How dvpombo/solete works
SOLETE combines a version-4 dataset cleaning and quality-control pipeline with a time-series forecasting platform and reproducible benchmarks. The data-preparation path cleans raw measurements, resamples and exports them; the forecasting path loads SOLETE data, prepares features, trains models and evaluates predictions. The README documents the benchmark and data-quality/diagnostic capabilities as additional branches. Some subsystem links rely on the SOURCE INDEX or README because their source was not excerpted; uncertain implementation details are omitted.
Main entry points: clean_solete_1sec.py 1-sec cleaning · MLForecasting.py Forecast quickstart · io.py Data loading and expansion · preprocessing.py Feature preparation
Components 21
Dataset pipeline
- 1-sec cleaning
dataset/pipeline/clean_solete_1sec.py - Cleaning QC rules
dataset/pipeline/qc_flags.py - Solar position
dataset/pipeline/solar_position.py - Resolution resampling
dataset/pipeline/resample_solete.py - Parquet export
dataset/pipeline/export_parquet.py - Pipeline reports
dataset/pipeline/solete_report.py - Data diagnostics
dataset/diagnostics
Forecast platform
- Forecast quickstart
scripts/quickstart/MLForecasting.py - Data loading and expansion
solete/io.py - Feature preparation
solete/preprocessing.py - Forecast models
solete/modeling.py - Forecast metrics
solete/metrics.py - Forecast postprocessing
solete/postprocess.py
Shared utilities
- Data path resolution
solete/paths.py - Platform QC
solete/qc.py - PV physics
solete/physics.py
Analysis and benchmarks
- Dataset inspection
scripts/inspect_dataset.py - Forecast benchmarks
benchmarks
Other components
- Researcher
- Figshare dataset
- Raw and cleaned data
data/hdf5
Connections 28
- Researcher → Figshare dataset: downloads data
README.md - Figshare dataset → Raw and cleaned data: provides files
README.md - Researcher → 1-sec cleaning: runs cleaning
README.md - 1-sec cleaning → Cleaning QC rules: uses QC rules
dataset/pipeline/clean_solete_1sec.py - 1-sec cleaning → Solar position: computes solar position
dataset/pipeline/clean_solete_1sec.py - 1-sec cleaning → Pipeline reports: prints report
dataset/pipeline/clean_solete_1sec.py - 1-sec cleaning → Raw and cleaned data: reads and writes
README.md - Researcher → Resolution resampling: runs resampling
README.md - Resolution resampling → Raw and cleaned data: reads and writes
README.md - Resolution resampling → Pipeline reports: uses report helper
dataset/pipeline/resample_solete.py - Researcher → Parquet export: runs export
README.md - Parquet export → Raw and cleaned data: exports files
README.md - Researcher → Data diagnostics: investigates quality
README.md - Researcher → Forecast quickstart: configures forecasts
README.md - Forecast quickstart → Data loading and expansion: loads data
scripts/quickstart/MLForecasting.py - Forecast quickstart → Feature preparation: prepares features
scripts/quickstart/MLForecasting.py - Forecast quickstart → Forecast models: trains models
scripts/quickstart/MLForecasting.py - Forecast quickstart → Forecast postprocessing: processes predictions
scripts/quickstart/MLForecasting.py - Data loading and expansion → Data path resolution: resolves files
solete/io.py - Data loading and expansion → Feature preparation: expands data
solete/io.py - Data loading and expansion → Platform QC: applies QC
solete/io.py - Data loading and expansion → Forecast postprocessing: uses helpers
solete/io.py - Feature preparation → PV physics: uses physics features
solete/preprocessing.py - Feature preparation → Platform QC: uses QC precedence
solete/preprocessing.py - Forecast postprocessing → Data path resolution: writes outputs
solete/postprocess.py - Researcher → Dataset inspection: inspects dataset
README.md - Dataset inspection → Data path resolution: resolves data files
scripts/inspect_dataset.py - Researcher → Forecast benchmarks: reproduces benchmarks
README.md