📄 PDF to JSON Parser

A Python tool that extracts structured content (paragraphs, tables, charts) from PDF documents and converts it into a hierarchical JSON format.

Built using:

• Streamlit – Interactive UI
• pdfplumber – Text & table extraction
• PyMuPDF (fitz) – Chart/image detection
• pandas – Table rendering in UI

---

🎯 Features

• Extracts:
  • 📝 Paragraphs with section/subsection context
  • 📊 Tables (structured data)
  • 📈 Charts (embedded images)
• Maintains page-level hierarchy
• Exports clean, well-structured JSON
• Interactive Streamlit web interface
• Download extracted JSON directly

---

⚙️ Installation

1. Clone the repository

git clone <your-repo-url>
cd <your-repo-folder>

2. Create & activate virtual environment (recommended)

python -m venv venv
source venv/bin/activate   # On Linux/Mac
venv\Scripts\activate      # On Windows

3. Install dependencies

pip install -r requirements.txt

---

🚀 Usage

Run the Streamlit app

streamlit run app.py

Steps

1. Upload a PDF file (resume, factsheet, etc.)
2. Click Process PDF
3. Preview extracted content (page-wise)
4. Download the JSON output

---

📂 Project Structure

.
├── app.py              # Streamlit web app
├── pdf_parser.py       # Core parser (paragraph, table, chart extraction)
├── utils/
│   └── styling.py      # Custom CSS for Streamlit UI
├── requirements.txt    # Python dependencies
└── README.md           # Project instructions

---

📊 Example JSON Output

{
  "pages": [
    {
      "page_number": 1,
      "content": [
        {
          "type": "paragraph",
          "section": "FACTSHEET",
          "sub_section": null,
          "text": "June 2025 Mutual Fund investments are subject to market risks..."
        },
        {
          "type": "chart",
          "section": "FACTSHEET",
          "sub_section": null,
          "description": "Embedded image detected - possible chart",
          "image_index": 1
        },
        {
          "type": "table",
          "section": "Financial Data",
          "sub_section": null,
          "table_data": [
            ["Year", "Revenue", "Profit"],
            ["2022", "$10M", "$2M"]
          ]
        }
      ]
    }
  ]
}

---

📝 Notes

• Paragraphs are detected using regex-based section/subsection rules.
• Tables are extracted only if the PDF has true tabular structures.
• Charts are flagged when images are found (exported as metadata, not raw image files).
• For scanned PDFs, OCR (e.g., Tesseract) would be required (not included here).

---

✅ Assignment Requirements Covered

• Input: PDF → Output: JSON
• Maintains page hierarchy
• Detects paragraphs, tables, charts
• Preserves sections & subsections
• Exportable JSON + Streamlit UI