fix: filter user-provided clustering weights to available columns

[FBumann]

Summary

• When users pass a ClusterConfig with explicit weights to cluster(), those weights may reference time series that were removed before clustering (e.g., constant arrays dropped by drop_constant_arrays)
• tsam raises errors when weights contain keys not present in the DataFrame
• Now both user-provided and auto-calculated weights are filtered through the same available_columns parameter in _build_cluster_config_with_weights

Test plan

• All 245 existing clustering tests pass
• Manual test: pass ClusterConfig(weights={...}) with extra keys to cluster() on a multi-period system

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Clustering now tolerates weight settings that reference missing or dropped columns—weights are automatically ignored for unavailable features so clustering runs reliably.
• Tests
  • Added integration tests covering single-period, dropped-column, and multi-period clustering to ensure stable behavior and correct cluster counts.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: e7a72982-0844-44f1-a98d-f024514cc8df

📥 Commits

Reviewing files that changed from the base of the PR and between dc5a220 and 445a478.

📒 Files selected for processing (1)

• tests/test_clustering/test_integration.py

---

📝 Walkthrough

Walkthrough

Refactors clustering weight resolution by adding an optional available_columns parameter to _build_cluster_config_with_weights(), so resolved weights (user-provided or auto) are filtered inside the helper. Removes local filtered_weights from cluster() and centralizes filtering logic.

Changes

Cohort / File(s)	Summary
Weight filtering centralization flixopt/transform_accessor.py	Added optional available_columns parameter to _build_cluster_config_with_weights(); resolve weights by priority (user cluster.weights or auto_weights), then filter to available columns and apply consistently for cluster present or absent. Removed local filtered_weights construction from cluster().
Clustering integration tests tests/test_clustering/test_integration.py	Added three integration tests (TestClusterAdvancedOptions) ensuring ClusterConfig(weights=...) tolerates keys absent from clustering input across single-period, constant-column, and multi-period scenarios; asserts cluster counts and period-dimension behavior.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• fix: filter user-provided clustering weights to available columns #625 — Same change to _build_cluster_config_with_weights centralizing weight filtering logic; directly related.

Poem

🐰
I hopped through code with careful paws,
Gathered weights and trimmed the claws,
Now columns pruned in one small nook,
No stray keys slipping past the hook,
A playful refactor—done! 🥕

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the main change: filtering user-provided clustering weights to match available columns, which is the core fix addressed in this PR.
Description check	✅ Passed	The description covers the problem, solution, and test status, but is missing key template sections like 'Type of Change' and 'Related Issues'.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/filter-user-clustering-weights

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (1)

flixopt/transform_accessor.py (1)

1776-1780: Please add a regression test for extra user weight keys.

This call path is the right place for the fix; adding one automated test for ClusterConfig(weights=...) containing keys absent from df_for_clustering.columns (especially with constant-array dropping and multi-slice clustering) would lock the behavior in.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@flixopt/transform_accessor.py` around lines 1776 - 1780, Add a regression
test that ensures extra keys in ClusterConfig(weights=...) that are not present
in df_for_clustering.columns are ignored/filtered out when building
cluster_config via _calculate_clustering_weights and
_build_cluster_config_with_weights; specifically construct a test that passes a
ClusterConfig with extra weight keys, exercise the same call path around
df_for_clustering used in transform_accessor (including a case where some
columns are constant and dropped, and a multi-slice clustering scenario) and
assert the resulting cluster_config only contains weights for
available/non-constant columns and that no exception is raised for unknown keys.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@flixopt/transform_accessor.py`:
- Around line 1776-1780: Add a regression test that ensures extra keys in
ClusterConfig(weights=...) that are not present in df_for_clustering.columns are
ignored/filtered out when building cluster_config via
_calculate_clustering_weights and _build_cluster_config_with_weights;
specifically construct a test that passes a ClusterConfig with extra weight
keys, exercise the same call path around df_for_clustering used in
transform_accessor (including a case where some columns are constant and
dropped, and a multi-slice clustering scenario) and assert the resulting
cluster_config only contains weights for available/non-constant columns and that
no exception is raised for unknown keys.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: a5830606-0044-4d77-b928-99e07c45e14b

📥 Commits

Reviewing files that changed from the base of the PR and between 1615ee8 and dc5a220.

📒 Files selected for processing (1)

• flixopt/transform_accessor.py