Issues with Minified Version of docker images

[Rajat72]

I have a complex FastAPI app, which uses haystack and pytorch under the hood, I was looking for a way to reduce the size of the docker image but when I run the minified version it throws out an error, a lot of modules are supposed to be there are not included in optimized image
@kcq Can you help me with this issue

[kcq]

Need more info here. Where's the app repo?

[Rajat72]

command use:
slim build --target eapl_search:v2 --http-probe-start-wait 60 --include-workdir true --include-path "/venv/lib/python3.10/site-packages:777"

Error in slimmed Image:

Error processing line 1 of /venv/lib/python3.10/site-packages/protobuf-3.20.2-py3.10-nspkg.pth:

  Traceback (most recent call last):
    File "/usr/local/lib/python3.10/site.py", line 186, in addpackage
      exec(line)
    File "<string>", line 1, in <module>
    File "<frozen importlib._bootstrap>", line 568, in module_from_spec
  AttributeError: 'NoneType' object has no attribute 'loader'

Remainder of file ignored
Error processing line 1 of /venv/lib/python3.10/site-packages/protobuf-3.20.2-py3.10-nspkg.pth:

  Traceback (most recent call last):
    File "/usr/local/lib/python3.10/site.py", line 186, in addpackage
      exec(line)
    File "<string>", line 1, in <module>
    File "<frozen importlib._bootstrap>", line 568, in module_from_spec
  AttributeError: 'NoneType' object has no attribute 'loader'

Remainder of file ignored
Traceback (most recent call last):
  File "/eapl_search/app/[api.py](https://www.rediffmail.com/cgi-bin/red.cgi?red=http://api.py&isImage=0&BlockImage=0&rediffng=0&rdf=UHdRMgRpXztTelNqBTdUaAUzA3RdbgNxAzoIdwY5BTpWMQ%3D%3D&rogue=038c71e90d58f909a61ce4f1db5dd4975171f0f8)", line 21, in <module>
    from search import es, get_app_settings, search_driver, search_log_index, setup_app_settings
  File "src/[search.py](https://www.rediffmail.com/cgi-bin/red.cgi?red=http://search.py&isImage=0&BlockImage=0&rediffng=0&rdf=UHdRMgRpXztTelNqBTdUaAUzA3RdbgNxAzoIdwY5BTpWMQ%3D%3D&rogue=9184e773486ebc47ea91ce73bb845b8925134aed)", line 9, in init src.search
  File "/venv/lib/python3.10/site-packages/haystack/__init__.py", line 92, in <module>
    from haystack.nodes.base import BaseComponent
  File "/venv/lib/python3.10/site-packages/haystack/nodes/__init__.py", line 5, in <module>
    from haystack.nodes.answer_generator import BaseGenerator, RAGenerator, Seq2SeqGenerator, OpenAIAnswerGenerator
  File "/venv/lib/python3.10/site-packages/haystack/nodes/answer_generator/__init__.py", line 5, in <module>
    from haystack.nodes.answer_generator.openai import OpenAIAnswerGenerator
  File "/venv/lib/python3.10/site-packages/haystack/nodes/answer_generator/openai.py", line 8, in <module>
    from haystack.nodes.prompt import PromptTemplate
  File "/venv/lib/python3.10/site-packages/haystack/nodes/prompt/__init__.py", line 1, in <module>
    from haystack.nodes.prompt.prompt_node import PromptNode
  File "/venv/lib/python3.10/site-packages/haystack/nodes/prompt/prompt_node.py", line 15, in <module>
    from haystack.nodes.prompt.prompt_model import PromptModel
  File "/venv/lib/python3.10/site-packages/haystack/nodes/prompt/prompt_model.py", line 7, in <module>
    from haystack.nodes.prompt.invocation_layer import PromptModelInvocationLayer
  File "/venv/lib/python3.10/site-packages/haystack/nodes/prompt/invocation_layer/__init__.py", line 1, in <module>
    from haystack.nodes.prompt.invocation_layer.azure_open_ai import AzureOpenAIInvocationLayer
  File "/venv/lib/python3.10/site-packages/haystack/nodes/prompt/invocation_layer/azure_open_ai.py", line 3, in <module>
    from haystack.nodes.prompt.invocation_layer.open_ai import OpenAIInvocationLayer
  File "/venv/lib/python3.10/site-packages/haystack/nodes/prompt/invocation_layer/open_ai.py", line 8, in <module>
    from haystack.utils.openai_utils import (
  File "/venv/lib/python3.10/site-packages/haystack/utils/openai_utils.py", line 36, in <module>
    import tiktoken  # pylint: disable=import-error
  File "/venv/lib/python3.10/site-packages/tiktoken/__init__.py", line 2, in <module>
    from .model import encoding_for_model as encoding_for_model
  File "/venv/lib/python3.10/site-packages/tiktoken/model.py", line 4, in <module>
    from .registry import get_encoding
  File "/venv/lib/python3.10/site-packages/tiktoken/registry.py", line 8, in <module>
    import tiktoken_ext
ModuleNotFoundError: No module named 'tiktoken_ext'

[Rajat72]

Requirements:

git+https://github.com/UmarEm/haystack-emp.git@add_semantic_highlight_emp
fastapi==0.78.0
mkdocs==1.3.0
mkdocs-jupyter==0.20.1
mkdocs-material==8.2.9
mkdocs-material-extensions==1.0.3
mkdocstrings[python]==0.19
python-dotenv==0.20.0
protobuf==3.20.2
uvicorn==0.17.6
fuzzywuzzy==0.18.0
black==23.1.0
flake8==4.0.1
isort==5.10.1
pre-commit==2.18.1
loguru==0.6.0
spacy==3.5.3
elasticsearch==7.17.9
frontend==0.0.3
pymupdf==1.22.3
opensearch-py==2.2.0
SQLAlchemy==2.0.15
selenium==4.1.0

[Rajat72]

Docker File:

# Using Python3.10-slim-bullseye as Base Docker image
FROM python:3.10-slim-bullseye

# Defining Environment Variables
ENV DEBIAN_FRONTEND=noninteractive
ENV PYTHONDONTWRITEBYTECODE 1
ENV PYTHONUNBUFFERED 1
ENV APP_HOME=/eapl_search

# Create Working-Directory
RUN mkdir -p $APP_HOME
WORKDIR $APP_HOME

# System Library Installation
RUN apt update && \
    apt upgrade -y && \
    apt install -y git

# Virtual Environment Creation & requirements.txt Installation
RUN python3 -m venv /venv

# Python Environment Path
ENV VIRTUAL_ENV /venv
ENV PATH /venv/bin:$PATH

# Create the app user called emplay
RUN adduser --disabled-password --gecos '' emplay

# Install python packages
COPY requirements.txt .
RUN pip install --upgrade pip setuptools wheel && \
pip install --no-cache-dir -r requirements.txt

RUN python3 -m spacy download en_core_web_sm

# Copy Repo & Dependencies
COPY . /$APP_HOME

# Remove unwanted packages
RUN rm -rf /var/lib/apt/lists/* && \
apt-get autoremove -y && apt-get autoclean -y && \
rm -rf ~/.cache/pip

# Change the Folder Permissions to emplay user & Log Folder Creation
RUN chown -R emplay:emplay $APP_HOME && \
chmod -R 755 $APP_HOME

# Compile python modules in binary files
ARG build_cython=true

RUN if [ "$build_cython" = "true" ] ; then \
   # System Library Installation
       apt update && apt install gcc g++ --no-install-recommends -y && pip install --upgrade pip cython && python3 setup.py build_ext --inplace --delpy --delc;\
    else \
       echo "Skipping Binary Conversion"; \
    fi


# Change the user to start the app
USER emplay
EXPOSE 9090
ENTRYPOINT ["python3","/eapl_search/app/api.py"]

[Rajat72]

@kcq can you help me with these?