`, ``, `

`), and citations, with automatic fallback to PyPDF layout parsing. 3. **Structured Semantic Schema Extraction (Step 2 Stage 2):** Maps raw section text into validated Pydantic `PaperSchema` JSON files via `Gemini 2.0 Flash` or heuristic grounding. 4. **LangGraph Multi-Agent Orchestration:** Stateful multi-agent pipeline (*Extraction, Summarizer, Comparator, Gap Identification, and Report Writer*) operating on a centralized state graph. 5. **Zero-Hallucination Citation Guardrail:** Vector embedding cosine similarity matching (`similarity >= 0.85`) to guarantee page-level provenance. 6. **Command-Line & Web Dashboard Interface:** Includes both a lightweight terminal script (`demo.py`) and an interactive split-screen web interface (`app.py`). --- ## 📁 Repository Structure ``` Agentic-AI-for-literature-review/ │── README.md # Project Documentation │── requirements.txt # Python Dependencies │── app.py # Streamlit Web Application │── demo.py # Command-Line CLI Live Demo Script │── data/ │ ├── pdf_corpus/ # Downloaded Benchmark Research Paper PDFs │ └── extracted_schemas/ # Extracted Pydantic Paper JSONs & Parsed Layouts └── src/ ├── ingestion/ # Ingestion & PDF Parser Module │ ├── downloader.py # Automated arXiv Paper Downloader (Step 1) │ ├── grobid_parser.py # GROBID TEI-XML REST Service Client │ └── parser.py # 2-Tier Hybrid Layout Parser (Step 2 Stage 1) └── agents/ # Semantic Extraction & Agent Workflows └── extraction.py # Pydantic Schema Extraction Agent (Step 2 Stage 2) ``` --- ## ⚙️ Setting Up GROBID (Optional but Recommended) GROBID provides structural TEI-XML extraction for multi-column academic paper PDFs. ### Option A: Using Docker (Recommended) ```bash docker run --rm --init -p 8070:8070 lfoppiano/grobid:0.8.0 ``` ### Option B: Local Java Service ```bash ./gradlew run ``` *Note: If the GROBID server (`http://localhost:8070`) is not running, the system automatically falls back to PyPDF layout parsing.* --- ## 🚀 Quick Start Guide ### 1. Clone the Repository ```bash git clone https://github.com/Chaitanya-G41/Agentic-AI-for-literature-review.git cd Agentic-AI-for-literature-review ``` ### 2. Set Up Virtual Environment & Install Dependencies ```bash python -m venv venv # On Windows: venv\Scripts\activate # On Linux/macOS: source venv/bin/activate pip install -r requirements.txt ``` ### 3. Set API Keys (Optional for Gemini LLM Extraction) ```bash # On Windows PowerShell: $env:GEMINI_API_KEY="your_api_key_here" # On Linux/macOS: export GEMINI_API_KEY="your_api_key_here" ``` ### 4. Step 1: Download Benchmark Research Corpus ```bash python src/ingestion/downloader.py ``` ### 5. Step 2: Run PDF Layout Parser & Schema Extraction ```bash python src/ingestion/parser.py ``` ### 6. Run the CLI Demo (No Web Server Needed) ```bash python demo.py ``` ### 7. Run the Streamlit Web Application ```bash streamlit run app.py ``` --- ## 📜 License This project is open-source under the [MIT License](LICENSE).