fix: embed stanza resources in container since free tier has no persistent storage on huggingface spaces

Dockerfile

@@ -6,8 +6,8 @@ COPY ./requirements.txt /code/requirements.txt
 
 RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 
-ENV STANZA_RESOURCES_DIR /data/stanza_resources
-
 ADD . .
 
+RUN python download.py
+
 CMD ["gunicorn", "-k", "sync", "-w", "1", "-b", "0.0.0.0:7860", "main:app"]

download.py

@@ -0,0 +1,3 @@
+import stanza
+
+stanza.download("en", processors="tokenize")

main.py

@@ -2,9 +2,7 @@ from flask import Flask, request, jsonify
 import os
 import stanza
 
-os.makedirs(os.getenv("STANZA_RESOURCES_DIR", "/tmp/stanza_resources"), exist_ok=True)
-
-nlp = stanza.Pipeline("en", processors="tokenize")
+nlp = stanza.Pipeline("en", processors="tokenize", download_method=None)
 
 app = Flask(__name__)