flake8 testing of https://github.com/juand-r/entity-recognition-datasets on Python 3.7.0

$ flake8 . --count --select=E901,E999,F821,F822,F823 --show-source --statistics

./data/NIST_IEER/CONLL-format/utils/quick_comma_fix.py:41:37: E999 SyntaxError: invalid syntax
                    print annotations
                                    ^
./data/NIST_IEER/

@stevenbird
Initial release of the corpus available at: https://doi.org/10.7488/ds/1411

Suggested NLTK name: ARCOSG

I have updated and corrected the corpus for inclusion in NLTK. (The one at the link above is older and shouldn't be used).

Corpus reader code verified:
`arcosg = LazyCorpusLoader(
'arcosg',
CategorizedTaggedCorpusReader,
r'.*.txt',
cat_file='cats.prn',

Feature just like "Suggest" in PML-TQ web, which allows a user to pick nodes in a tree in search results (words in a concordance line, in case of Kontext), visualise the nodes (words) with all their attributes and pick the ones that should be part of the new query.

See e.g. http://hdl.handle.net/11346/PMLTQ-ELDJ, execute the query and click on some nodes in a resulting tree. Suggest button w

Database says 66k reviews while UI says 15k,

As of now, the codebase lacks a proper set of offline documentation (docstrings). I think it may be a good idea to add it.

The current version of corpus provides raw text of tweets which violates Twitter policy. We should provide raw message ids instead and re-construct the corpus from these ids.