🧾 AWS Automated Receipt Processing System

A serverless AWS pipeline that automatically extracts, stores, and notifies you about receipt data using AI-powered document analysis.

📐 Architecture

S3 (Upload Receipt)
      ↓
AWS Lambda (Orchestrator)
      ↓
Amazon Textract (Extract Data)
      ↓
Amazon DynamoDB (Store Results)
      ↓
Amazon SES (Email Notification)

✨ Features

• 📤 Upload receipts (PDF, JPG, PNG) to S3 and everything happens automatically
• 🤖 AI-powered data extraction using Amazon Textract's Expense Analysis
• 🗄️ Structured storage of receipt data in DynamoDB
• 📧 Instant email notification with extracted receipt details
• 🛡️ Robust error handling and CloudWatch logging

🛠️ AWS Services Used

Service	Purpose
Amazon S3	Receipt file storage and pipeline trigger
AWS Lambda	Serverless orchestration (Python 3.x)
Amazon Textract	AI receipt data extraction
Amazon DynamoDB	NoSQL receipt data storage
Amazon SES	Email notifications
Amazon CloudWatch	Logging and monitoring

🚀 How It Works

1. A receipt image or PDF is uploaded to the incoming/ folder in the S3 bucket
2. S3 triggers the Lambda function automatically
3. Lambda calls Textract's analyze_expense API to extract structured data
4. Extracted data (vendor, date, total, line items) is stored in DynamoDB
5. A formatted HTML email notification is sent via SES with all receipt details

📦 Project Structure

aws-receipt-processor/
│
├── lambda_function.py      # Main Lambda function code
├── README.md               # Project documentation
└── screenshots/            # Architecture and demo screenshots
    ├── s3-trigger.png
    ├── dynamodb-results.png
    └── email-notification.png

⚙️ Setup & Deployment

Prerequisites

• AWS Account
• Python 3.x
• AWS CLI configured

1. S3 Bucket

• Create an S3 bucket (e.g. your-receipts-bucket)
• Create an incoming/ folder inside it

2. DynamoDB Table

• Create a table with receipt_id as the partition key (String)

3. SES Email Verification

• Verify your sender and recipient email addresses in Amazon SES

4. Lambda Function

• Create a new Lambda function (Python 3.x runtime)
• Paste the code from lambda_function.py
• Set the following environment variables:

Variable	Description
DYNAMODB_TABLE	Your DynamoDB table name
SES_SENDER_EMAIL	Verified sender email address
SES_RECIPIENT_EMAIL	Verified recipient email address

5. IAM Permissions

Attach the following permissions to your Lambda execution role:

• s3:GetObject
• textract:AnalyzeExpense
• dynamodb:PutItem
• ses:SendEmail

6. S3 Trigger

• Add an S3 trigger to your Lambda
• Event type: s3:ObjectCreated:*
• Prefix: incoming/

📧 Sample Email Output

The system sends a formatted HTML email containing:

• Receipt ID
• Vendor name
• Date
• Total amount
• Full list of line items with prices and quantities
• Link to original S3 file

📊 DynamoDB Schema

{
  "receipt_id": "uuid-string",
  "vendor": "Store Name",
  "date": "2024-01-15",
  "total": "42.99",
  "items": [
    { "name": "Item 1", "price": "10.00", "quantity": "2" }
  ],
  "s3_path": "s3://bucket/incoming/receipt.pdf",
  "processed_timestamp": "2024-01-15T10:30:00"
}

🔧 Troubleshooting

Issue	Solution
's3' KeyError	Check your Lambda test event uses the correct S3 event format
SubscriptionRequiredException	Enable Textract Expense Analysis in your AWS account
Email not received	Check spam folder; verify SES identities are confirmed
AccessDenied	Check Lambda IAM role has all required permissions

👤 Author

Khamidullokhon A
GitHub · LinkedIn

📄 License

MIT License - feel free to use this project as a reference or starting point.