Update README.md

README.md

@@ -4,7 +4,7 @@ emoji: 💬
 colorFrom: yellow
 colorTo: purple
 sdk: gradio
-sdk_version: 5.42.0
+sdk_version: 6.0.0
 app_file: app.py
 pinned: false
 hf_oauth: true
@@ -13,4 +13,137 @@ hf_oauth_scopes:
 license: mit
 ---
 
-An example chatbot using [Gradio](https://gradio.app), [`huggingface_hub`](https://huggingface.co/docs/huggingface_hub/v0.22.2/en/index), and the [Hugging Face Inference API](https://huggingface.co/docs/api-inference/index).
+# APJ Threat Intelligence System  
+Mobile-First, Multilingual Cybercrime Intelligence Platform
+
+## Overview
+This project delivers a full-stack threat-intelligence console focused on
+Asia-Pacific & Japan (APJ) cybercrime ecosystems. The system ingests
+Mandarin/Cantonese underground-market chatter, interprets idioms and cultural
+nuances, classifies threats using transformer models, and presents insights to
+analysts through a mobile-first Gradio interface.
+
+The design is modular, allowing you to use:
+- Local Transformers models (HuggingFace)
+- External LLM APIs
+- Custom datasets
+- A growing slang / idiom lexicon
+- Marketplace monitoring pipelines
+
+This repository is optimized for **GitHub** and **HuggingFace Spaces**.
+
+---
+
+## ✨ Features
+
+### 🔍 Intelligence Layer
+- Threat classification (Transformers)
+- Vendor graph modeling
+- Marketplace and trend analysis
+- Slang & idiom identification
+
+### 🌏 Multilingual Processing
+- Mandarin + Cantonese dialect detection  
+- Literal + functional translation  
+- Cultural interpretation for cybercrime slang  
+
+### 📱 Mobile-First UI
+Built with Gradio 4.x:
+- Single-column layout  
+- Mode switcher (Threat Intel / Translation / Marketplace Watch / Analyst Tools)  
+- File upload for logs, screenshots, raw text  
+- Downloadable chat transcripts  
+- Clean UX optimized for mobile operators  
+
+### 🔧 Built With
+- `transformers`
+- `datasets`
+- `gradio`
+- Python 3.10+
+
+---
+
+## 🧩 Repository Structure
+
+.
+├── app.py                   # Main Gradio app
+├── prompt_engine.py         # Centralized prompt construction
+├── model_inference.py       # Transformers-based inference wrapper
+├── datasets_loader.py       # HuggingFace datasets loader utilities
+├── slang_lexicon.json       # Evolving idiom/slang dictionary
+├── PROJECT_SPEC.md          # Architectural overview
+├── HUGGINGFACE.md           # HF Spaces deployment instructions
+├── GITHUB_SETUP.md          # Repo usage & development guide
+├── requirements.txt
+└── README.md
+
+---
+
+## 🚀 Running Locally
+
+### 1. Clone the repo
+```bash
+git clone https://github.com/<yourname>/apj-threat-intel
+cd apj-threat-intel
+
+2. Install requirements
+
+pip install -r requirements.txt
+
+3. Launch the app
+
+python app.py
+
+The interface opens automatically in your browser.
+
+⸻
+
+🧠 Model Integration
+
+You can plug your own HuggingFace model into model_inference.py:
+
+ThreatModel(model_path="your-model-name")
+
+Or use an API-driven LLM via prompt_engine.py.
+
+⸻
+
+🌐 Deploy on HuggingFace Spaces
+
+See HUGGINGFACE.md in this repo for step-by-step instructions.
+
+⸻
+
+📄 Docs
+	•	Project Spec
+	•	HF Spaces Setup
+	•	GitHub Setup Guide
+
+⸻
+
+📜 License
+
+MIT (You may swap this with your preferred license.)
+
+⸻
+
+🤝 Contributing
+
+Pull requests and issue reports are welcome.
+
+⸻
+
+🧭 Roadmap
+	•	Vendor identity resolution model
+	•	Marketplace scraping connectors
+	•	Cantonese pretrained language model fine-tuning
+	•	In-browser graph explorer
+	•	Real-time alerting engine
+
+⸻
+
+✉️ Contact
+
+For questions, enhancements, or collaboration, open an issue.
+
+---