General list of tasks of merging nalaf, nala, relna

[juanmirocks]

• We merge nalaf / split branch into develop
• We merge ashish branch into develop
• We make sure that relna depends on develop branch of nalaf
• Implement bootstrapping of relationships as it's done for entities / annotations #151 : Move bootstrapping to nalaf as it is (later we gradually move things)
• Remove dependencies on IDs #155 : Leave Mutation constants in nalaf but mark as deprecated & in nalaf recreate the constants and use them from nala only
• Rename in nalaf default_model to example_entity_moel
• Make nalaf depend on textblob or even spacy #163 Make nalaf depend on textblob (nltk)
• Make nalaf depend on textblob or even spacy #163 Maybe/Someday make nalaf use and depend on spacy rather than textblob
• Move relna relations.py into nalaf #164 Replace relna relations.py in nalaf — Move to relations.py to nalaf: EdgeFeatureGenerator, TokenFeatureGenerator, NamedEntityCountFeatureGenerator, BagOfWords, StemmedBag, SentenceFeature, WOrlfFilter
• Move relna Document Level evaluator to nalaf #165 Move Document Level evaluator to nalaf
• Move from relna the standard error and deviation to evaluators #166 Move the standard error and deviation to evaluators (SE not calculated for docRelationsEvaluator at the moment)
• Move from relna the standard error and deviation to evaluators #166 nala team will make the standard error component a general one
• [ ] Remove GNormPlusTagger from taggers.py (doesn't belong to nalaf really or otherwise put into separated file) --> No priority
• [ ] Rename taggers.py to annotators.py > Rename taggers.py to annotators.py #171
• Make StubSameSentenceRelationExtractor general for list of entities Make StubSameSentenceRelationExtractor general for list of entities #172
• [ ] Remove SameSentence stub from nala already done
• ~~Annotator Hierarchy #77 Make PostProcessor & PreProcessor subclasses of Annotator and have that in nalaf
• Remove edges.py from relna
• [ ] move label_edges to a preprocessing file (in nalaf) to labelers.py move label_edges (data.py) to a preprocessing file to labelers.py #173
• Rename all functions & variables of annotation to entity (like pred_annotations) Rename all functions & variables of annotation to entity (like pred_annotations) #174
• [ ] Merge form_pred_relations from ashish branch to nalaf DID NOT FOUND THIS
• [ ] Merge fold with dev & test from ashish branch to nalaf DID NOT FOUND THIS
• Merge ashish branch list of parse trees to Part class and write down comment that this could be perhaps too specific and perhaps be removed from here
• [ ] Merge ashish entity as a list of tokens to nalaf Likely remove Parts' tokens #167
• [ ] Move edge / toke score-like features from ashish to features in nalaf Move relna relation features to nalaf #170 and Move relna relations.py into nalaf #164
• Put comment down in Entity head to specify that this is not the first token, just the head of the entity as declared by parsing
• [ ] Move Dataset_pipelines.py class make it implement Annotator and move it to annotators.py No priority
• [ ] In nala rename DocumentSelectorPipeline to DocumentSelectorPMIDPipeline & make subclass of DocumentSelector & Define DocumentSelector interface in nalaf No priority
• Move corpora from nalaf to nala

[juanmirocks]

#155

[juanmirocks]

Reviewed all items. Left undone or without specific assigned task for those things currently not crucial for nala

[juanmirocks]

(SE not calculated for docRelationsEvaluator at the moment)