Work Log

Week 1 (17.02.15 to 23.02.15): getting started

Week 2 (24.02.15 to 02.03.15): packaging

Week 3 (03.03.15 to 09.03.15): packaging

• √ choose a tool for report
• √ package sentence splitter from begin (Bluima) to end (Sherlok) and test it

Week 4 (10.03.15 to 16.03.15):

• √ fix issues from last week
• √ transform PoS, Token, Chunck annotator
• x (opt) transform Parser annotator

Week 5 (17.03.15 to 23.03.15):

• x In case of ClassNotFound in ModelProxy, print all reachable classes that implement ModelResource for debug purpose. This was discarded.
• √ Instead a simple error message should be added to the log.
• √ report: how are annotator and resources (typically) converted
• √ Add warning log when using OpenNLP
• √ Add some tests in pipelines
• √ Add engines' description in bundle
• √ Convert BlueBioLemmatizer

Week 6 (24.03.15 to 30.03.15):

• ~ Convert MeasureRegexAnnotators
  • proxy to RegExAnnotator which is not yet compatible with JCas
• ~ Convert PruneMeasuresAnnotator
  • similar to KeepLargestAnnotationAnnotator
• ~ Convert LinnaeusAnnotator
  • the dependency linnaeus.jar is not really flexible.
• ~ Convert BrainRegionAnnotator
  • need to devise strategy for writing data in train mode
• √ Convert KeepLargestAnnotationAnnotator
• ~ Wiki: how to create resources
  • depends on pending decision about multiple files resources
• √ Test REST POST annotate query
  • doesn't work atm since the body of the query is not used FIXED.

Week 7 (31.03.15 to 06.04.15):

• Packaging Issue and New Strategy

Week 8 (07.04.15 to 13.04.15): Holidays

Week 9 (14.04.15 to 20.04.15):

• √ Add issue with duplicating resources when using sherlok as a multi-instance server
• √ Replace PruneMeasuresAnnotator with KeepLargestAnnotationAnnotator
• √ Create the different KeepLargestAnnotationAnnotator engines
• ~ Convert ExtractSameCoocurrences
• √ Convert JsreFilterAnnotator
• ~ Report (fixes)

Week 10 (21.04.15 to 27.04.15):

• ~ Convert MeasureRegexAnnotators.getAllAED() and RegExAnnotator.
  • combined effort with Renaud
• √ Explore usage of vars in Sherlok config (e.g. for paths)
• √ Report

Week 11 (28.04.15 to 04.05.15):

• √ Config var: more refactoring in order to support git://
• √ Report
• ~ Write unit test for POST query for tests (not enough time)
• ~ Fix ReferencesClassifierAnnotatorTest (not enough time)

Week 12 (05.05.15 to 11.05.15):

• √ More report writing
• √ Write unit test for POST query for tests
• ~ Add support for http resource file
  • need improvement
• √ Write unit tests for configuration variables
  • √ GitConfigVariable
  • √ ConfigVariableManager
  • √ HttpConfigVariable (partial)
• √ fix RegExAnnotator

Week 13 (12.05.15 to 18.05.15):

• √ fully support http config var (preserve name)
• ~ test on Ubuntu
  • see here
  • Fix ReferencesClassifierAnnotatorTest (will follow shortly)
• √ investigate coverage
  • relatively ok but neuroner prevent tests to pass and therefore to get proper coverage report.
  • without it, coverage good.
• √ using bluima v1.0.1
• ~ test on Windows
  • not conclusive at all
  • see here
• √ add licence
• √ add DELETE /clean/runtime_resources/{type} query
• √ work on report

Week 14 (19.05.15 to 25.05.15):

• √ Fix ReferencesClassifierAnnotatorTest
• √ TOP-Down based brain regions
• √ Rules based brain regions
• ~ Lex-NER
  • see here for feasability analysis
• √ More report

Week 15+:

• √ report & doc
• √ defence