Python 2: One syntax error and two undefined names / Python 3: Five undefined names

[cclauss]

Python 2

See: https://travis-ci.org/explosion/spaCy/builds/304678969

python2 -m flake8 . --count --exclude=spacy/lang --select=E901,E999,F821,F822,F823 --show-source --statistics

./examples/information_extraction/entity_relations.py:39:18: E999 SyntaxError: invalid syntax
    for span in [*list(doc.ents), *list(doc.noun_chunks)]:
                 ^

./spacy/tests/lang/fr/test_lemmatization.py:25:14: F821 undefined name 'FR'
    tokens = FR("il y a des Costaricienne.")
             ^

./spacy/tests/regression/test_issue910.py:81:11: F821 undefined name 'Adam'
    sgd = Adam(nlp.entity.model[0].ops, 0.001)
          ^

1     E999 SyntaxError: invalid syntax
2     F821 undefined name 'FR'
3

Python 3

See: https://travis-ci.org/explosion/spaCy/builds/304678969

flake8 . --count --exclude=spacy/lang --select=E901,E999,F821,F822,F823 --show-source --statistics

./spacy/compat.py:56:16: F821 undefined name 'unicode'
    unicode_ = unicode
               ^

./spacy/compat.py:57:19: F821 undefined name 'basestring'
    basestring_ = basestring
                  ^

./spacy/compat.py:58:14: F821 undefined name 'raw_input'
    input_ = raw_input
             ^

./spacy/tests/lang/fr/test_lemmatization.py:25:14: F821 undefined name 'FR'
    tokens = FR("il y a des Costaricienne.")
             ^

./spacy/tests/regression/test_issue910.py:81:11: F821 undefined name 'Adam'
    sgd = Adam(nlp.entity.model[0].ops, 0.001)
          ^

5     F821 undefined name 'unicode'

[ines]

Thanks a lot for the detailed report – I hope you don't mind that I merged the two issues, since they're essentially related to the same thing.

The spacy/compat.py errors should probably be excluded via # noqa: F821. Those are the Python 2/3-specific versions of the built-ins.

I'm surprised that the tests passed... will check, but I assume test_issue910.py is likely xfailed, and the French lemmatization tests require the models and are only run if --models --fr is set when running pytest (which we're not doing on Travis, as it's too time consuming and potentially flakey at the moment). Should be easy to fix by adding the FR fixture to the test arguments.

for span in [*list(doc.ents), *list(doc.noun_chunks)]:

Damn, this is a good catch. Looks like we only ever ran this example in Python 3, so will fix that to make it compatible.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.