SYMBOL INDEX (222 symbols across 29 files)

FILE: mdr/qa/basic_tokenizer.py
  class Tokens (line 16) | class Tokens(object):
    method __init__ (line 25) | def __init__(self, data, annotators, opts=None):
    method __len__ (line 30) | def __len__(self):
    method slice (line 34) | def slice(self, i=None, j=None):
    method untokenize (line 40) | def untokenize(self):
    method words (line 44) | def words(self, uncased=False):
    method offsets (line 55) | def offsets(self):
    method pos (line 59) | def pos(self):
    method lemmas (line 67) | def lemmas(self):
    method entities (line 75) | def entities(self):
    method ngrams (line 83) | def ngrams(self, n=1, uncased=False, filter_fn=None, as_strings=True):
    method entity_groups (line 110) | def entity_groups(self):
  class Tokenizer (line 132) | class Tokenizer(object):
    method tokenize (line 137) | def tokenize(self, text):
    method shutdown (line 140) | def shutdown(self):
    method __del__ (line 143) | def __del__(self):
  class RegexpTokenizer (line 153) | class RegexpTokenizer(Tokenizer):
    method __init__ (line 172) | def __init__(self, **kwargs):
    method tokenize (line 196) | def tokenize(self, text):
  class SimpleTokenizer (line 236) | class SimpleTokenizer(Tokenizer):
    method __init__ (line 240) | def __init__(self, **kwargs):
    method tokenize (line 254) | def tokenize(self, text):

FILE: mdr/qa/config.py
  class ClusterConfig (line 12) | class ClusterConfig(NamedTuple):
  function common_args (line 16) | def common_args():
  function train_args (line 57) | def train_args():

FILE: mdr/qa/data_utils.py
  function explore (line 11) | def explore(path):
  function load_corpus (line 23) | def load_corpus(corpus_path="/private/home/xwhan/data/hotpot/tfidf/abstr...

FILE: mdr/qa/hotpot_evaluate_v1.py
  function normalize_answer (line 13) | def normalize_answer(s):
  function f1_score (line 31) | def f1_score(prediction, ground_truth):
  function exact_match_score (line 54) | def exact_match_score(prediction, ground_truth):
  function update_answer (line 57) | def update_answer(metrics, prediction, gold):
  function update_sp (line 66) | def update_sp(metrics, prediction, gold):
  function eval (line 88) | def eval(prediction_file, gold_file):

FILE: mdr/qa/qa_dataset.py
  function collate_tokens (line 17) | def collate_tokens(values, pad_idx, eos_idx=None, left_pad=False, move_e...
  function prepare (line 38) | def prepare(item, tokenizer, special_toks=["[SEP]", "[unused1]", "[unuse...
  class QAEvalDataset (line 108) | class QAEvalDataset(Dataset):
    method __init__ (line 110) | def __init__(self,
    method __len__ (line 150) | def __len__(self):
    method __getitem__ (line 153) | def __getitem__(self, index):
  class QADataset (line 188) | class QADataset(Dataset):
    method __init__ (line 190) | def __init__(self,
    method __len__ (line 302) | def __len__(self):
    method __getitem__ (line 305) | def __getitem__(self, index):
  class MhopSampler (line 391) | class MhopSampler(Sampler):
    method __init__ (line 396) | def __init__(self, data_source, num_neg=9, n_gpu=8):
    method __len__ (line 408) | def __len__(self):
    method __iter__ (line 411) | def __iter__(self):
  function qa_collate (line 424) | def qa_collate(samples, pad_id=0):

FILE: mdr/qa/qa_model.py
  class BertPooler (line 13) | class BertPooler(nn.Module):
    method __init__ (line 14) | def __init__(self, config):
    method forward (line 19) | def forward(self, hidden_states):
  class QAModel (line 27) | class QAModel(nn.Module):
    method __init__ (line 29) | def __init__(self,
    method forward (line 49) | def forward(self, batch):

FILE: mdr/qa/qa_trainer.py
  class TrainerState (line 40) | class TrainerState:
    method save (line 52) | def save(self, filename: str) -> None:
    method load (line 61) | def load(cls, filename: str, default: "TrainerState", gpu: int) -> "Tr...
  class Trainer (line 78) | class Trainer:
    method __init__ (line 79) | def __init__(self, train_cfg: NamedTuple, cluster_cfg: ClusterConfig) ...
    method __call__ (line 83) | def __call__(self) -> Optional[float]:
    method log (line 93) | def log(self, log_data: dict):
    method checkpoint (line 101) | def checkpoint(self, rm_init=True) -> submitit.helpers.DelayedSubmission:
    method _setup_process_group (line 117) | def _setup_process_group(self) -> None:
    method _init_state (line 128) | def _init_state(self) -> None:
    method _train (line 209) | def _train(self) -> Optional[float]:
    method _eval (line 283) | def _eval(self) -> dict:

FILE: mdr/qa/train_ranker.py
  function load_saved (line 30) | def load_saved(model, path):
  function main (line 37) | def main():
  function predict (line 212) | def predict(args, model, eval_dataloader, device, logger):

FILE: mdr/qa/utils.py
  function set_global_logging_level (line 13) | def set_global_logging_level(level=logging.ERROR, prefices=[""]):
  function load_saved (line 29) | def load_saved(model, path, exact=True):
  function move_to_cuda (line 45) | def move_to_cuda(sample):
  function convert_to_half (line 65) | def convert_to_half(sample):
  class AverageMeter (line 85) | class AverageMeter(object):
    method __init__ (line 88) | def __init__(self):
    method reset (line 91) | def reset(self):
    method update (line 97) | def update(self, val, n=1):
  function normalize (line 104) | def normalize(text):
  function para_has_answer (line 109) | def para_has_answer(answer, para, tokenizer):
  function match_answer_span (line 124) | def match_answer_span(p, answer, tokenizer, match="string"):
  function _is_whitespace (line 145) | def _is_whitespace(char):
  function _improve_answer_span (line 160) | def _improve_answer_span(doc_tokens, input_start, input_end, tokenizer,
  function whitespace_tokenize (line 173) | def whitespace_tokenize(text):
  function find_ans_span_with_char_offsets (line 182) | def find_ans_span_with_char_offsets(detected_ans, char_to_word_offset, d...
  function convert_to_unicode (line 212) | def convert_to_unicode(text):
  function _is_control (line 232) | def _is_control(char):
  function _is_punctuation (line 243) | def _is_punctuation(char):
  class BasicTokenizer (line 259) | class BasicTokenizer(object):
    method __init__ (line 262) | def __init__(self, do_lower_case=True):
    method tokenize (line 269) | def tokenize(self, text):
    method _run_strip_accents (line 284) | def _run_strip_accents(self, text):
    method _run_split_on_punc (line 295) | def _run_split_on_punc(self, text):
    method _clean_text (line 315) | def _clean_text(self, text):
  function get_final_text (line 329) | def get_final_text(pred_text, orig_text, do_lower_case=False, verbose_lo...

FILE: mdr/retrieval/config.py
  class ClusterConfig (line 10) | class ClusterConfig(NamedTuple):
  function common_args (line 14) | def common_args():
  function train_args (line 71) | def train_args():
  function encode_args (line 107) | def encode_args():

FILE: mdr/retrieval/criterions.py
  function mhop_loss (line 114) | def mhop_loss(model, batch, args):
  function mhop_eval (line 153) | def mhop_eval(outputs, args):
  function unified_loss (line 185) | def unified_loss(model, batch, args):
  function unified_eval (line 212) | def unified_eval(outputs, batch):

FILE: mdr/retrieval/decomposed_analysis.py
  function decomposed_errors (line 9) | def decomposed_errors():
  function collect_gold_decomposition (line 62) | def collect_gold_decomposition():
  function qdmr_utils (line 99) | def qdmr_utils():
  function analyze_results (line 128) | def analyze_results():

FILE: mdr/retrieval/mhop_trainer.py
  class TrainerState (line 42) | class TrainerState:
    method save (line 54) | def save(self, filename: str) -> None:
    method load (line 63) | def load(cls, filename: str, default: "TrainerState", gpu: int) -> "Tr...
  class Trainer (line 80) | class Trainer:
    method __init__ (line 81) | def __init__(self, train_cfg: NamedTuple, cluster_cfg: ClusterConfig) ...
    method __call__ (line 85) | def __call__(self) -> Optional[float]:
    method log (line 95) | def log(self, log_data: dict):
    method checkpoint (line 103) | def checkpoint(self, rm_init=True) -> submitit.helpers.DelayedSubmission:
    method _setup_process_group (line 119) | def _setup_process_group(self) -> None:
    method _init_state (line 130) | def _init_state(self) -> None:
    method _train (line 204) | def _train(self) -> Optional[float]:
    method _eval (line 271) | def _eval(self) -> float:

FILE: mdr/retrieval/single_trainer.py
  class TrainerState (line 41) | class TrainerState:
    method save (line 53) | def save(self, filename: str) -> None:
    method load (line 62) | def load(cls, filename: str, default: "TrainerState", gpu: int) -> "Tr...
  class Trainer (line 79) | class Trainer:
    method __init__ (line 80) | def __init__(self, train_cfg: NamedTuple, cluster_cfg: ClusterConfig) ...
    method __call__ (line 84) | def __call__(self) -> Optional[float]:
    method log (line 94) | def log(self, log_data: dict):
    method checkpoint (line 102) | def checkpoint(self, rm_init=True) -> submitit.helpers.DelayedSubmission:
    method _setup_process_group (line 118) | def _setup_process_group(self) -> None:
    method _init_state (line 129) | def _init_state(self) -> None:
    method _train (line 203) | def _train(self) -> Optional[float]:
    method _eval (line 276) | def _eval(self) -> float:

FILE: mdr/retrieval/train_single.py
  function main (line 111) | def main():
  function predict (line 295) | def predict(args, model, eval_dataloader, device, logger):

FILE: mdr/retrieval/utils/basic_tokenizer.py
  class Tokens (line 18) | class Tokens(object):
    method __init__ (line 27) | def __init__(self, data, annotators, opts=None):
    method __len__ (line 32) | def __len__(self):
    method slice (line 36) | def slice(self, i=None, j=None):
    method untokenize (line 42) | def untokenize(self):
    method words (line 46) | def words(self, uncased=False):
    method offsets (line 57) | def offsets(self):
    method pos (line 61) | def pos(self):
    method lemmas (line 69) | def lemmas(self):
    method entities (line 77) | def entities(self):
    method ngrams (line 85) | def ngrams(self, n=1, uncased=False, filter_fn=None, as_strings=True):
    method entity_groups (line 112) | def entity_groups(self):
  class Tokenizer (line 134) | class Tokenizer(object):
    method tokenize (line 139) | def tokenize(self, text):
    method shutdown (line 142) | def shutdown(self):
    method __del__ (line 145) | def __del__(self):
  class RegexpTokenizer (line 155) | class RegexpTokenizer(Tokenizer):
    method __init__ (line 174) | def __init__(self, **kwargs):
    method tokenize (line 198) | def tokenize(self, text):
  class SimpleTokenizer (line 238) | class SimpleTokenizer(Tokenizer):
    method __init__ (line 242) | def __init__(self, **kwargs):
    method tokenize (line 256) | def tokenize(self, text):
  function normalize (line 303) | def normalize(text):
  function filter_word (line 308) | def filter_word(text):
  function filter_ngram (line 317) | def filter_ngram(gram, mode='any'):

FILE: mdr/retrieval/utils/mhop_utils.py
  function pick_bridge_v0 (line 16) | def pick_bridge_v0(title2linked, title2doc, titles, q, ans):
  function load_annotated (line 31) | def load_annotated(path="/private/home/xwhan/data/hotpot/tfidf/abstracts...
  function normalize_answer (line 37) | def normalize_answer(s):
  function hotpot_sp_data (line 55) | def hotpot_sp_data(raw_path):
  function add_qid (line 106) | def add_qid(raw_path):
  function add_bridge_ann (line 135) | def add_bridge_ann(raw_path):
  function check_2hop (line 166) | def check_2hop(raw_path):
  function add_sp_labels (line 173) | def add_sp_labels(raw_path, input_file, save_path,
  function explore_QDMR (line 212) | def explore_QDMR(path="/private/home/xwhan/data/Break-dataset/QDMR-high-...
  function add_sents_to_corpus_dict (line 249) | def add_sents_to_corpus_dict():

FILE: mdr/retrieval/utils/tokenizer.py
  function convert_tokens_to_ids (line 32) | def convert_tokens_to_ids(vocab, tokens):
  function whitespace_tokenize (line 39) | def whitespace_tokenize(text):
  function convert_to_unicode (line 48) | def convert_to_unicode(text):
  function _is_whitespace (line 68) | def _is_whitespace(char):
  function _is_control (line 80) | def _is_control(char):
  class BasicTokenizer (line 91) | class BasicTokenizer(object):
    method __init__ (line 94) | def __init__(self, do_lower_case=True):
    method tokenize (line 101) | def tokenize(self, text):
    method _run_strip_accents (line 116) | def _run_strip_accents(self, text):
    method _run_split_on_punc (line 127) | def _run_split_on_punc(self, text):
    method _clean_text (line 147) | def _clean_text(self, text):
  function _is_punctuation (line 161) | def _is_punctuation(char):
  function process (line 177) | def process(s, tokenizer):

FILE: mdr/retrieval/utils/utils.py
  function load_saved (line 10) | def load_saved(model, path, exact=True):
  function move_to_cuda (line 24) | def move_to_cuda(sample):
  function convert_to_half (line 43) | def convert_to_half(sample):
  class AverageMeter (line 63) | class AverageMeter(object):
    method __init__ (line 66) | def __init__(self):
    method reset (line 69) | def reset(self):
    method update (line 75) | def update(self, val, n=1):
  function normalize (line 82) | def normalize(text):
  class DocDB (line 87) | class DocDB(object):
    method __init__ (line 93) | def __init__(self, db_path=None):
    method __enter__ (line 97) | def __enter__(self):
    method __exit__ (line 100) | def __exit__(self, *args):
    method close (line 103) | def close(self):
    method get_doc_ids (line 107) | def get_doc_ids(self):
    method get_doc_text (line 115) | def get_doc_text(self, doc_id):
  function para_has_answer (line 126) | def para_has_answer(answer, para, tokenizer):
  function complex_ans_recall (line 142) | def complex_ans_recall():

FILE: scripts/demo.py
  function init_retrieval (line 28) | def init_retrieval(args):
  function init_reader (line 54) | def init_reader(args):

FILE: scripts/encode_corpus.py
  function main (line 41) | def main():
  function predict (line 95) | def predict(model, eval_dataloader):

FILE: scripts/end2end.py
  function convert_hnsw_query (line 49) | def convert_hnsw_query(query_vectors):

FILE: scripts/eval/eval_mhop_retrieval.py
  function convert_hnsw_query (line 44) | def convert_hnsw_query(query_vectors):

FILE: scripts/eval/eval_retrieval.py
  function init (line 61) | def init():
  function get_score (line 66) | def get_score(answer_doc, topk=20):
  function add_marker_q (line 85) | def add_marker_q(tokenizer, q):

FILE: scripts/train_mhop.py
  function main (line 54) | def main():
  function predict (line 233) | def predict(args, model, eval_dataloader, device, logger):

FILE: scripts/train_momentum.py
  function main (line 28) | def main():
  function predict (line 214) | def predict(args, model, eval_dataloader, device, logger):

FILE: scripts/train_qa.py
  function load_saved (line 33) | def load_saved(model, path):
  function main (line 40) | def main():
  function predict (line 220) | def predict(args, model, eval_dataloader, logger, fixed_thresh=None):
  function eval_final (line 380) | def eval_final(args, model, eval_dataloader, weight=0.8, gpu=True):

FILE: submitit/submitit_train.py
  function get_shared_folder (line 20) | def get_shared_folder() -> Path:
  function get_init_file (line 23) | def get_init_file() -> Path:
  function grid_parameters (line 31) | def grid_parameters(grid: Dict):
  function grid_search (line 43) | def grid_search(args):

FILE: submitit/submitit_train_qa.py
  function get_shared_folder (line 19) | def get_shared_folder() -> Path:
  function get_init_file (line 22) | def get_init_file() -> Path:
  function grid_parameters (line 30) | def grid_parameters(grid: Dict):
  function grid_search (line 42) | def grid_search(args):