ayatelawej/fault-detection-and-reliability-monitoring-for-a-newly-deployed-weather-station-network-in-libya

How ayatelawej/fault-detection-and-reliability-monitoring-for-a-newly-deployed-weather-station-network-in-libya works

This repository implements station-data quality and reliability analytics: it derives rule-based detector evidence and hourly fault models, classifies availability and outage events, scores station health, forecasts health and risk, and presents results through a dashboard. Model training and tuning are a substantial supporting branch. The graph follows relationships named in the SOURCE INDEX or visible in excerpts; unsampled scripts and modules are represented only where their role is clear, and uncertain connections are omitted.

Main entry points: build_availability_events.py Availability events · risk_dataset.py Risk datasets · hourly_detection.py Hourly detection · health_score.py Health scoring

Components 26

Data Quality

Data audit src/features/run_data_audit.py
Row state src/features/row_state.py
Rule evidence features src/features/detection_features.py
Channel handling src/rules/channel_handlers.py
Rule configuration src/rules/config.py

Fault Detection

Hourly detection src/model/hourly_detection.py
Hourly baseline src/model/hourly_baseline.py
Model feature specification src/model/feature_spec.py
Reliability-aware fusion model src/model/hourly_rgfn.py
RGFN training src/model/hourly_rgfn_training.py
RGFN tuning src/model/hourly_rgfn_tuning.py
Hourly tuning workflow src/workflows/tune_hourly_detection.py
RGFN training workflow src/workflows/train_hourly_rgfn.py

Availability & Health

Availability events src/availability/build_availability_events.py
Network outage windows src/availability/build_network_outage_windows.py
Station reliability summary src/availability/build_station_reliability_summary.py
Health scoring src/availability/health_score.py
Operational scorecard src/availability/operational_scorecard.py

Risk Forecasting

Risk datasets src/availability/risk_dataset.py
Risk model src/availability/risk_model.py
Risk evaluation src/availability/risk_eval.py
Health forecasting src/availability/health_forecast.py

Results & Dashboard

Dashboard replay src/dashboard/replay.py
Dashboard presentation src/dashboard/presentation.py
Weather context src/dashboard/weather_context.py

Other components

Analyst / operator

Connections 41

Analyst / operator → Dashboard presentation: views results
Data audit → Row state: uses row states src/features/run_data_audit.py
Rule evidence features → Rule configuration: uses settings src/features/detection_features.py
Rule evidence features → Channel handling: uses channel handling src/features/detection_features.py
Availability events → Row state: uses row states src/availability/build_availability_events.py
Availability events → Channel handling: groups channels src/availability/build_availability_events.py
Availability events → Rule configuration: uses settings src/availability/build_availability_events.py
Availability events → Network outage windows: classifies outages src/availability/build_availability_events.py
Availability events → Station reliability summary: builds summary src/availability/build_availability_events.py
Risk datasets → Row state: uses row states src/availability/risk_dataset.py
Risk datasets → Channel handling: uses channel groups src/availability/risk_dataset.py
Risk datasets → Rule configuration: uses settings src/availability/risk_dataset.py
Risk datasets → Hourly detection: uses detector features src/availability/risk_dataset.py
Risk datasets → Availability events: uses availability src/availability/risk_dataset.py
Health scoring → Availability events: uses events src/availability/health_score.py
Health scoring → Risk datasets: uses risk data src/availability/health_score.py
Risk evaluation → Risk datasets: evaluates datasets src/availability/risk_eval.py
Risk evaluation → Risk model: evaluates model src/availability/risk_eval.py
Risk model → Risk datasets: uses datasets src/availability/risk_model.py
Health forecasting → Availability events: uses events src/availability/health_forecast.py
Health forecasting → Health scoring: uses health scores src/availability/health_forecast.py
Health forecasting → Risk datasets: uses risk data src/availability/health_forecast.py
Hourly detection → Model feature specification: uses feature spec src/model/hourly_detection.py
Hourly detection → Rule evidence features: uses rule labels
Hourly baseline → Hourly detection: uses hourly detection src/model/hourly_baseline.py
Reliability-aware fusion model → Model feature specification: uses feature spec src/model/hourly_rgfn.py
Reliability-aware fusion model → Hourly detection: uses detection features src/model/hourly_rgfn.py
RGFN training → Reliability-aware fusion model: trains model src/model/hourly_rgfn_training.py
Hourly tuning workflow → Hourly baseline: uses baseline src/workflows/tune_hourly_detection.py
Hourly tuning workflow → Hourly detection: uses detector src/workflows/tune_hourly_detection.py
Hourly tuning workflow → RGFN training: runs training src/workflows/tune_hourly_detection.py
Hourly tuning workflow → RGFN tuning: runs tuning src/workflows/tune_hourly_detection.py
RGFN training workflow → Hourly baseline: uses baseline src/workflows/train_hourly_rgfn.py
RGFN training workflow → Hourly detection: uses detector src/workflows/train_hourly_rgfn.py
RGFN training workflow → RGFN training: runs training src/workflows/train_hourly_rgfn.py
Dashboard replay → Availability events: loads events src/dashboard/replay.py
Dashboard replay → Health forecasting: loads forecasts src/dashboard/replay.py
Dashboard replay → Health scoring: loads health scores src/dashboard/replay.py
Dashboard presentation → Dashboard replay: presents replay src/dashboard/presentation.py
Dashboard presentation → Health scoring: presents health src/dashboard/presentation.py
Weather context → Channel handling: context mapping

More Python diagrams: codeofaxel/kiln · commune-ai/commune · cuhk-eda/neural-ilt · evcc-io/optimizer · fredzzhang/pvic

Drawn by GitDiagram on October 7, 2026 from the repository's file tree, README and source files. Read as Markdown ·