Skip to content

CONFIGURATION.md

Latest commit

 

History

History
196 lines (142 loc) · 5.37 KB

CONFIGURATION.md

File metadata and controls

196 lines (142 loc) · 5.37 KB

Configuration Guide

This guide covers all configuration options for the YouTube Value Extractor.

Environment Variables

The tool uses environment variables for configuration. You can set these in a .env file or export them directly.

Core Configuration

Variable Description Default Required
LLM_MODEL The LLM model to use for analysis gpt-4o-mini No
OPENAI_API_KEY OpenAI API key (required for GPT models) - For OpenAI models
ANTHROPIC_API_KEY Anthropic API key (required for Claude models) - For Anthropic models

Output Configuration

Variable Description Default Required
DEFAULT_OUTPUT_DIR Default directory for output files ./notes No
REPORT_TZ Timezone for report timestamps America/Costa_Rica No

Caching Configuration

Variable Description Default Required
ENABLE_CACHE Enable/disable caching true No
CACHE_DIR Directory for cache files ./.cache No

Processing Configuration

Variable Description Default Required
MAX_CONCURRENT_VIDEOS Maximum concurrent video processing 3 No

Whisper Configuration (Optional)

Only needed if you want to use Whisper for fallback transcription when YouTube transcripts aren't available.

Variable Description Default Required
WHISPER_MODEL Whisper model size (tiny, base, small, medium, large) base No
WHISPER_DEVICE Processing device (auto, cuda, cpu) auto No
WHISPER_COMPUTE_TYPE Compute precision (float16, float32, int8) float16 No

LLM Model Options

OpenAI Models (Recommended)

  • gpt-4o-mini - Fast and cost-effective (default)
  • gpt-4o - Higher quality, more expensive
  • gpt-4-turbo - Good balance of speed and quality
  • gpt-3.5-turbo - Fastest, least expensive

Anthropic Models

  • claude-3-5-sonnet-latest - Latest Sonnet model
  • claude-3-5-haiku-latest - Fastest Claude model
  • claude-3-opus-latest - Highest quality Claude model

Local Models (via Ollama)

  • ollama/llama3.1:8b - Meta's Llama 3.1 8B
  • ollama/llama3.1:70b - Meta's Llama 3.1 70B (requires significant RAM)
  • ollama/qwen2.5:7b - Alibaba's Qwen 2.5 7B

Configuration Examples

Example 1: OpenAI with Custom Output

# .env file
LLM_MODEL=gpt-4o-mini
OPENAI_API_KEY=sk-your-openai-key-here
DEFAULT_OUTPUT_DIR=./video-insights
REPORT_TZ=America/New_York
ENABLE_CACHE=true

Example 2: Anthropic with Whisper Fallback

# .env file
LLM_MODEL=claude-3-5-sonnet-latest
ANTHROPIC_API_KEY=your-anthropic-key-here
DEFAULT_OUTPUT_DIR=./reports
WHISPER_MODEL=small
WHISPER_DEVICE=cuda

Example 3: Local Ollama Setup

# .env file
LLM_MODEL=ollama/llama3.1:8b
DEFAULT_OUTPUT_DIR=./local-analysis
ENABLE_CACHE=false
MAX_CONCURRENT_VIDEOS=1

CLI Configuration

You can override environment variables using CLI flags:

# Override output directory
python -m yt_extractor.cli process VIDEO_URL --output-dir ./custom-output

# Use verbose mode
python -m yt_extractor.cli process VIDEO_URL --verbose

# Override concurrent processing
python -m yt_extractor.cli batch videos.txt --concurrent 5

Configuration Management Commands

Check Current Configuration

python -m yt_extractor.cli config check

This command will:

  • Show all current configuration values
  • Validate API keys
  • Check model accessibility
  • Verify directory permissions

Initialize New Configuration

python -m yt_extractor.cli config init

This creates a new .env file with default values and prompts for required settings.

Best Practices

Performance Optimization

  1. Model Selection:

    • Use gpt-4o-mini for most use cases (fast, cost-effective)
    • Upgrade to gpt-4o for complex technical content
    • Consider claude-3-5-sonnet-latest for detailed analysis

  2. Caching:

    • Keep caching enabled (ENABLE_CACHE=true)
    • Transcripts are cached for 7 days
    • LLM responses are cached for 30 days

  3. Concurrent Processing:

    • Default MAX_CONCURRENT_VIDEOS=3 works well
    • Increase for faster batch processing (if API limits allow)
    • Decrease for rate-limited APIs or limited resources

Security

  1. API Keys:

    • Never commit .env files to version control
    • Use environment variables in production
    • Rotate keys regularly

  2. File Permissions:

    • Ensure output directories have appropriate permissions
    • Cache directory should be writable

Troubleshooting

Common Issues

  1. Missing API Key:

    ConfigurationError: OPENAI_API_KEY required for OpenAI models

    Solution: Set the appropriate API key for your chosen model.

  2. Model Not Available:

    LLMProcessingError: Model gpt-5 not found

    Solution: Check model name spelling and availability.

  3. Cache Permission Error:

    CacheError: Failed to initialize cache

    Solution: Ensure cache directory is writable or change CACHE_DIR.

Validation

Run the config check command to validate your setup:

python -m yt_extractor.cli config check

This will identify configuration issues and suggest fixes.