[Bug] docker-compose-transformers.yml: NLP_CONF_FILE set as build ARG but missing from runtime environment

[sriram687]

Summary

docker-compose-transformers.yml passes NLP_CONF_FILE only as a Docker
build-time ARG but omits it from the container's runtime environment: block.

Since app.py reads this value via os.environ.get("NLP_CONF_FILE") at
startup, the build ARG alone is not sufficient to guarantee the correct
NLP engine config is loaded at runtime — especially when the image is run
directly or in environments where build ENVs are not forwarded.

Change

Added NLP_CONF_FILE to the environment: section of presidio-analyzer
in docker-compose-transformers.yml so the build intent is explicitly
mirrored at runtime.

File changed: docker-compose-transformers.yml

     environment:
       - PORT=5001
+      - NLP_CONF_FILE=presidio_analyzer/conf/transformers.yaml

Why this matters

• app.py reads NLP_CONF_FILE via os.environ.get() at runtime
• Without this line, the env var is absent at runtime in override/run scenarios
• Makes the config explicit and overridable without rebuilding the image

Related

• Issue Docker support for Stanza NLP Engine #1028 — Docker support for Stanza NLP Engine
• Issue Can't configure analyzer to use the StanzaRecognizer in docker #1641 — Can't configure StanzaRecognizer in Docker