added

app.py

@@ -13,6 +13,7 @@ import pytesseract
 import re
 from openai import OpenAI
 import os
+from pdf2image import convert_from_bytes
 from dotenv import load_dotenv
 from chatbot_utils import ask_receipt_chatbot
 import time
@@ -178,7 +179,12 @@ def main():
     
     if uploaded_file is not None:
         # Display uploaded image
-        image = Image.open(uploaded_file)
+        if uploaded_file.type == "application/pdf":
+    # Convert first page of PDF to image
+            images = convert_from_bytes(uploaded_file.read())
+            image = images[0]  # Use the first page
+        else:
+            image = Image.open(uploaded_file)
         st.image(image, caption="Uploaded Document", width=600)
 
         # Process button

insert_dummy_data.py

@@ -1,51 +0,0 @@
-import boto3
-
-# Initialize DynamoDB resource (ensure AWS credentials and region are set)
-dynamodb = boto3.resource('dynamodb', region_name='us-east-1')  # Change region if needed
-table = dynamodb.Table('Receipts')  # Replace with your table name
-
-# List of dummy items to insert with meaningful receipt numbers
-dummy_items = [
-    {
-        'receipt_no': 'RCPT-2024-0001',
-        'amount_paid': '100.00',
-        'date': '2024-01-01',
-        'name': 'John Doe',
-        'product': 'Widget A'
-    },
-    {
-        'receipt_no': 'RCPT-2024-0002',
-        'amount_paid': '250.50',
-        'date': '2024-02-15',
-        'name': 'Jane Smith',
-        'product': 'Gadget B'
-    },
-    {
-        'receipt_no': 'RCPT-2024-0003',
-        'amount_paid': '75.25',
-        'date': '2024-03-10',
-        'name': 'Alice Johnson',
-        'product': 'Thingamajig C'
-    },
-    {
-        'receipt_no': 'RCPT-2024-0004',
-        'amount_paid': '180.00',
-        'date': '2024-04-05',
-        'name': 'Bob Lee',
-        'product': 'Gizmo D'
-    },
-    {
-        'receipt_no': 'RCPT-2024-0005',
-        'amount_paid': '320.75',
-        'date': '2024-05-20',
-        'name': 'Carol King',
-        'product': 'Device E'
-    }
-]
-
-# Insert each item
-for item in dummy_items:
-    table.put_item(Item=item)
-    print(f"Inserted: {item['receipt_no']}")
-
-print("Dummy data inserted successfully.") 

packages.txt

@@ -1 +1,2 @@
-tesseract-ocr
+tesseract-ocr
+poppler-utils

requirements.txt

@@ -33,3 +33,4 @@ streamlit
 plotly==5.18.0
 matplotlib
 scikit-learn
+pdf2image