[Bug] Server CI mistakenly infers presence of failing keymap based on .py file.

[mmccoyd]

Describe the Bug

Server CI mistakenly infers presence of a failing keymap in keyboards/foo/keymaps/ from existence of a .py file.

This was triggered in #18751 by presence of an existing pretty print script keyboards/hillside/48/keymaps/json2hill48.py. That file, attached, is now removed in that PR, but adding it back should trigger the issue.
json2hill48.py.zip

Quoting from sigprof's analysis in the above PR:
"""
The problem is this part of the CI code:
https://github.com/qmk/qmk_ci_executor/blob/88605f9607f8554535c8b8d168f9797ce950b3a0/runner/app/ci_worker.py#L288-L300
together with the presence of keyboards/hillside/48/keymaps/json2hill48.py in this PR.

The CI code mistakenly parses the keyboards/hillside/48/keymaps/json2hill48.py path and infers keymap="keymaps", possible_keyboard="hillside" from it, then that nonexistent keymap gets added for all hillside/* keyboards.
"""

Keyboard Used

n/a

Link to product page (if applicable)

No response

Operating System

n/a

qmk doctor Output

n/a

Is AutoHotKey / Karabiner installed

• AutoHotKey (Windows)
• Karabiner (macOS)

Other keyboard-related software installed

n/a

Additional Context

above

[drashna]

@tzarc