Tasks

  • POS
  • morphological analysis
  • transliteration

Library Interface

  • Sentiment
  • NER
  • Frequency based comparison

Command Line interface

  • Sentiment
  • Reading stdin column format

Infrastructure

  • Cache models
  • Add normalization to the embeddings
  • Detect supported languages
    • added task/lang as part of the identifier, what is left is to iterate over the collections.
  • Throw different exception for missing package than undownloaded one
  • Define NotSupportedLanguage/Task Exception for the downloader
  • Remove noun phrases support.
  • Train more/new POS taggers

Documentation

  • Add a quick tutorial
  • Embed demos in our documentation
  • pycld2 README
  • Update rtdcs with the new submodules.