Context-Aware Islamic Knowledge Assistant 🤖📖

Overview 🌍

A specialized AI system that combines web scraping, semantic search, and a large language model to provide authentic, context-grounded answers about Islamic practices. Maintains theological accuracy while leveraging modern AI capabilities.

Key Innovations 🚀

Domain-Specific Architecture

# answerer/generate_answer.py
def generate_answer(question, context):
    prompt = f"""
    Based on Islamic content from islamveihsan.com:
    {context}
    
    Answer: {question}
    
    Include:
    - Quranic verses (with Surah:Verse)
    - Hadith references
    - Scholarly consensus
    - Practical applications
    """

Technical Differentiation

Technical Architecture 🧩

graph TD
    A[User Question] --> B(Web Crawler)
    B --> C[Raw HTML]
    C --> D{Content Processor}
    D --> E[Clean Text]
    E --> F[Whoosh Index]
    A --> G[Semantic Search]
    F --> G
    G --> H[Relevant Context]
    H --> I[LLM Processor]
    I --> J[Validated Answer]

Core Components 🔨

1. Islamic Knowledge Pipeline

# create_sample_data.py
sample_data = [
    {
        "title": "Abdestte Unutarak Su Yutmak Orucu Bozar mı?",
        "content": "Peygamber Efendimiz (s.a.v): 'Ümmetimden hata...'",
        "category": "Oruç",
        "sources": ["Sahih Bukhari 1234", "Quran 2:286"]
    }
]

2. Context-Aware AI Engine

# main.py
def ask_question(question, context):
    prompt = f"""
    You are an Islamic scholar assistant. Rules:
    1. Answer ONLY using provided context
    2. Cite sources as [Source #]
    3. If unsure, state "Allah knows best"
    
    Context: {context}
    Question: {question}
    """

Installation & Setup 💻

Requirements

• Python 3.9+
• Node.js (for future web interface)
• 2GB RAM minimum

Windows Setup

run_project.bat

Linux/macOS Setup

# 1. Create virtual environment
python -m venv .venv
source .venv/bin/activate

# 2. Install dependencies
pip install -r requirements.txt

# 3. Run pipeline
python create_sample_data.py && \
python indexer/build_index.py && \
python main.py

Why This Project Stands Out 🌟

1. Theological Integrity Guardrails

• Content validation pipeline
• Source citation requirements
• Fallback mechanisms for uncertain answers

2. Technical Innovations

# fix_data_json.py
def islamic_content_validator(text):
    required_keys = ["quran_refs", "hadith_refs", "scholar_consensus"]
    return all(key in text for key in required_keys)

3. Scalable Knowledge Base

# Crawler execution
python crawler/run_crawler.py \
  --start_url https://www.islamveihsan.com \
  --max_depth 3 \
  --islamic_domains_only True

Validation Process ✅

1. Content Indexing Check
2. Prompt Engineering Layer
3. Response Validation

# answer_validation.py
def validate_answer(answer):
    checks = [
        "This answer is Islamically appropriate",
        "Contains verifiable sources",
        "No speculative content"
    ]
    return all(llm_check(check, answer) for check in checks)

Scalability Roadmap 🛣️

1. Distributed Crawling System
2. Multilingual Support (Arabic/Urdu)
3. Mobile App Integration
4. Cloud Storage Sync

Contribution Guidelines 🤝

# 1. Clone repo
git clone https://github.com/iBz-04/Hudgent.git

# 2. Create feature branch
git checkout -b feature/your-choice

# 3. Setup dev environment
python -m pip install -r requirements-dev.txt

Ethical Considerations ⚖️

• All content verified by Islamic scholars
• No personal opinions in answers
• Clear source attribution
• Regular theological audits

FAQ ❓

Q: How is this different from ChatGPT?
A: Constrained to verified sources with academic citation requirements

Q: Can I add my own content sources?
A: Yes, through the config/custom_sources.json file

Q: What about different madhahib?
A: Current version follows Hanafi fiqh with plans for other schools

Author

Ibrahim Rayamah