >>> from redis.cluster import RedisCluster as Redis\n >>> # run cluster-meet command on all of the cluster's nodes\n >>> rc.cluster_meet('127.0.0.1', 6379, target_nodes=Redis.ALL_NODES)\n >>> # ping all replicas\n >>> rc.ping(target_nodes=Redis.REPLICAS)\n >>> # ping a random node\n >>> rc.ping(target_nodes=Redis.RANDOM)\n >>> # get the keys from all cluster nodes\n >>> rc.keys(target_nodes=Redis.ALL_NODES)\n [b'foo1', b'foo2']\n >>> # execute bgsave in all primaries\n >>> rc.bgsave(Redis.PRIMARIES)\n\nYou could also pass ClusterNodes directly if you want to execute a\ncommand on a specific node / node group that isn’t addressed by the\nnodes flag. However, if the command execution fails due to cluster\ntopology changes, a retry attempt will not be made, since the passed\ntarget node/s may no longer be valid, and the relevant cluster or\nconnection error will be returned.\n\n.. code:: python\n\n >>> node = rc.get_node('localhost', 6379)\n >>> # Get the keys only for that specific node\n >>> rc.keys(target_nodes=node)\n >>> # get Redis info from a subset of primaries\n >>> subset_primaries = [node for node in rc.get_primaries() if node.port > 6378]\n >>> rc.info(target_nodes=subset_primaries)\n\nIn addition, the RedisCluster instance can query the Redis instance of a\nspecific node and execute commands on that node directly. The Redis\nclient, however, does not handle cluster failures and retries.\n\n.. code:: python\n\n >>> cluster_node = rc.get_node(host='localhost', port=6379)\n >>> print(cluster_node)\n [host=127.0.0.1,port=6379,name=127.0.0.1:6379,server_type=primary,redis_connection=Redis>>]\n >>> r = cluster_node.redis_connection\n >>> r.client_list()\n [{'id': '276', 'addr': '127.0.0.1:64108', 'fd': '16', 'name': '', 'age': '0', 'idle': '0', 'flags': 'N', 'db': '0', 'sub': '0', 'psub': '0', 'multi': '-1', 'qbuf': '26', 'qbuf-free': '32742', 'argv-mem': '10', 'obl': '0', 'oll': '0', 'omem': '0', 'tot-mem': '54298', 'events': 'r', 'cmd': 'client', 'user': 'default'}]\n >>> # Get the keys only for that specific node\n >>> r.keys()\n [b'foo1']\n\nMulti-key Commands\n------------------\n\nRedis supports multi-key commands in Cluster Mode, such as Set type\nunions or intersections, mset and mget, as long as the keys all hash to\nthe same slot. By using RedisCluster client, you can use the known\nfunctions (e.g. mget, mset) to perform an atomic multi-key operation.\nHowever, you must ensure all keys are mapped to the same slot, otherwise\na RedisClusterException will be thrown. Redis Cluster implements a\nconcept called hash tags that can be used in order to force certain keys\nto be stored in the same hash slot, see `Keys hash\ntag `__. You can\nalso use nonatomic for some of the multikey operations, and pass keys\nthat aren’t mapped to the same slot. The client will then map the keys\nto the relevant slots, sending the commands to the slots’ node owners.\nNon-atomic operations batch the keys according to their hash value, and\nthen each batch is sent separately to the slot’s owner.\n\n.. code:: python\n\n # Atomic operations can be used when all keys are mapped to the same slot\n >>> rc.mset({'{foo}1': 'bar1', '{foo}2': 'bar2'})\n >>> rc.mget('{foo}1', '{foo}2')\n [b'bar1', b'bar2']\n # Non-atomic multi-key operations splits the keys into different slots\n >>> rc.mset_nonatomic({'foo': 'value1', 'bar': 'value2', 'zzz': 'value3')\n >>> rc.mget_nonatomic('foo', 'bar', 'zzz')\n [b'value1', b'value2', b'value3']\n\n**Cluster PubSub:**\n\nWhen a ClusterPubSub instance is created without specifying a node, a\nsingle node will be transparently chosen for the pubsub connection on\nthe first command execution. The node will be determined by: 1. Hashing\nthe channel name in the request to find its keyslot 2. Selecting a node\nthat handles the keyslot: If read_from_replicas is set to true or\nload_balancing_strategy is provided, a replica can be selected.\n\nKnown PubSub Limitations\n------------------------\n\nPattern subscribe and publish do not currently work properly due to key\nslots. If we hash a pattern like fo\\* we will receive a keyslot for that\nstring but there are endless possibilities for channel names based on\nthis pattern - unknowable in advance. This feature is not disabled but\nthe commands are not currently recommended for use. See\n`redis-py-cluster\ndocumentation `__\nfor more.\n\n.. code:: python\n\n >>> p1 = rc.pubsub()\n # p1 connection will be set to the node that holds 'foo' keyslot\n >>> p1.subscribe('foo')\n # p2 connection will be set to node 'localhost:6379'\n >>> p2 = rc.pubsub(rc.get_node('localhost', 6379))\n\n**Read Only Mode**\n\nBy default, Redis Cluster always returns MOVE redirection response on\naccessing a replica node. You can overcome this limitation and scale\nread commands by triggering READONLY mode.\n\nTo enable READONLY mode pass read_from_replicas=True or define\na load_balancing_strategy to RedisCluster constructor.\nWhen read_from_replicas is set to true read commands will be assigned between\nthe primary and its replications in a Round-Robin manner.\nWith load_balancing_strategy you can define a custom strategy for\nassigning read commands to the replicas and primary nodes.\n\nREADONLY mode can be set at runtime by calling the readonly() method\nwith target_nodes=‘replicas’, and read-write access can be restored by\ncalling the readwrite() method.\n\n.. code:: python\n\n >>> from cluster import RedisCluster as Redis\n # Use 'debug' log level to print the node that the command is executed on\n >>> rc_readonly = Redis(startup_nodes=startup_nodes,\n ... read_from_replicas=True)\n >>> rc_readonly.set('{foo}1', 'bar1')\n >>> for i in range(0, 4):\n ... # Assigns read command to the slot's hosts in a Round-Robin manner\n ... rc_readonly.get('{foo}1')\n # set command would be directed only to the slot's primary node\n >>> rc_readonly.set('{foo}2', 'bar2')\n # reset READONLY flag\n >>> rc_readonly.readwrite(target_nodes='replicas')\n # now the get command would be directed only to the slot's primary node\n >>> rc_readonly.get('{foo}1')\n" }, { "path": "docs/commands.rst", "content": "Redis Commands\n##############\n\nCore Commands\n*************\n\nThe following functions can be used to replicate their equivalent `Redis command `_. Generally they can be used as functions on your redis connection. For the simplest example, see below:\n\nGetting and settings data in redis::\n\n import redis\n r = redis.Redis(decode_responses=True)\n r.set('mykey', 'thevalueofmykey')\n r.get('mykey')\n\n.. autoclass:: redis.commands.core.CoreCommands\n :inherited-members:\n\nSentinel Commands\n*****************\n.. autoclass:: redis.commands.sentinel.SentinelCommands\n :inherited-members:\n\nRedis Cluster Commands\n**********************\n\nThe following `Redis commands `_ are available within a `Redis Cluster `_. Generally they can be used as functions on your redis connection.\n\n.. autoclass:: redis.commands.cluster.RedisClusterCommands\n :inherited-members:\n" }, { "path": "docs/conf.py", "content": "# redis-py documentation build configuration file, created by\n# sphinx-quickstart on Fri Feb 8 00:47:08 2013.\n#\n# This file is execfile()d with the current directory set to its containing\n# dir.\n#\n# Note that not all possible configuration values are present in this\n# autogenerated file.\n#\n# All configuration values have a default; values that are commented out\n# serve to show the default.\n\nimport datetime\nimport os\nimport sys\n\n# If extensions (or modules to document with autodoc) are in another directory,\n# add these directories to sys.path here. If the directory is relative to the\n# documentation root, use os.path.abspath to make it absolute, like shown here.\n# sys.path.insert(0, os.path.abspath('.'))\nsys.path.append(os.path.abspath(os.path.pardir))\n\n# -- General configuration ----------------------------------------------------\n\n# If your documentation needs a minimal Sphinx version, state it here.\n# needs_sphinx = '1.0'\n\n# Add any Sphinx extension module names here, as strings. They can be\n# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom ones.\nextensions = [\n \"nbsphinx\",\n \"sphinx_gallery.load_style\",\n \"sphinx.ext.autodoc\",\n \"sphinx.ext.viewcode\",\n \"sphinx.ext.autosectionlabel\",\n \"sphinx.ext.napoleon\",\n]\n\n# Napoleon settings. We only accept Google-style docstrings.\nnapoleon_google_docstring = True\nnapoleon_numpy_docstring = False\n\n# AutosectionLabel settings.\n# Uses a :