How codingnuz/innovation-g2 works
Innovation-G2 is a spam, malware, and process-detection application: users submit text, files, or process counters through a React interface; FastAPI dispatches requests to the corresponding prediction path and returns results. The Assistant adds tool-driven chat and attachment analysis, while training and data preparation produce the models used at runtime. The graph preserves the documented product branches, but some backend endpoint wiring and unsampled training details are represented only where the README or source index supports them.
Main entry points: App.jsx App shell · client.js API client · predict.py Prediction entry · main.py FastAPI service
Components 24
React interface
- App shell
frontend-react/src/App.jsx - Spam screen
frontend-react/src/views/SpamView.jsx - Malware screen
frontend-react/src/views/MalwareView.jsx - Process screen
frontend-react/src/views/ProcessView.jsx - Assistant chat
frontend-react/src/views/ChatView.jsx - History view
frontend-react/src/views/HistoryView.jsx - API client
frontend-react/src/api/client.js
API and assistant
- FastAPI service
innovation-g2/main.py - Chat agent
innovation-g2/agent.py - Attachment analysis
innovation-g2/attachments.py - Upload scan
innovation-g2/malware_scan.py
Prediction runtime
- Prediction entry
spam_malware_detection_model/predict.py - Spam inference
spam_malware_detection_model/inference/spam.py - Malware inference
spam_malware_detection_model/inference/malware.py - Process inference
spam_malware_detection_model/inference/process.py - Format router
spam_malware_detection_model/malware_engine/inference/router.py
Model development
- Training orchestrator
spam_malware_detection_model/training/run_all.py - Spam training
spam_malware_detection_model/spam/train.py - Tabular malware training
spam_malware_detection_model/malware_tabular/train.py - File model training
spam_malware_detection_model/malware_engine/training/train_pipeline.py - Process training
spam_malware_detection_model/regression/train.py - Saved model bundle
spam_malware_detection_model/config/paths.py
Other components
- User
- OpenAI-compatible server
Connections 35
- User → App shell: uses
- App shell → Spam screen: shows
frontend-react/src/App.jsx - App shell → Malware screen: shows
frontend-react/src/App.jsx - App shell → Process screen: shows
frontend-react/src/App.jsx - App shell → Assistant chat: shows
frontend-react/src/App.jsx - App shell → History view: shows
frontend-react/src/App.jsx - App shell → API client: checks health
frontend-react/src/App.jsx - Spam screen → API client: requests predictions
- Malware screen → API client: requests scans
frontend-react/src/views/MalwareView.jsx - Process screen → API client: requests predictions
- Assistant chat → API client: streams chat
- Malware screen → API client: uploads files
frontend-react/src/views/MalwareView.jsx - API client → FastAPI service: calls API
- FastAPI service → Prediction entry: dispatches predictions
innovation-g2/main.py - FastAPI service → Chat agent: runs chat
innovation-g2/main.py - FastAPI service → Attachment analysis: passes attachments
innovation-g2/main.py - FastAPI service → Upload scan: dispatches uploads
innovation-g2/main.py - Chat agent → Attachment analysis: uses attachment tools
innovation-g2/agent.py - Chat agent → OpenAI-compatible server: requests completions
- Attachment analysis → Format router: routes file formats
innovation-g2/attachments.py - Attachment analysis → Prediction entry: scores attachment content
innovation-g2/attachments.py - Upload scan → Format router: identifies formats
- Prediction entry → Spam inference: delegates spam
spam_malware_detection_model/predict.py - Prediction entry → Malware inference: delegates malware
spam_malware_detection_model/predict.py - Prediction entry → Process inference: delegates process
spam_malware_detection_model/predict.py - User → Training orchestrator: runs training
- Training orchestrator → Spam training: invokes
spam_malware_detection_model/training/run_all.py - Training orchestrator → Tabular malware training: invokes
spam_malware_detection_model/training/run_all.py - Training orchestrator → File model training: invokes
spam_malware_detection_model/training/run_all.py - Training orchestrator → Process training: invokes
spam_malware_detection_model/training/run_all.py - Spam training → Saved model bundle: saves model
spam_malware_detection_model/spam/train.py - Tabular malware training → Saved model bundle: saves models
spam_malware_detection_model/malware_tabular/train.py - File model training → Saved model bundle: saves specialists
spam_malware_detection_model/malware_engine/training/train_pipeline.py - Process training → Saved model bundle: saves model
spam_malware_detection_model/regression/train.py - Prediction entry → Saved model bundle: loads models
spam_malware_detection_model/predict.py