python代码进去docker容器内_在Docker容器中导入python文件

I'm executing a python file inside a docker container, and need to import all the functions that I put into a separate python file called base_functions. However, writing from base_functions import * throws the error, that ModuleNotFoundError: No module named 'base_functions', even though base_functions.py is in the same directory as the main python file. How can I do this? Do I need to specify which python scripts I want to import beforehand, in the settings.ini or something?

This is the content of the Dockerfile:

FROM amancevice/pandas:0.24.1-alpine

RUN apk update

RUN apk add build-base

RUN apk add gcc musl-dev libc-dev util-linux-dev linux-headers python3-dev postgresql-libs postgresql-dev git libffi-dev libmemcached-dev zlib-dev \

ca-certificates zlib-dev jpeg-dev freetype-dev libpng

RUN pip3 install --upgrade pip

COPY requirements.txt .

RUN pip3 install -r requirements.txt

COPY src /vdp

WORKDIR /

ENTRYPOINT ["python3", "-m", "vdp"]

These are all the files in the directory of the project:

/home/cr/docker/71119/.

├── docker-compose.yaml

├── Dockerfile

├── Makefile

├── README.md

├── requirements.txt

├── settings.ini

└── src

├── base_functions.py

├── influx.py

├── __init__.py

└── __main__.py

解决方案

You're practically renaming your package (it smells like a package, having __init__.py and __main__.py) from "src" to "vdp" when you're creating the dockerfile.

I'd recommend:

Rename src/... to src/vdp/... (or vdp/...)

– at this point you should be able to run python -m vdp within the src directory on your host machine and things should work.

Change the Dockerfile stanzas to

COPY src /src

WORKDIR /src

ENTRYPOINT ["python3", "-m", "vdp"]

– this'd be equivalent to how you're running it on the host machine.

Another option would be to drop the /src in the repository and change things to COPY vdp /src/vdp.

The third option, of course, would be to set up proper Python packaging, have your setup.py build a proper wheel, then simply install that in the Docker container.