Fix : ValueError: numpy.ndarray size changed, may indicate binary inc…

…ompatibility. Expected 88 from C header, got 80 from PyObject

See Similar to piskvorky/gensim#3097

.gitignore

@@ -83,6 +83,7 @@ celerybeat-schedule
 
 # additional files/directories
 glove.6B.50d.txt
+glove.6B.50d.txt*tmp
 par_vec_trained_400.pkl.docvecs.vectors_docs.npy
 data/data/
 data/work/
@@ -107,6 +108,7 @@ venv*/
 ENV/
 env.bak/
 venv.bak/
+.conda37
 
 # IDE
 .spyderproject
@@ -129,4 +131,5 @@ venv.bak/
 tmp
 wheel
 /sherlock/features/*.npy
+/sherlock/features/*.npy*tmp
 *.csv

requirements.txt

@@ -8,7 +8,7 @@ chardet==3.0.4
 docutils==0.14
 gast==0.2.2
 gdown==4.3.0
-gensim==3.8.0
+gensim==3.8.3
 google-pasta==0.1.7
 grpcio==1.22.0
 h5py==2.9.0
@@ -36,7 +36,7 @@ requests==2.22.0
 s3transfer==0.2.1
 scikit-learn==0.24.2
 six==1.12.0
-smart-open==1.8.4
+smart-open==1.11.1
 scipy==1.5.4
 setuptools~=47.1.0
 tensorboard==1.15.0

sherlock/features/paragraph_vectors.py

@@ -10,7 +10,7 @@
 from datetime import datetime
 from sherlock.global_state import is_first
 
-assert gensim.models.doc2vec.FAST_VERSION > -1, "This will be painfully slow otherwise"
+assert gensim.models.word2vec_inner.FAST_VERSION > -1, "This will be painfully slow otherwise"
 
 
 def tokenise(values):