Bump nltk from 3.8.1 to 3.9.1 (#1993)

docker/Dockerfile.baseimage

@@ -11,7 +11,7 @@
 #
 # Main repo for this image is here:
 # https://github.com/phusion/baseimage-docker
-FROM phusion/baseimage:jammy-1.0.1 As baseimage
+FROM phusion/baseimage:jammy-1.0.1 AS baseimage
 
 # Make sure base system is up to date
 RUN apt-get update && \
@@ -63,7 +63,10 @@ RUN python3 -m venv env && \
     . env/bin/activate && \
     pip install --upgrade pip && \
     poetry install --only main,pg --sync --no-root && \
+    # TODO: This can be removed once NLTK merges https://github.com/sloria/TextBlob/pull/469
+    python3 -m nltk.downloader punkt_tab && \
     python3 -m textblob.download_corpora lite && \
+    rm -rf /root/nltk_data/tokenizers/punkt && \
     mv /root/nltk_data /usr/lib/ && \
     find /usr/lib/nltk_data -name *.zip -delete && \
     rm -Rf /root/.cache && \

pyproject.toml

@@ -246,7 +246,6 @@ lxml = {extras = ["html-clean"], version = "^5.2.1"}
 money = "1.3.0"
 multipledispatch = "^1.0"
 nameparser = "^1.1"  # nameparser is for author name manipulations
-nltk = "3.8.1"  # nltk is a textblob dependency.
 opensearch-dsl = "~1.0"
 opensearch-py = "~1.1"
 palace-webpub-manifest-parser = "^4.0.0"

tox.ini

@@ -5,6 +5,8 @@ skipsdist = true
 [testenv]
 commands_pre =
     poetry install --without ci --sync -v
+    # TODO: This can be removed once NLTK merges https://github.com/sloria/TextBlob/pull/469
+    python -m nltk.downloader punkt_tab
     python -m textblob.download_corpora
 commands =
     pytest {posargs:tests}