Create app.py

app.py

@@ -0,0 +1,193 @@
+import gradio as gr
+import os
+from dotenv import load_dotenv
+import pandas as pd
+from groq import Groq
+from PIL import Image
+import base64
+import io
+import openpyxl
+from datetime import datetime
+
+# Load environment variables from .env file
+load_dotenv()
+
+# Initialize Groq client
+client = Groq(
+    api_key=os.environ.get("GROQ_API_KEY")
+)
+
+def encode_image_to_base64(image_path):
+    """Convert image to base64 string"""
+    with open(image_path, "rb") as image_file:
+        return base64.b64encode(image_file.read()).decode('utf-8')
+
+def extract_invoice_details(image):
+    """Extract invoice details using Groq's vision model"""
+    # Save the uploaded image temporarily
+    temp_path = "temp_invoice.png"
+    image.save(temp_path)
+    
+    # Convert image to base64
+    base64_image = encode_image_to_base64(temp_path)
+    
+    # Remove temporary file
+    os.remove(temp_path)
+    
+    # Prepare the prompt
+    prompt = """Analyze this invoice image and provide ONLY ONE dictionary with the following format, including all line items. Remove any special characters (*, #, $) and format numbers as plain decimal values:
+
+    {
+        "Invoice Number": "inv-00",  # Remove special chars, keep alphanumeric only
+        "Invoice Date": "07/07/2025",  # Use MM/DD/YYYY format
+        "Items": [
+            {
+                "Item Name": "Product 1",  # Clean text only
+                "Price/Rate": "40.00",  # Numeric only
+                "Quantity": "2",  # Numeric only
+                "Amount": "80.00"  # Numeric only
+            },
+            {
+                "Item Name": "Product 2",
+                "Price/Rate": "19.00",
+                "Quantity": "6",
+                "Amount": "114.00"
+            }
+        ],
+        "Total Invoice Value": "2555.00"  # Numeric only, total amount
+    }
+
+    Provide ONLY the dictionary, no additional text or formatting."""
+    
+    # Make API call to Groq
+    response = client.chat.completions.create(
+        model="llama-3.2-90b-vision-preview",
+        messages=[
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "image_url",
+                        "image_url": {
+                            "url": f"data:image/png;base64,{base64_image}"
+                        }
+                    },
+                    {
+                        "type": "text",
+                        "text": prompt
+                    }
+                ]
+            }
+        ]
+    )
+    
+    return response.choices[0].message.content
+
+def parse_response(response_text):
+    """Parse the model's response into structured data"""
+    try:
+        # Find the dictionary part of the response
+        start_idx = response_text.find('{')
+        end_idx = response_text.rfind('}') + 1
+        if start_idx != -1 and end_idx != -1:
+            dict_str = response_text[start_idx:end_idx]
+            # Safely evaluate the dictionary string
+            data = eval(dict_str)
+            
+            # Create rows for each item in the items list
+            rows = []
+            for item in data.get('Items', []):
+                row = {
+                    'Invoice Number': data.get('Invoice Number', ''),
+                    'Invoice Date': data.get('Invoice Date', ''),
+                    'Item Name': item.get('Item Name', ''),
+                    'Price/Rate': item.get('Price/Rate', ''),
+                    'Quantity': item.get('Quantity', ''),
+                    'Amount': item.get('Amount', ''),
+                    'Total Invoice Value': data.get('Total Invoice Value', '')
+                }
+                rows.append(row)
+            
+            return rows
+    except Exception as e:
+        print(f"Error parsing response: {e}")
+        return [{
+            'Invoice Number': '',
+            'Invoice Date': '',
+            'Item Name': '',
+            'Price/Rate': '',
+            'Quantity': '',
+            'Amount': '',
+            'Total Invoice Value': ''
+        }]
+
+def save_to_excel(data_rows):
+    """Save cleaned extracted data to Excel file"""
+    excel_file = "invoice_data.xlsx"
+    
+    # Create new DataFrame with the current data only
+    df = pd.DataFrame(data_rows, columns=[
+        'Invoice Number', 'Invoice Date', 'Item Name',
+        'Price/Rate', 'Quantity', 'Amount', 'Total Invoice Value'
+    ])
+    
+    # Apply number formatting for currency columns
+    currency_columns = ['Price/Rate', 'Amount', 'Total Invoice Value']
+    for col in currency_columns:
+        df[col] = pd.to_numeric(df[col], errors='ignore')
+    
+    # Save to Excel with formatting
+    with pd.ExcelWriter(excel_file, engine='openpyxl') as writer:
+        df.to_excel(writer, index=False, sheet_name='Invoice Data')
+        
+        # Get the workbook and worksheet
+        workbook = writer.book
+        worksheet = writer.sheets['Invoice Data']
+        
+        # Apply currency formatting to relevant columns
+        for col_idx, col_name in enumerate(df.columns):
+            if col_name in currency_columns:
+                for row in range(2, len(df) + 2):  # Start from row 2 to skip header
+                    cell = worksheet.cell(row=row, column=col_idx + 1)
+                    cell.number_format = '$#,##0.00'
+    
+    return excel_file
+
+def process_invoice(image):
+    """Main function to process invoice image"""
+    try:
+        # Extract text from image
+        extracted_text = extract_invoice_details(image)
+        
+        # Parse the response
+        data = parse_response(extracted_text)
+        
+        # Save to Excel
+        excel_path = save_to_excel(data)
+        
+        return (
+            f"Successfully processed invoice!\n\n"
+            f"Extracted Information:\n{extracted_text}",
+            excel_path
+        )
+    
+    except Exception as e:
+        return f"Error processing invoice: {str(e)}", None
+
+# Create Gradio interface
+iface = gr.Interface(
+    fn=process_invoice,
+    inputs=gr.Image(type="pil", label="Upload Handwritten Invoice"),
+    outputs=[
+        gr.Textbox(label="Processing Result"),
+        gr.File(label="Download Excel File")
+    ],
+    title="Handwritten Invoice Processor",
+    description="Upload a handwritten invoice to extract key information and save it to Excel.",
+    examples=[],
+    theme=gr.themes.Base()
+)
+
+# Launch the application
+if __name__ == "__main__":
+    iface.launch()