[Presidio Analyzer Bug] Dockerfile.transformers: Build and startup issues

[gerardduch]

Describe the bug
The current Dockerfile.transformers in the Presidio Analyzer project fails to build and start correctly.

To Reproduce
Steps to reproduce the behavior:

1. Attempt to build the Docker image using the Dockerfile.transformers:

docker build -f Dockerfile.transformers -t presidio-analyzer-transformers .

2. Observe the build errors.

Expected behavior
The Docker image should build successfully and the container should start properly, with Presidio Analyzer running and accessible as expected.

Additional context

• The Dockerfile is missing important environment variables (PORT and WORKERS), which are required for proper configuration and operation.
• The order of COPY commands causes issues.
• The use of poetry install --no-lock should be removed.

[omri374]

thanks, could you please add more information on the error you're getting? also, in which environment are you building the docker?

[gerardduch]

Hello @omri374, thanks for your message.

With the current Dockerfile.transformers, I've the following error:

[...]
27.48   - Installing tldextract (5.3.0)
29.70 
29.70 Writing lock file
29.70 
29.70 Installing the current project: presidio_analyzer (2.2.360)
29.70 
29.70 Error: The current project could not be installed: Readme path `/app/README.md` does not exist.
29.70 If you do not want to install the current project use --no-root.
29.70 If you want to use Poetry only for dependency management but not for packaging, you can disable package mode by setting package-mode = false in your pyproject.toml file.
29.70 If you did intend to install the current project, you may need to set `packages` in your pyproject.toml file.
29.70 
------
Dockerfile.transformers:18
--------------------
  16 |     
  17 |     COPY ./pyproject.toml /app/
  18 | >>> RUN pip install poetry && poetry install -E server -E transformers
  19 |     
  20 |     RUN poetry add torch transformers huggingface_hub --no-lock
--------------------
ERROR: failed to build: failed to solve: process "/bin/sh -c pip install poetry && poetry install -E server -E transformers" did not complete successfully: exit code: 1

View build details: docker-desktop://dashboard/build/desktop-linux/desktop-linux/p4x7z6th65sx19mfi9h4ur6xi

If I move the COPY . /app/ line right before RUN pip install poetry && poetry install -E server -E transformers, the README.md error goes away. However, I then get a different error regarding --no-lock:

 > [ 8/10] RUN poetry add torch transformers huggingface_hub --no-lock:
0.586 
0.586 The option "--no-lock" does not exist
------
Dockerfile.transformers:20
--------------------
  18 |     RUN pip install poetry && poetry install -E server -E transformers
  19 |     
  20 | >>> RUN poetry add torch transformers huggingface_hub --no-lock
  21 |     
  22 |     # install nlp models specified in NLP_CONF_FILE
--------------------
ERROR: failed to build: failed to solve: process "/bin/sh -c poetry add torch transformers huggingface_hub --no-lock" did not complete successfully: exit code: 1

View build details: docker-desktop://dashboard/build/desktop-linux/desktop-linux/ozsh82rjci613zw1zdmg6wdqv

Context:
I'm building this on a Mac M1 Pro, running macOS 26.0.1, using Docker.

Build command:
bash docker build -f Dockerfile.transformers -t presidio-analyzer-transformers .

Let me know if you need more information!

[gerardduch]

Hello @omri374, I managed to get it working using the following Dockerfile:

FROM python:3.12-slim

ARG NLP_CONF_FILE=presidio_analyzer/conf/transformers.yaml
ARG ANALYZER_CONF_FILE=presidio_analyzer/conf/default_analyzer.yaml
ARG RECOGNIZER_REGISTRY_CONF_FILE=presidio_analyzer/conf/default_recognizers.yaml

ENV PIP_NO_CACHE_DIR=1
ENV ANALYZER_CONF_FILE=${ANALYZER_CONF_FILE}
ENV RECOGNIZER_REGISTRY_CONF_FILE=${RECOGNIZER_REGISTRY_CONF_FILE}
ENV NLP_CONF_FILE=${NLP_CONF_FILE}
ENV PORT=3000
ENV WORKERS=1

WORKDIR /app

COPY ${ANALYZER_CONF_FILE} /app/${ANALYZER_CONF_FILE}
COPY ${RECOGNIZER_REGISTRY_CONF_FILE} /app/${RECOGNIZER_REGISTRY_CONF_FILE}
COPY ${NLP_CONF_FILE} /app/${NLP_CONF_FILE}

COPY . /app/
RUN pip install poetry && \
    poetry install -E server -E transformers && \
    poetry run pip install torch transformers huggingface_hub

RUN poetry run python install_nlp_models.py --conf_file ${NLP_CONF_FILE}

EXPOSE ${PORT}
CMD ["./entrypoint.sh"]

However, there’s an issue with the Presidio server. When I run this CURL command:

curl --location 'http://localhost:5002/analyze' \
--header 'Content-Type: application/json' \
--data '{
  "text": "Mi name is John Doe and I live in Barcelona",
  "language": "en"
}
 '

I get the following error:

{"error":"'numpy.float32' object has no attribute 'to_dict'"}

Any advice would be appreciated, maybe I missed something while updating the Dockerfile.

Thanks!