spacy-french-models icon indicating copy to clipboard operation
spacy-french-models copied to clipboard

Morphology `KeyError` while processing '\t'

Open f11r opened this issue 7 years ago • 2 comments

Great to see a first (community created) french model.

When analysing '\t' I get an KeyError in spacy.morphology.Morphology.assign_tag_id (spacy/morphology.cpp:4681)().

Steps to reproduce:

docker run -ti fedora:25 bash
dnf install python3-devel python3-pip gcc-c++ redhat-rpm-config
pip3 install spacy==1.8.0
pip3 install https://github.com/pagesjaunes/spacy-french-models/releases/download/v0.0.1-alpha/fr_model-0.0.1.tar.gz
python3 -m spacy link fr_model fr_default --force

python3 -c "import spacy; nlp = spacy.load('fr_default'); nlp('\t')"

Output:

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/usr/lib64/python3.5/site-packages/spacy/language.py", line 320, in __call__
    proc(doc)
  File "spacy/tagger.pyx", line 212, in spacy.tagger.Tagger.__call__ (spacy/tagger.cpp:6245)
  File "spacy/morphology.pyx", line 76, in spacy.morphology.Morphology.assign_tag_id (spacy/morphology.cpp:4681)
KeyError: 172049

Not sure whether that is a problem with the french model or actually with spacy itself. Maybe @honnibal or @ines can take a look at this as well?

f11r avatar Apr 19 '17 14:04 f11r