Call Recording System

This system handles CDR (Call Detail Record) fetching and recording queries, providing a robust solution for managing and retrieving call recordings across multiple storage backends.

System Overview

The application consists of several worker components that handle different aspects of call recording management:

1. CDR Fetcher Worker: Retrieves Call Detail Records from various sources and processes them for storage
2. Recording Query Worker: Handles queries for call recordings and manages retrieval from different storage backends
3. Recording Handler Worker: Manages the storage of recordings across multiple backend systems
4. Transcription Worker: Processes audio recordings to generate text transcriptions
5. Summary Worker: Generates concise summaries of call transcriptions using LLM technology

Features

• Multi-backend storage support (AWS S3, Backblaze B2, FTP, FTPS, SFTP)
• Automated call transcription
• AI-powered call summarization
• Configurable retry mechanisms for all operations
• Concurrent processing capabilities
• Flexible storage cleanup options

Project Structure

app/
├── config/               # Configuration files and handlers
├── workers/             # Worker components
│   ├── cdr_fetcher/     # CDR retrieval and processing
│   ├── recording_query/ # Recording retrieval management
│   ├── recording_handler/ # Storage backend management
│   ├── transcription/   # Audio transcription processing
│   └── summa/          # Summary generation
├── data/               # Local data storage
├── logs/               # Application logs
└── state/              # Worker state management

Setup

Windows

Run setup.bat

Linux/MacOS

Run ./setup.sh

Configuration

The system is configured through a config.json file. Key configuration areas include:

• Storage backend credentials and settings
• Worker-specific configurations:
  • Retry intervals
  • Polling intervals
  • Cleanup settings
• API endpoints and authentication

See config.json.example for a complete configuration template.

Worker Components

CDR Fetcher

• Retrieves call detail records from various sources
• Processes and stores records in the database
• Configurable polling and retry intervals

Recording Query

• Handles requests for call recordings
• Manages retrieval from multiple storage backends
• Implements concurrent processing

Transcription Worker

• Processes audio recordings to text
• Configurable API integration
• Supports transcript cleanup after processing

Summary Worker

• Generates AI-powered call summaries
• Configurable summary retention
• Supports multiple output formats

Storage Backends

Supported storage systems:

• AWS S3
• Backblaze B2
• FTP
• FTPS
• SFTP

Each backend supports:

• Individual enable/disable toggles
• Custom retry intervals
• Path configurations
• Authentication settings

Development

To set up the development environment:

1. Clone the repository
2. Run the appropriate setup script:
   • Windows: setup.bat
   • Linux/MacOS: ./setup.sh
3. Copy config.json.example to config.json and configure as needed
4. Install dependencies: pip install -r requirements.txt

Requirements

See requirements.txt for a complete list of dependencies.