Textblob

------------------
- Fix display (``__repr__``) of WordList slices on Python 3.
- Add download_corpora module. Corpora must now be downloaded using ``python -m textblob.download_corpora``.

------------------
- Sentiment analyzers return namedtuples, e.g. ``Sentiment(polarity=0.12, subjectivity=0.34)``.
- Memory usage improvements to NaiveBayesAnalyzer and basic_extractor (default feature extractor for classifiers module).
- Add ``textblob.tokenizers.sent_tokenize`` and ``textblob.tokenizers.word_tokenize`` convenience functions.
- Add ``textblob.classifiers.MaxEntClassifer``.
- Improved NLTKTagger.

------------------
- Fix bug in spelling correction that stripped some punctuation (Issue 48).
- Various improvements to spelling correction: preserves whitespace characters (Issue 12); handle contractions and punctuation between words. Thanks davidnk.
- Make ``TextBlob.words`` more memory-efficient.
- Translator now sends POST instead of GET requests. This allows for larger bodies of text to be translated (Issue 49).
- Update pattern tagger for better accuracy.

------------------
- Fix bug that caused ``ValueError`` upon sentence tokenization. This removes modifications made to the NLTK sentence tokenizer.
- Add ``Word.lemmatize()`` method that allows passing in a part-of-speech argument.
- ``Word.lemma`` returns correct part of speech for Word objects that have their ``pos`` attribute set. Thanks RomanYankovsky.

------------------
- *Backwards-incompatible*: Renamed package to ``textblob``. This avoids clashes with other namespaces called `text`. TextBlob should now be imported with ``from textblob import TextBlob``.
- Update pattern resources for improved parser accuracy.
- Update NLTK.
- Allow Translator to connect to proxy server.
- PerceptronTagger completely deprecated. Install the ``textblob-aptagger`` extension instead.

------------------
- Bugfix updates.
- Fix bug in feature extraction for ``NaiveBayesClassifier``.
- ``basic_extractor`` is now case-sensitive, e.g. contains(I) != contains(i)
- Fix ``repr`` output when a TextBlob contains non-ascii characters.
- Fix part-of-speech tagging with ``PatternTagger`` on Windows.
- Suppress warning about not having scikit-learn installed.