Invoice Data Extractor 📄👨‍🔧

This project is a Streamlit-based application for extracting data from invoices in various file formats, such as images (PNG, JPG), PDFs, and DOCX files. The extracted data is displayed in a tabular format, saved into a pandas DataFrame with standardized formatting, and exportable in multiple formats.

Features 🛠️

• File Format Support: Supports PNG, JPG, PDF, and DOCX files.
• Data Extraction: Extracts key information, such as:
  • Invoice Number
  • Invoice Date/Time
  • Invoice Amount
  • Total Amount
  • GST
  • Subtotal
  • Invoiced From
  • Invoiced To
• Standardized Output:
  • Dates in YYYY-MM-DD format.
  • Currency fields prefixed with INR.
  • Missing or invalid fields replaced with None.
• Export Options:
  • Export data as CSV, Excel, or JSON.
• Error Handling: Handles invalid inputs and unsupported file types gracefully.

Technologies Used 🤖

• Streamlit: For building the user interface.
• Google Gemini AI: For document understanding and data extraction.
• Python Libraries:
  • pandas: Data manipulation and storage.
  • Pillow: Image handling.
  • fitz (PyMuPDF): PDF processing.
  • re: Regular expressions for data parsing.

Installation ⚙️

1. Clone the repository:

   git clone https://github.com/your-repository/invoice-extractor.git
   cd invoice-extractor

2. Install dependencies:

   pip install -r requirements.txt

3. Configure Google Gemini API:

   • Obtain your API key from Google Gemini AI.
   • Set the API key in the project:

     import google.generativeai as genai
     genai.configure(api_key="your_actual_api_key")

Usage 🔄

1. Run the application:

   streamlit run app.py

2. Upload a file (PNG, JPG, PDF, or DOCX).

3. View extracted data in tabular format.

4. Export the data to your preferred format (CSV, Excel, JSON).

File Processing Workflow ⏳

1. Input Handling:
   • Uploaded files are identified by type.
   • Images are loaded with Pillow.
   • PDFs are converted to images using fitz.
2. Data Extraction:
   • Google Gemini AI processes the content to extract relevant fields.
   • Extracted data is parsed and formatted into a pandas DataFrame.
3. Standardization:
   • Dates are converted to a uniform format (YYYY-MM-DD).
   • Numeric fields are prefixed with INR and rounded to two decimal places.
   • Missing data is replaced with None.
4. Export Options:
   • Data can be exported as CSV, Excel, or JSON.

Example Code 🔧

Data Parsing and Standardization

import pandas as pd
import re

def standardize_date(date_str):
    try:
        return pd.to_datetime(date_str, format="%d.%m.%Y", errors="coerce")
    except Exception:
        return None

def standardize_inr(value):
    try:
        return f"INR {float(value):,.2f}" if value else None
    except ValueError:
        return None

# Apply standardization to DataFrame
currency_fields = ["invoiced_amount", "total", "gst", "subtotal"]
for field in currency_fields:
    df[field] = df[field].apply(standardize_inr)

df["invoiced_date_time"] = df["invoiced_date_time"].apply(standardize_date)

Example Output 📊

Invoice Number	Invoice Date/Time	Invoice Amount	Total	GST	Subtotal
IN-MAA4-3685	2020-08-25	INR 2,450.00	INR 2,450.00	INR 373.72	INR 2,076.28
IN-MAA4-3686	NaT	INR 3,737.50	INR 3,737.50	INR 567.82	INR 3,170.18
None	NaT	None	None	None	None

License 🔒

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

Contact 📧

For any queries, feel free to contact ramusiva77@gmail.com.