Client Data Extractor 📄➡️📊

A Python tool for automated client data extraction from PDF documents using GPT-4 Vision. It converts PDFs to images, analyzes them with GPT-4o, and extracts structured client/business information into a clean CSV file.

---

🚀 Features

• Convert PDFs into images with multiple fallback methods (pdf2image, pdftocairo, PyMuPDF)

• Encode PDF pages as base64 images for GPT-4 Vision

• Extract structured client and business details:

  • Page number, date, timestamp
  • Client ID, company name, address
  • Contact/attorneys, phone, fax, email
  • Status, manager, billing type
  • Accounts receivable balance and more

• Process multiple PDFs at once

• Save results into CSV with detailed logs and extraction statistics

• Handles failures gracefully with empty responses

---

📦 Installation

1. Clone this repository:

   git clone https://github.com/yourusername/client-data-extractor.git
   cd client-data-extractor

2. Install dependencies:

   pip install -r requirements.txt

   Core Dependencies:

   • openai

   • pdf2image

   • pillow

   • pandas

   • PyMuPDF (optional fallback)

   ⚠️ You may need to install Poppler for pdf2image on Windows.

---

🔑 Setup

1. Get an OpenAI API key.

2. Save your API key as an environment variable (or use google.colab.userdata if running in Colab):

   export OPENAI_API_KEY="your_api_key_here"

---

🛠 Usage

Run locally:

python main.py

Workflow:

1. Place your client PDF files inside:

   /content/drive/MyDrive/Client_Documents
   

2. Run the script.

3. Extracted client data will be saved as:

   extracted_client_data_update.csv
   

---

📊 Output Example

Each extracted entry will include fields like:

Document Name	Entry ID	Page Number	Client_Id	Company_Name	Address	Phone	Email	Status	A/R_Balance
client1.pdf	entry_1	2	12345	ABC Corp	123 Main St	(555) 123-4567	client@email.com	Active	$2,500

---

⚡ Notes

• Default limit: 20 pages per PDF (to control GPT-4 Vision costs).
• Each entry missing a field is labeled "Not Found".
• Logs and statistics are shown after each run, including success rate and entry distribution.

---

📈 Roadmap

• Add support for exporting to Excel and JSON
• Build a Streamlit dashboard for previewing extracted data
• Optimize batch processing with async calls to reduce runtime

---

📝 License

This project is licensed under the MIT License. See LICENSE for details.