nlp_architect.models.cross_doc_coref package

Subpackages

• nlp_architect.models.cross_doc_coref.system package
  • Subpackages
    • nlp_architect.models.cross_doc_coref.system.sieves package
      • Submodules
      • nlp_architect.models.cross_doc_coref.system.sieves.run_sieve_system module
      • nlp_architect.models.cross_doc_coref.system.sieves.sieves module
      • Module contents
  • Submodules
  • nlp_architect.models.cross_doc_coref.system.cdc_utils module
  • nlp_architect.models.cross_doc_coref.system.sieves_container_init module
  • Module contents

Submodules

nlp_architect.models.cross_doc_coref.sieves_config module

class nlp_architect.models.cross_doc_coref.sieves_config.EntitySievesConfiguration

Bases: nlp_architect.models.cross_doc_coref.sieves_config.SievesConfiguration

class nlp_architect.models.cross_doc_coref.sieves_config.EventSievesConfiguration

Bases: nlp_architect.models.cross_doc_coref.sieves_config.SievesConfiguration

class nlp_architect.models.cross_doc_coref.sieves_config.SievesConfiguration

Bases: object

run_evaluation

Should run evaluation (True/False)

sieves_order

Sieve definition and Sieve running order

Tuple[SieveType, RelationType, Threshold(float)] - define sieves to run, were

Strict- Merge clusters only in case all mentions has current relation between them, Relax- Merge clusters in case (matched mentions) / len(cluster_1.mentions)) >= thresh, Very_Relax- Merge clusters in case (matched mentions) / (all possible pairs) >= thresh

RelationType represent the type of sieve to run.

nlp_architect.models.cross_doc_coref.sieves_resource module

class nlp_architect.models.cross_doc_coref.sieves_resource.SievesResources

Bases: object

elastic_host

Elastic host, #Required when using Elastic evaluation using Wikipedia sieve

elastic_index

Elastic index name, #Required when using Elastic evaluation using Wikipedia sieve

elastic_port

Elastic port number, #Required when using Elastic evaluation using Wikipedia sieve

elmo_file

Location of Elmo mini data set file, #Required mini data set file location for Offline evaluation using GloVe sieve

embed_search_method

EmbeddingMethod.GLOVE, EmbeddingMethod.GLOVE_OFFLINE, EmbeddingMethod.ELMO, EmbeddingMethod.ELMO_OFFLINE

Type:Wordnet search method type, one of

eval_output_dir

The output dir of the evaluation files, here scorer file for cross doc coref spans will be saved

glove_file

Location of GloVe mini data set file, #Required mini data set file location for Offline evaluation using GloVe sieve

referent_dict_file

Location of Referent dic data set file, #Required mini data set file for Offline evaluation or original file for Online evaluation using Referent Dict sieve

referent_dict_method

OnlineOROfflineMethod.ONLINE, OnlineOROfflineMethod.OFFLINE

Type:Referent Dict search method type, one of

vo_dict_file

Location of VerbOcean data set file, #Required mini data set file for Offline evaluation or original file for Online evaluation using VerbOcean sieve

vo_search_method

OnlineOROfflineMethod.ONLINE, OnlineOROfflineMethod.OFFLINE

Type:VerbOcean search method type, one of

wd_file

Location of Within doc data set file, #Required when using Within doc sieve

wiki_folder

Location of Wikipedia mini data set file, #Required mini data set file location for Offline evaluation using Wikipedia sieve

wiki_search_method

WikipediaSearchMethod.ONLINE, WikipediaSearchMethod.OFFLINE, WikipediaSearchMethod.ELASTIC

Type:Wikipedia search method type, one of

wn_folder

Location of WordNet mini data set file, #Required mini data set file location for Offline evaluation using WordNet sieve

wn_search_method

OnlineOROfflineMethod.ONLINE, OnlineOROfflineMethod.OFFLINE

Type:Wordnet search method type, one of

Module contents