🛡️ Enterprise XGBoost Software Debt Underwriter

Created by Srinivasta

Fulfilling TCoE Domain VII: Gradient-Boosted Decision Tree Inference Application

Live Application URL: Enterprise XGBoost Software Debt Underwriter

This repository contains the complete codebase for an intelligent, automated technical debt and code compliance underwriting portal. The prototype utilizes a dual-layer validation engine—combining deterministic Abstract Syntax Tree (AST) scanning with classical machine learning inference via XGBoost—to quantify risk profiling models into actionable monetary financial exposure matrices.

---

🏗️ Core Repository Architecture

The platform isolates engine evaluation matrices from deployment scripts to ensure enterprise code hygiene:

my-grant-prototype/
│
├── .github/
│   └── workflows/
│       └── test.yml          # GitHub Actions CI automated pipeline integration
│
├── app.py                    # Streamlit Web UI, Core 4-Phase Underwriting Pipeline & Plotly Graphs
├── generate_test_data.py     # Unified Pytest-Native Unit Testing Suite & Fixture Controls
├── pytest.ini                # Custom collector rule linking pytest to the testing script
└── requirements.txt          # Defined application & verification matrix dependencies

---

⚡ Operational Pipeline Mechanics

The underwriting gateway processes ingestion packages (.zip archives) through four sequential phases:

1. Phase 1: Feature Engineering via AST Parser
   Strict, deterministic attribute-matching algorithms parse python files into tabular coordinate metrics while filtering infrastructure utilities (e.g., wake_app.py) to eliminate algorithmic false-positives.
2. Phase 2: XGBoost Machine Learning Inference
   Executes gradient-boosted decision trees over numeric feature coordinates [Total_Lines, Insecure_Calls, Secrets, Char_Length] to generate categorical high-risk probability bounds.
3. Phase 3: Financial Exposure Calculation Waterfall
   Translates flagged structural security vectors into localized financial liabilities ($\text{INR } ₹$) using custom widget parameters tracking base severity breach costs and engineer labor hours.
4. Phase 4: Plotly UI Visual Analytics Mapping
   Outputs interactive data dashboards mapping codebase profile traits beside a live stdout execution container trace log.

---

🚀 Quick Start Deployment

1. Local Environment Provisioning

Ensure you have Python 3.10+ installed. Clone the repository workspace and run pip installations:

pip install -r requirements.txt

2. Launch the Streamlit Web Application Portal

Boot up the process container natively to access the interactive web interface:

streamlit run app.py

3. Run the Automated Compliance Pytest Suite

Thanks to pytest.ini, you can trigger our entire test suit matrix simply by typing the base command:

pytest

---

⚙️ Automated CI/CD Lifecycle Trace

This system features integrated GitHub Actions (test.yml) automation loops. Every branch push or pulled merge request spins up an isolated virtual runner environment to execute full compliance matrix assertions against your generate_test_data.py definitions.

• Underwriting Policy Target: Verified Clean Code Ingestions drop computed portfolio penalties down to a baseline ₹0 Total Liability. --

📄 License & Copyright

⚠️ IMPORTANT COPYRIGHT NOTICE

All Rights Reserved © 2026 T A Srinivas. This repository is strictly for portfolio viewing purposes. DO NOT COPY, CLONE, OR REDISTRIBUTE this code. Stolen copies or unauthorized forks will be reported immediately for a GitHub copyright takedown.

• Lead Architect & Developer: Srinivasta

🌐 Let’s Connect

• 
• 
• 
• 
•