coreference_scores_file (jsonl) - { 'doc_id' : str, 'pairwise_coreference_scores' : List[(s_1, e_1), (s_2, e_2), float (3 sig

test_file contains atleast - doc_id, sections, sentences, ner in scirex format. output_file - { 'doc_id' : str, 'saliency' :

span_prediction_file (jsonl) needs atleast three fields - doc_id, words: List[str], field: List[Tuple[start_index, end_index, type]]

The method that processes the children of a BeautifulSoup tag

Save a list as text, one entry per line.

Get the tokens and NER tags. Ignore the POS tags.

Sorts the instances by their padding lengths, using the keys in ``sorting_keys`` (in the order in which they are provided). ``sorting_keys``

Returns the list of tokens that covers the given list of character spans

To compute coreference evaluation metrics, the evaluator needs access to the list of coreference clusters, given in the same form as

TODO(dwadden) document me.

Loop over all event argument predictions. Check whether the event trigger and the argument ner tag agree with the role of the argumen