Fix: PandasReporter with droplevels keeps one belief per event

[Flix6x]

Description

Closes #2594.

A PandasReporter with droplevels: true drops the source level of its inputs and asserts one row per event,
so it failed with AssertionError: BeliefDataframe has more than one row per event. as soon as an input event was reported by more than one data source.
This showed up when porting a reporting job to an automation gave its results a new data source, and a downstream job reporting on a rolling window saw both sources until the window moved on.

• With droplevels, each input is now searched with one_deterministic_belief_per_event=True, as AggregatorReporter and ProfitOrLossReporter already do.
• This is a default, set with setdefault: an input that sets one_deterministic_belief_per_event itself keeps its value, instead of the keyword reaching search_beliefs twice (TypeError).
• Configurations that work today are unaffected, because the assertion only passed where each event already had one deterministic belief, which the flag leaves as is.
• The assertion stays, as a safety net.

How to test

pytest flexmeasures/data/models/reporting/tests/test_pandas_reporter.py -k droplevels

• test_pandas_reporter_droplevels_with_several_sources uses the existing fixture data in which two sources report the same noon event. It fails on main with the AssertionError above.
• test_pandas_reporter_droplevels_respects_input_override checks that an input's own one_deterministic_belief_per_event: false is respected. It fails with a TypeError if the flag is passed as a fixed keyword instead of a default.

The whole reporting test package and the report CLI tests pass.

🤖 Generated with Claude Code

[read-the-docs-community]

Documentation build overview

📚 flexmeasures | 🛠️ Build #34754501 | 📁 Comparing bb9fc28 against latest (6dd0460)

  🔍 Preview build  

2 files changed

± changelog.html
± api/v3_0.html

[Copilot]

Copilot review overview

🟡 Changes recommended

The newly added tests introduce unused local variables (likely flake8 F841) that should be cleaned up to avoid lint failures.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

• Unused variables in test tuple unpacking · New

What changed in this PR

This PR fixes PandasReporter failures when droplevels: true is used and multiple data sources report the same event, by defaulting input searches to one_deterministic_belief_per_event=True (while still allowing per-input overrides). It adds regression tests covering the multi-source overlap case and the override behavior, plus a user-facing changelog entry.

Changes:

• Default one_deterministic_belief_per_event to droplevels via setdefault in PandasReporter.fetch_data.
• Add tests ensuring droplevels keeps one belief per event when multiple sources overlap, and that explicit per-input overrides are respected.
• Add a main changelog entry describing the bugfix.

File	Description
flexmeasures/​data/​models/​reporting/​pandas_reporter.py	Default deterministic belief selection when droplevels is enabled to prevent duplicate events after dropping index levels
flexmeasures/​data/​models/​reporting/​tests/​test_pandas_reporter.py	Add regression tests for multi-source overlap under droplevels and for respecting per-input overrides
documentation/​changelog.rst	Document the bugfix in the main changelog

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Copilot review overview

🟢 Approval recommended

The change is narrowly scoped, aligns with existing reporter behavior, and includes targeted regression tests plus a changelog entry.

Review effort: Lite
Findings: None

Resolved since last review (1)

• Unused variables in test tuple unpacking