Dummy Bank API Documentation

A comprehensive FastAPI-based banking API designed for portfolio analysis and customer data management, optimized for CustomerOID-based operations and MCP (Model Context Protocol) server integration.

🚀 Quick Start

Prerequisites

• Python 3.8+

• Conda environment (recommended)

Installation & Setup

1. Set up environment:

   conda create -n openbanking-backend python=3.11 conda activate openbanking-backend

2. Install dependencies:

   pip install fastapi uvicorn sqlmodel sqlite3

3. Start the API server:

   conda run -n openbanking-backend uvicorn main:app --host 127.0.0.1 --port 3000 --reload

4. Access API documentation:

   • Swagger UI: http://127.0.0.1:3000/docs

   • ReDoc: http://127.0.0.1:3000/redoc

📋 API Overview

Base URL

Core Features

• CustomerOID-centric design: All operations are based on unique Customer IDs (UUID format)

• Comprehensive portfolio data: Assets, transactions, bank accounts, spending patterns, derivatives

• Customer lifecycle management: Registration, deletion, and data retrieval

• MCP server integration ready: Optimized responses for analysis agents

• SQLite database: Local file-based storage with indexed CustomerOID fields

🔑 CustomerOID Format

The API uses UUID format for CustomerOIDs:

• Format: 550e8400-e29b-41d4-a716-446655440000

• Generation: Automatic UUID generation for new customers

• Validation: Strict UUID format validation with backward compatibility

📊 Data Models

User

• customer_oid (UUID): Unique customer identifier

• name (string): Customer name

Asset Holdings

• Asset type, symbol, quantity, current value

• Market data and performance metrics

Bank Accounts

• Account details, balances, institution information

• Multiple accounts per customer supported

Transactions

• Transaction history with amounts, dates, descriptions

• Categorized for analysis

Spending Patterns

• Spending categories and amounts

• Useful for financial behavior analysis

Derivative Transactions

• Options, futures, and other derivative instruments

• Advanced portfolio components

🛠 API Endpoints

Customer Management

GET /customers

List all registered customers.

Response:

POST /register-customer

Register a new customer in the system.

Request Body:

Response:

Features:

• Auto-generates UUID if customer_oid not provided

• Validates name (minimum 2 characters)

• Prevents duplicate registrations

• Returns comprehensive error messages

DELETE /customer/{customer_oid}

Delete a customer and all associated data.

⚠️ WARNING: This permanently deletes all portfolio data for the customer.

Response:

GET /customer/{customer_oid}/exists

Quick check if a CustomerOID exists in the system.

Response:

Portfolio Data

GET /user-portfolio/{customer_oid}

Get comprehensive portfolio data for a customer.

Response Structure:

System Health

GET /health

Health check endpoint for monitoring.

Response:

🗄 Database Schema

Tables

• users: Customer information with UUID CustomerOID (indexed)

• assets: Investment holdings per customer

• bank_accounts: Banking account details

• institutions: Bank and financial institution data

• transactions: Transaction history

• spending: Spending pattern data

• derivative_transactions: Derivative instrument trades

Key Features

• Indexed CustomerOID: All tables indexed on customer_oid for fast queries

• Referential integrity: Proper foreign key relationships

• SQLModel ORM: Type-safe database operations

• Automatic seeding: Sample data for testing and development

🧪 Testing

Test Files Included

1. test_api.py: Comprehensive API testing

2. test_registration.py: User registration workflow testing

3. reset_database.py: Database reset and reseeding

Running Tests

🔧 Configuration

Environment Variables

• Port: Default 3000 (configurable)

• Host: Default 127.0.0.1

• Database: SQLite file banking.db

Sample Data

The API includes comprehensive sample data for 5 customers:

• Complete portfolio data

• Multiple asset types

• Various transaction patterns

• Diverse spending categories

• Derivative instruments

🎯 MCP Server Integration

This API is optimized for MCP (Model Context Protocol) server integration:

Key Features for MCP

• CustomerOID-centric design: Easy customer identification

• Comprehensive data structure: All portfolio data in single endpoint

• Analysis-ready format: Pre-calculated summaries and metrics

• Efficient queries: Indexed database operations

• Standardized responses: Consistent JSON structure

Recommended MCP Usage

1. Customer Discovery: Use /customers to list available customers

2. Portfolio Analysis: Use /user-portfolio/{customer_oid} for complete data

3. Customer Management: Use registration/deletion endpoints for lifecycle management

4. Health Monitoring: Use /health for system status

🚨 Error Handling

Common Error Codes

• 400: Invalid CustomerOID format or validation errors

• 404: Customer not found

• 409: Duplicate customer registration

• 500: Internal server errors

Error Response Format

🔄 Development Workflow

Making Changes

1. Edit code: Modify main.py or db.py

2. Restart server: Server restart required for validation changes

3. Test changes: Run test scripts to verify functionality

4. Reset data: Use reset script if database schema changes

Adding New Customers

📝 Best Practices

CustomerOID Management

• Always use UUID format for new customers

• Validate CustomerOID before database operations

• Use the provided validation function consistently

Database Operations

• Use dependency injection for database sessions

• Implement proper error handling and rollbacks

• Close sessions properly to prevent leaks

API Design

• Follow RESTful conventions

• Provide comprehensive error messages

• Include validation for all inputs

• Document all endpoints clearly

🤝 Contributing

1. Follow the existing code structure

2. Add tests for new functionality

3. Update documentation for API changes

4. Ensure CustomerOID validation is maintained

5. Test with sample data before deployment

📄 License

This is a dummy/example implementation for development and testing purposes.

Need help? Check the Swagger documentation at http://127.0.0.1:3000/docs when the server is running.