Skip to content

Latest commit

 

History

History
143 lines (110 loc) · 4.91 KB

PROJECT_STRUCTURE.md

File metadata and controls

143 lines (110 loc) · 4.91 KB

Neural Network Project Structure

This project is organized into modular components for easy reuse and development of different neural network applications.

Folder Organization

📁 foundational_brain/

The core neural network implementation that can be used as a foundation for any pattern recognition task.

Contents:

  • NeuralNet.py - The foundational neural network with forward/backpropagation
  • README.md - Documentation for the foundational network
  • .gitignore - Version control configuration

Use this when: You want to build a new neural network application from scratch.

📁 medical_diagnosis_model/

A complete medical diagnosis system built on top of the foundational neural network.

Contents:

  • All medical diagnosis source files (20+ files)
  • Training architecture diagram
  • Complete documentation
  • Demo scripts
  • Requirements file

Use this when: You want to run or extend the medical diagnosis system.

How to Use

For New Projects

  1. Copy the foundational_brain folder
  2. Rename it to your project (e.g., image_recognition_model)
  3. Import and extend the base neural network: 
    from NeuralNet import initialize_network, train_network

For Medical Diagnosis

  1. Navigate to medical_diagnosis_model/
  2. Create virtual environment and install dependencies
  3. Run python enhanced_medical_system.py

Moving Between Projects

Each folder is self-contained with all necessary dependencies:

# Work on medical diagnosis
cd medical_diagnosis_model/
python enhanced_medical_system.py

# Start a new project
cp -r foundational_brain/ my_new_model/
cd my_new_model/
# Begin development...

Architecture Overview

PythonNeuralNet/
├── foundational_brain/          # Core neural network
├── medical_diagnosis_model/     # Medical application
└── future_model/               # Your next project

This modular structure allows you to:

  • Keep models separate and organized
  • Reuse the foundational code
  • Develop multiple applications in parallel
  • Easily share specific models without unnecessary files

Medical Model Directory Map (v2)

medical_diagnosis_model/
  backend/
    app.py                 # FastAPI app: /api/v2/diagnose, /export, /adaptive/*
    security/jwt_dep.py    # OIDC/JWT scaffold (feature-flagged)
    selector/eig_selector.py # EIG ranking util (math-only)
  configs/
    clinical_schema.yaml   # Clinical data mappings (scaffold)
    training.yaml          # Training/optimizer toggles (scaffold)
  data/
    case.schema.json       # JSON Schema for clinical cases
    dictionaries/          # Canonical symptoms/diseases
    samples/               # Example JSONL dataset
    validate_cases.py      # Schema validator CLI
  tests/
    test_api_phase1.py     # API TestClient expected primaries
    test_selector.py       # EIG selector unit test
    phase_1_backend/       # cURL-based test harness & outputs
  tools/
    sanity.py              # Modular sanity CLI (data/tests/api/export/rate/adaptive/suite)
  versions/
    v1/, v2/               # Model implementations and demos
  exports/                 # Generated reports (gitignored)
  models/                  # Saved models (dev artifacts)
  diagnosis_history/       # Session logs (gitignored)
  README.md                # Usage + Auth
  NEXT_STEPS.md            # Roadmap & acceptance criteria

Task Map (completed → pending)

  • Completed

    • Data scaffolding: data/case.schema.json, data/dictionaries/*, data/samples/*, data/validate_cases.py
    • Backend phase 1: /diagnose, /export endpoints; CORS; API key auth; request logging; rate limiting
    • Security scaffold: OIDC/JWT dependency (backend/security/jwt_dep.py) behind MDM_AUTH_MODE=oidc
    • CI: GitHub Actions runs dataset validation and unit tests via sanity CLI
    • Adaptive (alpha): /api/v2/adaptive/{start,answer,finish} + EIG selector; sanity CLI coverage

  • Pending (next)

    • Synthetic generator → emit JSONL to schema (medical_training_generator.py refactor)
    • Golden set (100 curated cases) with rationale/certainty
    • Data quality tests (sanity rules, leakage, balance) and DATA_CARD.md
    • Metrics & calibration module (AUROC/AUPRC/F1, reliability/ECE)
    • Batch scoring CLI (backend/tools/batch) and outputs
    • Adaptive end-to-end unit tests (FastAPI TestClient)
    • RAG PoC (knowledge/, index/, retriever/generator)
    • Frontend adaptive mode; persistence/workers; deployment

Quick Testing Commands

# From medical_diagnosis_model/
python tools/sanity.py data
python tools/sanity.py tests
python tools/sanity.py api --auto-start --api-key devkey
python tools/sanity.py export --auto-start --api-key devkey
python tools/sanity.py rate --auto-start --api-key devkey --count 140 --expect-over-limit
python tools/sanity.py adaptive --auto-start --api-key devkey
python tools/sanity.py suite --auto-start --api-key devkey --with-api --with-export --with-rate --with-adaptive