Merge branch 'master' of github.com:mwmbl/mwmbl

Dockerfile

@@ -26,8 +26,11 @@ COPY mwmbl /app/mwmbl
 # Use pip to install the mwmbl python package
 # PEP 518, PEP 517 and others have allowed for a standardized python packaging API, which allows
 # pip to be able to install poetry packages.
-RUN /venv/bin/pip install pip --upgrade && \
-    /venv/bin/pip install .
+# en-core-web-sm requires a compatible version of spacy
+RUN /venv/bin/pip install pip wheel --upgrade && \
+    /venv/bin/pip install . && \
+    /venv/bin/python -m spacy download en_core_web_sm-3.2.0 --direct && \
+    /venv/bin/python -m spacy validate
 
 FROM base as final
 

pyproject.toml

@@ -31,9 +31,6 @@ langdetect = {version= "==1.0.9", optional = true}
 pyarrow = {version= "==6.0.0", optional = true}
 pyspark = {version= "==3.2.0", optional = true}
 Levenshtein = {version= "==0.16.0", optional = true}
-# en-core-web-sm requires a compatible version of spacy
-en-core-web-sm = {url = "https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.2.0/en_core_web_sm-3.2.0.tar.gz"}
-
 
 [tool.poetry.extras]
 indexer = [
@@ -47,7 +44,6 @@ indexer = [
     "pyarrow",
     "pyspark",
     "Levenshtein",
-    # en-core-web-sm requires a compatible version of spacy
 ]
 
 [tool.poetry.dev-dependencies]