Text Summarization API

A Flask-based REST API that uses AWS Bedrock to generate concise summaries of customer messages. This service is particularly useful for quickly understanding the essence of customer communications.

Features

• RESTful API endpoint for text summarization
• Integration with AWS Bedrock's Claude model
• Simple and efficient text processing
• Error handling and input validation

Prerequisites

• Python 3.7+
• AWS Account with Bedrock access
• AWS credentials configured

Installation

1. Clone the repository:

git clone <repository-url>
cd <repository-name>

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Configure AWS credentials:

   • Place your AWS credentials in bedrock-user_accessKeys.csv
   • Ensure the credentials have access to AWS Bedrock

5. Configure the model settings in config.yaml:

model:
  modelId: "anthropic.claude-v2"
  contentType: "application/json"
  accept: "application/json"
  region: "us-east-1"
  parameters:
    max_tokens_to_sample: 100
    temperature: 0.7
    top_k: 250
    top_p: 1

Usage

1. Start the Flask server:

python app.py

2. The API will be available at http://localhost:5000

3. Send a POST request to /summarize endpoint:

curl -X POST http://localhost:5000/summarize \
  -H "Content-Type: application/json" \
  -d '{"message": "Your text to summarize here"}'

Example response:

{
  "summary": "Customer reports being overcharged 40 times for an invoice and expresses frustration."
}

API Endpoints

GET /

Health check endpoint that returns the API status.

POST /summarize

Summarizes the provided text.

Request body:

{
  "message": "Text to summarize"
}

Response:

{
  "summary": "Generated summary"
}

Error Handling

The API returns appropriate HTTP status codes:

• 200: Successful summarization
• 400: Missing or invalid input
• 500: Server error

Testing

Run the test suite:

python -m unittest test_bedrock_client.py

Security Notes

• Never commit AWS credentials to version control
• Keep your config.yaml secure
• Use environment variables for sensitive configuration in production

License

[TBD]