Arabic PDF Chat 📚💬

About 📖

Arabic Chat with PDF is an innovative tool designed to enable users to interactively query Arabic PDF documents. Powered by state-of-the-art language models and document processing libraries, this application extracts, processes, and retrieves meaningful insights from Arabic text documents. Users can ask questions in Arabic, and the system responds in a professional tone, making this an essential tool for Arabic language researchers, educators, and professionals.

---

Live Demo 🚀

Explore the hosted version of this project on Hugging Face Spaces:

👉

Upload an Arabic PDF and start chatting in seconds! 💬📚

---

Features ✨

• Seamless PDF Integration: Upload Arabic PDFs under 10 MB, and start chatting.
• Advanced Text Recognition: Utilizes OCR for Arabic text extraction from searchable PDFs.
• Conversational Interface: Interact via a chatbot with RTL (Right-To-Left) support for natural Arabic conversation.
• Multilingual Embeddings: Employs multilingual embeddings for precise text analysis.
• Text-to-Speech: Outputs audio responses in Arabic for accessibility.
• Customizable UI: Designed with Arabic aesthetics and user-friendly components.

---

Technologies Used 🛠️

• Python Libraries:
  • Gradio: User-friendly UI for interaction.
  • PyPDF2: PDF text extraction.
  • pytesseract: OCR for PDFs.
  • LangChain: Framework for conversational AI with retrieval-based querying.
  • gTTS: Arabic text-to-speech functionality.
• Machine Learning Models:
  • LLMs: Powered by ChatGroq using the gemma2-9b-it model.
  • Embeddings: Utilizes sentence-transformers/paraphrase-multilingual-mpnet-base-v2.
• Vector Store: FAISS for efficient similarity search and retrieval.

---

Getting Started 🚀

Prerequisites 📋

Ensure you have:

• Python 3.9+
• pip for package management
• Access to API keys:
  • GROQ_API_KEY
  • HF_TOKEN

Installation ⚙️

1. Clone the Repository:

   git clone https://github.com/your-repo/arabic-pdf-chat.git
   cd arabic-chat-with-pdf

2. Install Dependencies:

   pip install -r requirements.txt

3. Set Up Environment Variables:
   Create a .env file and add the required API keys:

   GROQ_API_KEY=your_groq_api_key
   HF_TOKEN=your_huggingface_token

Run the Application 🖥️

Launch the app with:

python app.py

The Gradio interface will open in your browser.

---

ETL Process 🔄

The system follows a structured ETL pipeline:

1. Extract: Reads Arabic PDFs using OCR (pytesseract) and PyPDF2.
2. Transform:
   • Splits text into manageable chunks with CharacterTextSplitter.
   • Converts raw text into vector embeddings using sentence-transformers.
3. Load: Stores transformed data in a FAISS vector database for efficient retrieval.

---

Limitations ⚠️

• File Size: Limited to PDFs under 10 MB.
• Language Support: Optimized only for Arabic text. Non-Arabic content is not supported.
• Scanned Documents: OCR may struggle with low-quality scans.
• Performance: Response times may vary depending on document size and complexity.

---

Author 🖋️

👤 M. N. Gaber
🔗 GitHub Profile
🔗 LinkedIn

---

License 📄

This project is licensed under the Apache License.

---

Acknowledgments 🙏

Special thanks to the developers of the libraries and frameworks that made this project possible.

---

Contributing 🤝

Contributions are welcome! Please fork the repository and submit a pull request. For major changes, open an issue first to discuss what you would like to change.

---

Future Plans 🛠️

• Add support for scanned and handwriting docs.
• Improve OCR accuracy for complex layouts.
• Enhance conversational capabilities with additional LLM models.

---

Enjoy exploring Arabic text in a whole new way! 🎉