Fix chart interpolation for rows mixing real sensors with flex-config constants

[sankalpsthakur]

Description

• Fix create_line_layer in flexmeasures/data/models/charts/belief_charts.py so that a chart row's interpolation mode (linear vs step-after) is decided only by the real sensors in that row, not by fixed-value/constant sensors sourced from flex-model / flex-context (e.g. soc-min, soc-max, capacities). Fixed-value sensors always report event_resolution == timedelta(0) by construction (they're synthetic, not backed by timed data), so a row mixing a real, non-instantaneous sensor with any such constant was incorrectly forced to linear interpolation instead of the correct step-after.
• Added changelog item in documentation/changelog.rst

Look & Feel

No screenshot from this environment — the change affects the Vega-Lite mark.interpolate value driving line rendering (straight-line "linear" vs staircase "step-after"), which is exactly what the two new regression tests assert on directly. The original bug report's before/after screenshots show a real (should be step-after) sensor's line rendering as smoothly interpolated/linear whenever it shares a row with a flex-model/flex-context constant.

How to test

Run the two new tests in flexmeasures/data/tests/test_belief_charts.py:

pytest flexmeasures/data/tests/test_belief_charts.py -k "test_mixed_row_with_fixed_value_sensor_keeps_step_after or test_real_instantaneous_sensor_still_uses_linear" -q

• test_mixed_row_with_fixed_value_sensor_keeps_step_after is the regression case: a real 15-min-resolution sensor sharing a row with a soc-min fixed-value constant must produce step-after. Fails on main (asserts linear == step-after → AssertionError), passes with this fix.
• test_real_instantaneous_sensor_still_uses_linear is a guard case proving a real instantaneous sensor alone still correctly gets linear.

Full file for no-regression check: pytest flexmeasures/data/tests/test_belief_charts.py -q (33 passed).

Further Improvements

None identified — this is a targeted fix to the interpolation-selection predicate.

Related Items

Fixes #2253

---

AI/LLM disclosure

• AI coding tools (Claude Code, Anthropic) were used to help draft/modify this code, test, changelog entry, and this PR description.
• I reviewed the complete change, understand the reasoning, and ran the reported local tests before submitting.
• This submission is original work of authorship under the project CLA / contributor terms; AI output was not pasted unreviewed.

---

Sign-off

• I agree to contribute to the project under Apache 2 License.
• To the best of my knowledge, the proposed patch is not based on code under GPL or other license that is incompatible with FlexMeasures

[read-the-docs-community]

Documentation build overview

📚 flexmeasures | 🛠️ Build #33831093 | 📁 Comparing 0ae1fec against latest (5bac1ab)

  🔍 Preview build  

31 files changed · + 1 added · ± 30 modified

+ Added

• _autodoc/OPERATION_MODES.html

± Modified

• changelog.html
• genindex.html
• _autosummary/flexmeasures.api.common.responses.html
• _autosummary/flexmeasures.api.common.schemas.generic_schemas.html
• _autosummary/flexmeasures.api.dev.sensors.html
• _autosummary/flexmeasures.api.v3_0.assets.html
• _autosummary/flexmeasures.api.v3_0.jobs.html
• _autosummary/flexmeasures.api.v3_0.sensors.html
• _autosummary/flexmeasures.data.models.forecasting.pipelines.base.html
• _autosummary/flexmeasures.data.models.forecasting.pipelines.predict.html
• and 20 more...

[Ahmad-Wahid]

@sankalpsthakur could you share some screenshots after the fixes?

[sankalpsthakur]

attached before/after — rendered straight from create_line_layer's real mark.interpolate output (linear vs step-after) using the same setup as the regression test's battery_with_interval_sensor_and_soc_min fixture, with representative sample values standing in for a live sensor since standing up a seeded deployment wasn't practical here.

before (main): soc-min forces the whole row to linear

after (this PR): real 15-min sensor correctly keeps step-after