Improve Development Accesibility - Devcontainer [part 1 / 2]

[cgarciae]

Whats new

This PR introduces some changes that improve the development experience via VSCode's devcontainers. Specifically:

• It removes the cloning of the pandas repo inside the container (it assumes that the user already cloned it on the host) and instead leverages the fact the the code is mounted to /workspaces/pandas.
• Defines the conda env inside the workspace itself such that it can be reused even when the container is deleted. This enables you to use the codebase outside the repo afterwards if you activate the same local conda env.
• Splits the previous Dockerfile in two: a simplified Dockerfile + a new container_create.sh that only runs the first time the container starts, making it possible to install the dependencies inside the workspace.

[cgarciae]

In the next PR I want to add support for Gitpod which is a nice way to develop when you don't have access to your PC.

[jreback]

can you:

• update the build instructions in the docs reflecting the changes
• can we create some sort of test for this (e.g. have a github action that does this)

[jreback]

add a final eol

[jreback]

cc @fangchenli @datapythonista @WillAyd

[datapythonista]

-1 on adding this file. I think ideally we shouldn't even have vscode stuff in the repo, and we should be agnostic of what editor or IDE devs are using in this repo. I understand that it can be convenient to have a settings file for vscode, and we already do, but adding more vscode scripts to the repo seems excesive. If ci/setup_env.sh can't be used, and this code can't be added directly to the vscode config file, maybe it's a good time to create a separate project for vscode stuff for pandas development.

[WillAyd]

Yea I agree with Mark here. I think adding stuff specific to an IDE is favoring a workflow that we should probably be agnostic to. Maybe better served as a dedicated pandas-vscode repo that just uses pandas as a submodule?

[cgarciae]

Hmmm. The .devcontainer.json is already an editor-specific file so that line has already been crossed.

How about we create a /.devcontainer folder and move both the .devcontianer.json and container_create.sh files into it? This way we don't have these files pollute the workspace.

[github-actions]

This pull request is stale because it has been open for thirty days with no activity. Please update or respond to this comment if you're still interested in working on this.

[datapythonista]

Closing this, since there is no agreement on moving forward, and it didn't have any progress for couple of months.

[cgarciae]

Hey @jreback and @datapythonista!
Sorry for the haitus, I got some time again and will be trying to move this forward.