Entify | Comparative NER Engine Lab

Entify is a high-performance research laboratory designed to benchmark and compare two fundamental architectural shifts in Natural Language Processing: Statistical (CRF) vs Neural (spaCy) Named Entity Recognition.

🚀 Key Features

• Dual-Engine Comparison: Real-time side-by-side analysis of CRF and spaCy models.
• Advanced Confidence Metrics: Probabilistic scores for both engines (CRF Marginals vs Neural Heuristics).
• Localized Presets: Specialized inference samples for News, Tech, Finance, and Business from a Pakistani perspective.
• Context-Aware Analytics: Intelligent metrics board that toggles between "Quality Leader" and "Unique Labels" based on mode.
• Export Capabilities: Seamlessly export findings in JSON or TXT formats.

🛠️ Technical Stack

• Backend: Flask (Python 3.9+)
• NLP Core:
  • Statistical: Custom Conditional Random Fields (CRF) trained on CoNLL2003 (90.3% F1).
  • Neural: spaCy en_core_web_sm (Efficient ~12MB CNN architecture).
• Frontend: Modern Glassmorphism UI (Tailwind CSS, Alpine.js, AOS).
• Deployment: AWS EC2 with Gunicorn + Apache.

📖 Getting Started

1. Clone the repository:

   git clone https://github.com/softdevhassan/entify.git

2. Install dependencies:

   pip install -r requirements.txt

3. Configure environment:
   • Create a .env file based on project metadata.
4. Run the lab locally:

   python app.py

🎓 Project Credit

• Institution: ILM College Sargodha (UOS Affiliated)
• Course: Artificial Intelligence (BSCS Semester 3)
• Mentor: Sir Abdur-Rehman
• Lead Developer: Hassan Ali
• Core Team: Mudassir Ali (Inference Analysis), Saad Ilyas (Design Systems).

---

For more details, visit the official documentation.