EDCopilot Conversation Refresher

An automated system for generating and updating custom chatter files for the EDCopilot platform using AI language models.

Features

• Smart AI Conversations: Creates fresh, engaging conversations using advanced AI (OpenAI and Claude) that sound natural and varied
• Personal Touch: Makes conversations feel personal by including your commander name, squadron, and preferences from your settings
• Natural Variety: Creates conversations that feel realistic by mixing in personal details, news references, and context tags at natural intervals
• Fresh News & Content: Pulls in recent Elite Dangerous Galnet to keep conversations current and relevant

Quick Start

1. Clone the repository

   git clone <repository-url>
   cd ConversationRefresher

2. Install dependencies

   py -m pip install -r requirements.txt

3. Configure environment

   copy env.example .env
   # Edit .env with your API keys and settings

4. Run the updater

   py src/main.py

Note: This project uses py to execute Python on Windows systems. If you're using a different operating system, replace py with python or python3 as appropriate.

Configuration

Environment Variables

Create a .env file with the following variables:

# API Keys
KEY_OPENAI=your_openai_api_key_here
KEY_ANTHROPIC=your_anthropic_api_key_here

# Model Configuration
MODEL_ANTHROPIC=claude-3-7-sonnet-20250219
MODEL_OPENAI=gpt-4.1-mini

# Provider Configuration
PROVIDER_PREFERRED=OPENAI

# Directory Configuration
DIR_CUSTOM=path/to/your/edcopilot/custom/directory  # Path to EDCopilot custom files directory (used in production mode)

# Optional Configuration
LOG_LEVEL=INFO
MAX_RETRIES=3
CONTENT_LENGTH=50
CONVERSATIONS_COUNT=30  # Default number of conversation pieces to generate (when not in debug mode)
CONVERSATIONS_CHANCE_PERSONALIZATION=25  # Percentage chance for personalization references
CONVERSATIONS_CHANCE_RSS=15  # Percentage chance for RSS feed references
CONVERSATIONS_CHANCE_CONDITIONALS=10  # Percentage chance for context tags

Output Routing

The system automatically routes output to the appropriate directory based on the mode:

• Production Mode (default): Writes directly to the EDCopilot custom directory specified in DIR_CUSTOM
• Debug Mode (--debug): Writes to the project's output/ directory for testing and review

This ensures that generated content goes directly to EDCopilot when ready for use, while allowing safe testing in debug mode.

Probability-Based Content Generation

The system uses configurable probability percentages to create natural, varied conversations:

• Context Tags (Conditionals): Only 10% of conversations include context tags like (not-station), (not-planet), (not-deep-space)
• Personalization References: Only 25% of conversations include personal references (commander name, squadron, fleet carrier)
• RSS Feed References: Only 15% of conversations reference recent news and events

This ensures conversations feel natural and varied, avoiding over-personalization or excessive context tagging.

Personalization Configuration

Create a personalization.md file to provide context for content generation:

# Personalization

## Specific Data
- Commander Name: Your Commander Name
- Squadron Name: Your Squadron Name
- Theme: Your preferred themes and style

## RSS Feeds
- https://www.elitedangerous.com/en-GB/rss/galnet.xml

## Quick Notes
- Your content preferences and guidelines
- Topics to include or avoid
- Humor style and tone preferences

Creating Custom Prompt Templates

The system uses template files in the prompts/ directory to generate content. You can create and customize these templates to control the style and content of generated conversations.

📖 For detailed template documentation, see prompts/README.md

Usage

Basic Usage

Generate content for all chatter types:

py src/main.py

Test Mode

Validate setup without generating content:

py src/main.py --test

Debug Mode

Output generated content to the shell:

py src/main.py --debug

Specific File Generation

Generate content for specific chatter types:

# Generate for one file type
py src/main.py --files chit_chat --debug

# Generate for multiple file types
py src/main.py --files space_chatter crew_chatter --debug

# Adjust the number of entries (default: 5)
py src/main.py --files deep_space_chatter --max-entries 10 --debug

Personalization Control

Disable personalization features:

# Disable all personalization
py src/main.py --no-personalization

# Disable RSS feeds only
py src/main.py --no-rss

# Disable web content only
py src/main.py --no-web

Content Management

Control content replacement:

# Keep existing content and merge with new content (default: replace entirely)
py src/main.py --keep-existing

Command Line Options

Option	Description
--test	Run in test mode (validate only)
--debug	Enable debug mode - output generated content to shell
--files	Generate content for specific file types only
--max-entries	Maximum number of entries to generate per file (default: from CONVERSATIONS_COUNT config, or 5 in debug mode)
--keep-existing	Keep existing content and merge with new content (default: replace entirely)
--no-personalization	Disable personalization context
--no-rss	Disable RSS feed fetching
--no-web	Disable web content fetching
--clear-cache	Clear RSS cache before running
--cache-info	Show RSS cache information and exit
--prompt-only	enable's debug mode & outputs the prompts that would have been sent to the LLM for review placing them in the output directory
--generate-prompt-template	Generate specific prompt files for each chatter type in the prompts directory (enables debug mode)

Available File Types

• chit_chat - General chit chat conversations
• space_chatter - Space exploration and astronomy discussions
• crew_chatter - Ship crew interactions and operations
• deep_space_chatter - Deep space exploration and mysteries

Examples

Generate 3 entries for chit chat with debug output:

py src/main.py --files chit_chat --debug --max-entries 3

Generate content using configuration default (30 entries):

py src/main.py --files chit_chat

Generate content for space and crew chatter without personalization:

py src/main.py --files space_chatter crew_chatter --no-personalization

Generate 10 entries for all files with debug mode:

py src/main.py --debug --max-entries 10

RSS Cache Management:

# Show cache information
py src/main.py --cache-info

# Clear cache and run
py src/main.py --clear-cache --files chit_chat

Prompt Template Generation:

# Generate specific prompt files for each chatter type
py src/main.py --generate-prompt-template

# Generate prompt files for specific chatter types only
py src/main.py --generate-prompt-template --files chit_chat crew_chatter

# Generate prompt file for one chatter type
py src/main.py --generate-prompt-template --files space_chatter

Project Structure

ConversationRefresher/
├── src/
│   ├── config.py              # Configuration management
│   ├── main.py                # Main orchestrator
│   ├── generators/            # Content generators
│   │   ├── base_generator.py
│   │   ├── chit_chat_generator.py
│   │   ├── space_chatter_generator.py
│   │   ├── crew_chatter_generator.py
│   │   └── deep_space_chatter_generator.py
│   └── utils/                 # Utility modules
│       ├── api_client.py      # API client for OpenAI/Anthropic
│       ├── file_manager.py    # File operations and backups
│       └── personalization.py # Personalization context management
├── prompts/                   # Prompt template files
│   ├── README.md             # Template documentation
│   ├── sample.prompt.md      # Example template
│   └── prompt_*.md           # Chatter type templates
├── scheduler.py               # Windows Task Scheduler integration
├── run_updater.bat           # Simple batch script
├── run_updater.ps1           # PowerShell script
├── test_setup.py             # Setup validation
├── requirements.txt          # Python dependencies
├── .env.example              # Environment template
├── personalization.md        # User personalization context
├── logs/                     # Log files
├── backups/                  # Backup files
├── cache/                    # RSS cache files
└── output/                   # Generated content files

Windows Integration

Task Scheduler Setup

Use the included PowerShell script to set up automated execution:

.\run_updater.ps1 -SetupScheduler -Frequency Weekly

Manual Execution

Run the updater manually using the batch script:

run_updater.bat

If you want EDCopilot to launch after the updater runs make sure you set the following in your .env file:

• START_EDCOPILOT_AFTER_UPDATE=TRUE
• DIR_EDCOPILOT=C:\Path\To\EDCopilot

Troubleshooting

Common Issues

1. API Key Errors: Ensure your API keys are correctly set in the .env file
2. Directory Not Found: Verify the DIR_CUSTOM path exists and is accessible (in production mode) or check the output/ directory (in debug mode)
3. Permission Errors: Run as administrator if needed for file operations
4. Network Issues: Check internet connectivity for API calls and RSS feeds

Logs

Check the logs/ directory for detailed execution logs. Log files are timestamped for easy tracking.

Validation

Run the test script to validate your setup:

py test_setup.py

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

License

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

Technical Constraints

Platform Requirements

• Operating System: Windows 10/11 with PowerShell
• Python Version: Python 3.8 or higher
• Python Execution: Uses py command on Windows systems
• API Dependencies: OpenAI API, Anthropic Claude API access