[rusetiq]

Summary

Fixes #312.

When a user ran fenn pull <bad-name> ... for a template that does not exist, the CLI suggested fenn pull --list. That flag does not exist on the pull subcommand. The correct command for listing templates is fenn list.

Changes

• fenn/cli/pull.py: point the template-not-found hint to fenn list.
• tests/unit/cli/test_pull_command.py: add regression assertions for the corrected hint.
• fenn/experimental/vision/normalize.py: replace 12 equivalent NumPy bound-method reductions (array.max() / array.min()) with np.max(array) / np.min(array). These mechanical substitutions resolve the ty diagnostics that caused the PR pipeline to fail; no normalization behavior is intentionally changed.

Scope

The PR is limited to 3 files with 17 additions and 13 deletions. Unrelated dashboard, template-test, normalization-validation, warning-handling, and lockfile changes have been removed.

Verification

• pre-commit run --all-files: passed
• Targeted CLI and normalization tests: 53 passed
• Configured Nox unit session: 918 passed, 7 optional-dependency skips

Reproduction

Run fenn pull DOES_NOT_EXIST temp.

Expected: the response suggests fenn list.

Previously observed: the response suggested the nonexistent fenn pull --list command.

[ApusBerliozi]

Hey @rusetiq! Thx 4 your work

Kindly check why pipeline fails

[rusetiq]

sure! give me a while, i'll work on it and commit

[rusetiq]

im pretty sure thatd fix it!
i changed five np calls in fenn/experimental/vision/normalize.py from instance methods like:
alpha_channel.max() to np.max(alpha_channel)
and similarly changed:
rgb_channels_float.max() to np.max(rgb_channels_float)

[rusetiq]

@ApusBerliozi fixed!

[ApusBerliozi]

Hey @rusetiq! Thx

I think, there are way to more changes in files for such an error. I will comeback with full report once I have some free time

[rusetiq]

Thanks for flagging the scope issue. I audited and reduced the PR from 12 files (+709/-144) to 3 files (+17/-13).

Removed from this PR:

• dashboard runner changes
• template-test redesign
• normalization behavior/edge-case changes
• image-batch and Ruff cleanup
• uv.lock regeneration

The branch now contains only the original CLI fix and regression assertions, plus the 12 mechanical ndarray.min/max to np.min/max substitutions required to clear the pipeline's ty diagnostics. Pre-commit passes, the targeted suite reports 53 passed, and the configured Nox unit session reports 918 passed with 7 optional-dependency skips.

Updated in commit 57a1b00.

[rusetiq]

@ApusBerliozi mind checking?

[ApusBerliozi]

Could you please elaborate, why we need to add changes here?

[rusetiq]

Hey @ApusBerliozi, sorry for the long delay getting back to you, and thanks for reviewing!

I included the changes in normalize.py because the pipeline was failing on ty diagnostics for the NumPy .max() / .min() method calls. Switching them to np.max(...) / np.min(...) cleared those diagnostics without intentionally changing the normalization behavior.

They aren't needed for the CLI hint fix itself; I added them to get the checks passing. I understand the concern about keeping this PR focused and am happy to move them into a separate PR if you'd prefer. Thanks for your patience!

[ApusBerliozi]

Hey @rusetiq!

Sorry 4 taking so long, was busy. Kindly check review comment