CONTRIBUTING.md

Development with Docker

To make it easier to contribute to PipelineDP, we've included a Dockerfile that can be used to run PipelineDP inside a Docker container.

First, fork the repo to your github account.

Then, clone the repo on your computer:

git clone https://github.com/HERE-IS-YOUR-ID/PipelineDP

Go to the project folder:

cd PipelineDP

Build the image using the Dockerfile provided in the contributing folder:

docker build -t devpipelinedp:latest -f contributing/Dockerfile . 

Now every time you want to develop, you have to:

1. Open the project folder (PipelineDP) in your favorite IDE as usual
2. Open the terminal and run:

(Windows): docker run -it --rm -v ${pwd}:/code -w /code devpipelinedp:latest zsh
(Unix): docker run -it --rm -v $(pwd):/code -w /code devpipelinedp:latest zsh

3. Edit your files in your IDE as usual: the local changes will be reflected in the container
4. Run commands in the terminal opened, to have them run in the container

Note that the docker run command above:

• Runs a container with a mapped volume with read/write permission on the project folder: so it can access the local/host project files and have the local changes reflected in the container.
• Removes the container every time you close the terminal (which is fine, since the changes remain in the local files)

Running end-to-end example

When developing it is convenient to run an end-to-end example.

The example below uses a sample from the combined_data_1.txt in the Netflix prize dataset.

To run the example, in the terminal:

python examples/movie_view_ratings/run_all_frameworks.py \
       --input_file=contributing/sample_combined_data_1.txt \
       --output_file=result_example.txt

And see the results by looking at the file result_example.txt.

Style guide

We use the Google Python Style Guide.

Submitting changes

Before submitting your changes, please make sure to auto-format, lint check, and run tests, by running:

make precommit

Individual targets are format, lint, test, clean, dev.