How devansh-567/sih-2026 works
RF Signal Analysis Platform is a local-first RF recording analysis backend with a React analyst GUI. The main workflow uploads IQ/WAV or SigMF recordings, ingests samples and metadata, runs staged DSP/ML analysis through demodulation, FEC/interleaving and bitstream hypotheses, then presents evidence-backed results. Analysts can also inspect waterfalls, batch and compare history, give feedback, and export reports. The README documents these capabilities; some subsystem links below rely on its statements or the SOURCE INDEX because their implementations were not excerpted.
Main entry points: stages.py Analysis orchestration · main.py HTTP API · App.tsx Analysis workspace · schema.py Evidence schema
Components 23
Analyst GUI
- Analysis workspace
frontend/src/App.tsx - API client
frontend/src/lib/api.ts - Evidence findings
frontend/src/components/AutomaticAnalysisPanel.tsx - Demod and FEC views
frontend/src/components/DemodPanel.tsx
API and records
- HTTP API
rfplatform/api/main.py - Analysis history
rfplatform/storage/db.py - History comparison
rfplatform/pipeline/compare.py
Signal analysis
- Recording ingestion
rfplatform/io/formats.py - Analysis orchestration
rfplatform/pipeline/stages.py - DSP estimation
rfplatform/dsp/estimators.py - Modulation classifier
rfplatform/dsp/classifier.py - Signal demodulation
rfplatform/dsp/demodulate.py - FEC decoding
rfplatform/fec/convolutional.py - Reed-Solomon
rfplatform/fec/reed_solomon.py - De-interleaving
rfplatform/interleave/interleavers.py - Bitstream framing
rfplatform/pipeline/bitstream.py - Confidence fusion
rfplatform/pipeline/fusion.py - Evidence schema
rfplatform/models/schema.py
ML and synthesis
- ML inference
rfplatform/ml/inference.py - CNN model
rfplatform/ml/model.py - Synthetic signals
rfplatform/synth/generator.py
Results and exports
- Report export
rfplatform/report/generator.py
Other components
- Analyst
Connections 27
- Analyst → Analysis workspace: uses
- Analysis workspace → API client: requests
frontend/src/App.tsx - API client → HTTP API: calls API
- HTTP API → Recording ingestion: loads recordings
rfplatform/api/main.py - HTTP API → Analysis orchestration: runs analysis
rfplatform/api/main.py - Analysis orchestration → DSP estimation: extracts features
rfplatform/pipeline/stages.py - Analysis orchestration → Modulation classifier: classifies modulation
rfplatform/pipeline/stages.py - Analysis orchestration → Signal demodulation: demodulates
rfplatform/pipeline/stages.py - Analysis orchestration → FEC decoding: tests convolutional codes
rfplatform/pipeline/stages.py - Analysis orchestration → Reed-Solomon: tests RS codes
rfplatform/pipeline/stages.py - Analysis orchestration → De-interleaving: tests interleavers
rfplatform/pipeline/stages.py - Analysis orchestration → Bitstream framing: analyzes framing
rfplatform/pipeline/stages.py - Analysis orchestration → ML inference: requests inference
rfplatform/pipeline/stages.py - Analysis orchestration → Confidence fusion: combines evidence
rfplatform/pipeline/stages.py - ML inference → CNN model: runs model
rfplatform/ml/inference.py - ML inference → Evidence schema: returns parameters
rfplatform/ml/inference.py - Confidence fusion → Evidence schema: uses evidence types
rfplatform/pipeline/fusion.py - Analysis orchestration → Evidence schema: builds results
rfplatform/pipeline/stages.py - HTTP API → Analysis history: persists analyses
rfplatform/api/main.py - HTTP API → Report export: exports results
rfplatform/api/main.py - HTTP API → History comparison: compares history
rfplatform/api/main.py - Analysis history → History comparison: provides stored analyses
- Synthetic signals → CNN model: supplies training signals
- Analysis workspace → Evidence findings: renders
frontend/src/App.tsx - Analysis workspace → Demod and FEC views: renders
frontend/src/App.tsx - Evidence findings → Evidence schema: displays parameters
- API client → Report export: requests export
frontend/src/lib/api.ts