dream2nix python package template (w/ optional PyCharm and direnv features)

I've created this template to create an isolated enviroment for developing python projects using Pycharm (Community Edition) from JetBrains.

Features:

• Uses dream2nix, to maximize reproduciblity of python packaging
  • Based on the example in examples/packages/single-language/python-project.
  • Modified to move most parameter configuration to pyproject.toml
• Published as flake
• (optional, checkout branch direnv) Added .envrc to automatically load shell environment using direnv
• (optional, checkout branch pycharm) Small bashscript to load PyCharm from the terminal, while suppressing stdout

How to use:

Initial set-up

1. Fork this repo and git checkout
2. cd into the folder
3. Modify pyproject.toml according to your project
4. Enable environment $ nix run --impure to start shell environment (will build all dependencies on first run and after any updates)
5. (or, optionally, instead of step 4.) Enable direnv by running direnv allow and auto-enable shell environment and update
6. Update the lock.json by running the command generated by dream2nix (see terminal output, bash -c /nix/... or something similar)
7. (optionally, to use with PyCharm)
   1. (prerequisite: pkgs.jetbrains.pycharm-community is installed in global package space (e.g in configuration.nix or via home-manager))
   2. From the terminal with the environment active, run where python and copy the path to the python bin
   3. un $ ./pycharm.sh (ensure it's executable by $ chmod +x pycharm.sh)
   4. (Close terminal, if you want)
   5. In PyCharm, add the new interpreter:
      • Settings > Add New Interpreter > Virtual Environment
      • Check Existing
      • Paste the copied path from step 7.2

Adding packages (⚠️ cumbersome, WIP 🚧)

Currently, pycharm needs to restarted after installing new python packages. I'm looking into ways to fix this, might be (very) complicated/impossible.

1. Modify `pyproject.toml
2. Re-enable environment (see step 4/5 above)
3. Update lock.json (see above)
4. Save your work and close PyCharm
5. Re-run PyCharm from the terminal

Please let me know in case something is not clear, doesn't work as described or can be improved.