PodcastInsights

PodcastInsights is a Python-based tool for processing podcast RSS feeds, downloading episodes, and extracting transcripts for further analysis. This project aims to create a foundation for building insights and analytics from podcast content.

Features

Core Functionality

• Parse podcast RSS feeds to extract episode metadata
• Download podcast episodes from feeds
• Transcribe audio to text using Whisper speech recognition
• Store organized data with proper file management

Usage Examples

from podcast_processor import PodcastProcessor

# Initialize the processor
processor = PodcastProcessor()

# Process a podcast feed (with a limit of 5 episodes)
results = processor.process_feed("https://feeds.megaphone.fm/darknetdiaries", max_episodes=5)

# Print results
for result in results:
    print(f"\nTitle: {result['title']}")
    print(f"Audio: {result['audio_path']}")
    print(f"Transcript: {result['transcript_path']}")

Installation

Prerequisites

• Python 3.12+
• FFmpeg (required for audio processing)

Setup

Using Conda (recommended)

# Clone the repository
git clone https://github.com/yourusername/podcast-insights.git
cd podcast-insights

# Create the conda environment
conda env create -f environment.yaml

# Activate the environment
conda activate podcast-processor

Using pip

# Clone the repository
git clone https://github.com/yourusername/podcast-insights.git
cd podcast-insights

# Create a virtual environment
python -m venv venv

# Activate the environment
# On Windows:
venv\Scripts\activate
# On macOS/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Project Structure

podcast-insights/
├── podcast_processor/      # Core package
│   ├── __init__.py
│   ├── processor.py        # Main processing functionality
│   ├── feed_parser.py      # RSS feed parsing utilities
│   ├── audio_handler.py    # Audio download and processing
│   └── transcriber.py      # Speech-to-text transcription
├── scripts/                # Utility scripts
├── data/                   # Default data storage
│   ├── audio/              # Downloaded audio files
│   └── transcripts/        # Generated transcripts
└── tests/                  # Test suite

Configuration

Set custom paths and preferences by editing the config.yaml file:

output_directory: "data"
whisper_model: "base"  # Options: tiny, base, small, medium, large
max_episodes: 10  # Default limit for batch processing

Planned Features

Near-term Roadmap

• Episode metadata extraction (show notes, timestamps, etc.)
• Basic content analysis (topic identification, keyword extraction)
• Simple web interface for browsing podcasts and transcripts

Future Enhancements

• Speaker diarization (identifying different speakers in transcripts)
• Topic segmentation (dividing episodes into thematic sections)
• Sentiment analysis of podcast content
• Cross-episode thematic analysis
• Automated summarization and key point extraction
• Content-based recommendation engine
• Transcript search and indexing
• Potential integration with LLM-based RAG systems for intelligent querying

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.