How duck091256/honda-power-reconstruction works
Honda is a local Spark-based workflow for recovering and evaluating active power across gaps in meter readings, with a second bounded-scaling workflow and a Streamlit results demo. Users supply source data, create validated panels, select masked evaluation cases, score imputations, and inspect or compare exported results. The README documents the T02–T15 run sequence; sampled code confirms key panel, evaluation, scoring, scaling, and artifact-verification responsibilities. Some task internals were not sampled, so those subsystems and their links are represented only at documented granularity.
Main entry points: run.py Run artifacts · scale.py Per-meter processing · panel.py Spark panel · __main__.py Task CLI
Components 19
Source and panel
- Source processing
src/honda/source.py - Fixture input
src/honda/fixture.py - Spark panel
src/honda/panel.py - Source inventory
src/honda/inventory.py
Evaluation and scoring
- Case selection
src/honda/cases.py - Imputation evaluation
src/honda/evaluate.py - Scoring
src/honda/scoring.py - Development runs
src/honda/development.py
Scaling workflow
- Per-meter processing
src/honda/scale.py - Scaled evaluation
src/honda/scale_eval.py - Engine benchmark
src/honda/scale_benchmark.py - Scale report export
src/honda/scale_report.py
Results and handoff
- Demo bundle
src/honda/bundle.py - Streamlit demo
app/main.py - Run comparison
src/honda/reproduce.py
Run coordination
- Task CLI
src/honda/__main__.py - Run artifacts
src/honda/run.py
Other components
- Operator
- Honda source data
Connections 31
- Operator → Task CLI: runs tasks
README.md - Task CLI → Source processing: dispatches tasks
src/honda/__main__.py - Honda source data → Source processing: supplies readings
README.md - Source processing → Spark panel: stages source
src/honda/panel.py - Task CLI → Fixture input: dispatches fixture
src/honda/__main__.py - Fixture input → Spark panel: provides fixture data
src/honda/fixture.py - Spark panel → Run artifacts: writes panel
src/honda/panel.py - Operator → Source inventory: starts inventory
README.md - Honda source data → Source inventory: enumerates files
README.md - Operator → Case selection: starts evaluation
README.md - Spark panel → Case selection: provides panel values
src/honda/cases.py - Case selection → Imputation evaluation: supplies masked cases
src/honda/evaluate.py - Imputation evaluation → Scoring: provides predictions
src/honda/scoring.py - Case selection → Run artifacts: writes evaluation inputs
src/honda/cases.py - Scoring → Run artifacts: writes scores
src/honda/scoring.py - Operator → Development runs: runs pilot tasks
README.md - Development runs → Run artifacts: records run
src/honda/development.py - Operator → Per-meter processing: runs scale tasks
README.md - Source inventory → Per-meter processing: provides source manifest
src/honda/scale.py - Per-meter processing → Spark panel: reuses panel logic
src/honda/scale.py - Per-meter processing → Scaled evaluation: dispatches evaluation
src/honda/scale.py - Per-meter processing → Engine benchmark: dispatches benchmark
src/honda/scale.py - Per-meter processing → Scale report export: dispatches report
src/honda/scale.py - Scaled evaluation → Run artifacts: writes metrics
src/honda/scale_eval.py - Engine benchmark → Run artifacts: writes measurements
src/honda/scale_benchmark.py - Scale report export → Demo bundle: exports bundle
src/honda/scale_report.py - Operator → Streamlit demo: opens demo
README.md - Demo bundle → Streamlit demo: provides demo data
README.md - Operator → Run comparison: compares runs
README.md - Run comparison → Run artifacts: verifies artifacts
src/honda/reproduce.py - Task CLI → Demo bundle: dispatches export
src/honda/__main__.py