Changelog¶

Version 0.3¶

Code name: Packaging

type checking and remove support for Py2
add tox.ini, requirements.txt, .travis.yml, LICENSE, MANIFEST, .pypirc
doc v0.3 [OK 50%]
- changing doc to API style of yellowbrick[OK]
- include data for examples[OK]
- include datasets rst
test v0.3
separate aligning and helpers in utils.aling.py [OK]
doc ngrams.py + Basic docstring of sn_gram [OK]
separated pack for ngrams, because sngrams is to long [OK]
embedding notebook examples to the sphinx doc [OK]
translate to English the notebooks involved.
include the test data from NLP course for Text Normalization Jupyter ntb
define if the hypenation functions must be in normalization [OK]
Write the Quickstart [OK, 10%]
Finish @Appender to link docstrings to .py with close to 200 LOC

Code name: Testing in Reality

str input restriction management with only one decorator (if: isinstance(input,str) do, else: raise error)[NO, replaced by type checking en release 0.3]
Revisar los init y comprobar que no se cargan variables y funciones que no deben.
Poner privadas todas las variables y funciones que deben estarlo.
Put the normalization funcs in shallow module[NO]
Set punctuation in shallow module
add collocations with nltk
made some changes to regular expressions in normalize.py after PSTS experiments
add first version of pdftotext
test configurations for documentation with Sphinx
ES support for Regular Expressions y los text normalization methods.
Revisar sí las funciones viejas donde se especificaba la codificación han sido cambiadas a Py3.

Code name: preproc-tech

first version of pipeline
lemmatization, stemming
ngrams added, also stowords_ngrams, contextual_ngrams, and syntactic_ngrams (as sngrams)
Notebook of POS
Syntactic Dependency Parsing con Stanford
Syntactic Tree Parsing con Stanford
multioutputs for every tagger func
NER con Stanford
POS con Stanford[OK]