FlexMeasures / flexmeasures / 21801020926 / 2

78%

main: 78%

Pull #1953

github

Fix parameter cleaning after Marshmallow kebab-case migration (#1961)

* Initial plan

* Add unified copilot environment setup file

Co-authored-by: Flix6x <30658763+Flix6x@users.noreply.github.com>

* docs: add changelog entries for bug fix and environment setup

Co-authored-by: Flix6x <30658763+Flix6x@users.noreply.github.com>

* fix: include last forecast event in get_forecast query

Context:
- Test test_trigger_and_fetch_forecasts was failing with 'P%P' instead of 'PT2H'
- Root cause: event_ends_before excluded the last event from the results
- When forecast period is short, this results in empty/incomplete data

Change:
- Add sensor.event_resolution to event_ends_before parameter
- Ensures all expected forecast events are included in the query
- Matches pattern already used correctly in the test code

Co-authored-by: Flix6x <30658763+Flix6x@users.noreply.github.com>

* revert: remove unnecessary event_ends_before adjustment in forecast API

Context:
- Previous fix added '+ sensor.event_resolution' to event_ends_before parameter
- Investigation revealed this was unnecessary - API behavior was correct
- event_ends_before is INCLUSIVE, so end_date correctly includes events ending at that time

Analysis:
- API correctly returns events with event_end = end_date
- For end_date=02:00 and 1h resolution, events 00:00-01:00 and 01:00-02:00 are correct
- Last event ends exactly at end_date (02:00), which is the expected behavior

Change:
- Reverted event_ends_before back to job.meta.get("end_date")
- Removed the '+ sensor.event_resolution' adjustment

See /tmp/DEBUGGING_FINDINGS.md for full investigation details

* fix(tests): test_forecasting_api now queries correct sensor

Context:
- Test was failing because it queried sensor_1 (solar-sensor-1)
- But API writes forecasts to sensor_0 (solar-sensor, the trigger sensor)
- This caused mismatch: API returned data from sensor_0, test expected data from sensor_1

Root Cause:
- Test used sensor_1 for manual pipel... (continued)

1621 of 2144 branches covered (75.61%)

Branch coverage included in aggregate %.

22396 of 28706 relevant lines covered (78.02%)

0.78 hits per line