feat(cli): Add post-processing steps to optimize the bundle size

[ryanking13]

This adds a post-processing step after installing python packages that reduces the size of the python_modules directory by applying several optimizers.

By default, following optimizers are on:

• remove_docstrings: cleans up all the docstrings
• remove_comments: cleans up all the comments
• minify_whitespace: 4 spaces => 1 space
• remove_pycache: Removes all pycache directory

while users can addtionally turn on other optimizers if they face size limit issues.

These optimizers can be configured by the pyproject.toml file:

[tool.pywrangler.optimize]
remove_docstrings = false
remove_comments = true
...

[ryanking13]

A very simple worker that uses pandas library gives the following size difference (uncompressed)

before

Attaching additional modules:
┌──────────────────────┬────────┬──────────────┐
│ Name                 │ Type   │ Size         │
├──────────────────────┼────────┼──────────────┤
│ submodule.py         │ python │ 0.05 KiB     │
├──────────────────────┼────────┼──────────────┤
│ Vendored Modules     │        │ 36097.02 KiB │
├──────────────────────┼────────┼──────────────┤
│ Total (2083 modules) │        │ 36097.07 KiB │
└──────────────────────┴────────┴──────────────┘

after

Attaching additional modules:
┌──────────────────────┬────────┬──────────────┐
│ Name                 │ Type   │ Size         │
├──────────────────────┼────────┼──────────────┤
│ submodule.py         │ python │ 0.05 KiB     │
├──────────────────────┼────────┼──────────────┤
│ Vendored Modules     │        │ 29963.09 KiB │
├──────────────────────┼────────┼──────────────┤
│ Total (2083 modules) │        │ 29963.14 KiB │
└──────────────────────┴────────┴──────────────┘

[ask-bonk]

I'm Bonk, and I've done a quick review of your PR.

This PR adds a post-processing optimization step using wheel-optimizer to reduce python_modules bundle size, with sensible defaults and user-configurable overrides via pyproject.toml.

Issues

1. Unpinned wheel-optimizer dependency (medium): Every other dependency in pyproject.toml has version constraints (e.g., click>=8.0.0,<9.0.0). Leaving wheel-optimizer unpinned risks breakage if a future release changes the OptimizerConfig/OptimizerPipeline API.

2. Integration test assumes py_files is non-empty (low): py_files[0] will raise IndexError if no .py files exceed 100 bytes, producing a confusing traceback instead of a clear test failure.

[ask-bonk]

Review posted successfully on PR #91 with two inline suggestion comments:

1. Unpinned wheel-optimizer dependency (packages/cli/pyproject.toml:22) -- suggested adding version bounds consistent with the rest of the project's dependencies.

2. Integration test IndexError risk (packages/cli/tests/test_optimize.py:229) -- suggested adding a guard assertion before indexing py_files[0] to produce a clear failure message instead of a cryptic traceback.

github run