V1 refactor scadadf

[aclerc]

Fully remove wind-up v0 method data column assumptions from the rest of the benchmarking code.

[Copilot]

Pull request overview

Refactors the benchmarking/synthetic pipeline and harness to operate on source-native SCADA column names (rather than wind-up v0 DataColumns aliases), keeping v0 aliasing confined to the v0 baseline on-ramp.

Changes:

• Introduces a ColumnSchema abstraction and threads it through synthetic generation, plotting, and ground-truth uplift computation.
• Updates the Hill of Towie adapter to load/cache source-native wtc_* tags, reshape wide→long for method-facing data, and adds a v0-only long_to_wind_up_format conversion.
• Makes the harness/method seam carry the turbine identifier column (turbine_col) and updates baselines/tests accordingly (including making NaiveRatioMethod configured by active-power column).

Reviewed changes

Copilot reviewed 30 out of 30 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/benchmarking/synthetic/test_upgrades.py	Updates synthetic upgrade tests to use HOT_COLUMNS source-native schema.
tests/benchmarking/synthetic/test_plots.py	Updates plot tests to use source-native column schema.
tests/benchmarking/synthetic/test_make_example_datasets.py	Updates example dataset tests to use source-native column schema.
tests/benchmarking/synthetic/test_ground_truth.py	Updates ground-truth tests to use source-native column schema.
tests/benchmarking/synthetic/test_generator.py	Updates generator tests to use source-native column schema and turbine id column.
tests/benchmarking/synthetic/sources/test_hill_of_towie.py	Splits tests between source-native reshape and v0 on-ramp conversion.
tests/benchmarking/synthetic/sources/test_hill_of_towie_cache.py	Adds offline tests for per-(year,turbine) parquet caching behavior.
tests/benchmarking/harness/test_scoring.py	Updates harness scoring tests to use HOT_COLUMNS and drop DataColumns.
tests/benchmarking/harness/test_replicates.py	Updates replicate tests to use HOT_COLUMNS and turbine id column.
tests/benchmarking/harness/stubs.py	Updates stub/oracle helpers to use seam-provided turbine column + source-native power.
tests/benchmarking/baselines/test_v0_binned.py	Adjusts v0 baseline tests for source-native input + conversion on-ramp.
tests/benchmarking/baselines/test_toggle_end_to_end.py	Updates end-to-end toggle test wiring for configured naive method + HOT_COLUMNS.
tests/benchmarking/baselines/test_naive_ratio.py	Reworks naive-ratio tests to prove no wind_up imports and use configured columns.
benchmarking/synthetic/upgrades.py	Refactors upgrades to be keyed by ColumnSchema (defaulting to HoT schema).
benchmarking/synthetic/sources/hill_of_towie.py	Loader now keeps wtc_* tags; adds scada_wide_to_long + v0-only long_to_wind_up_format; adds per-turbine-year cache.
benchmarking/synthetic/schema.py	New ColumnSchema dataclass defining semantic roles for source-native columns.
benchmarking/synthetic/plots.py	Plotting now selects columns via ColumnSchema instead of DataColumns.
benchmarking/synthetic/ground_truth.py	Ground-truth uplift now selects columns via ColumnSchema instead of DataColumns.
benchmarking/synthetic/generator.py	Generator now operates on source-native long SCADA and threads ColumnSchema through outputs.
benchmarking/synthetic/init.py	Re-exports ColumnSchema and HOT_COLUMNS from the synthetic package API.
benchmarking/harness/scoring.py	Threads ColumnSchema through replicate-building; injects turbine_col into MethodInput.
benchmarking/harness/replicates.py	Adds columns parameter and subsets by schema’s turbine column.
benchmarking/harness/method.py	Extends MethodInput to carry the turbine-identifier column name.
benchmarking/harness/example_hot_study.py	Updates oracle/example wiring to use source-native HOT_COLUMNS.
benchmarking/baselines/v0_binned.py	Converts source-native long SCADA to wind-up format inside the v0 baseline only.
benchmarking/baselines/naive_ratio.py	Makes naive method source-agnostic via configured active-power column + seam-provided turbine column.
benchmarking/baselines/inspect_v0_run.py	Updates inspection tool to provide turbine_col and use source-native schema.
benchmarking/baselines/inspect_naive.py	Updates inspection tool to configure naive method with source-native power col + turbine_col.
benchmarking/baselines/example_toggle_study.py	Updates example toggle study to configure naive method with source-native power col.
benchmarking/baselines/example_prepost_study.py	Updates example prepost study to configure naive method with source-native power col.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.