Corrected model list based on official Perplexity API documentation:
REMOVED (not available via API):
- llama-3.1-sonar-small-128k-online
- llama-3.1-sonar-large-128k-online
- llama-3.1-sonar-huge-128k-online
ADDED (actual API models):
- sonar-reasoning-pro (powered by DeepSeek-R1)
- sonar-deep-research (long-form research)
AVAILABLE MODELS:
1. sonar-pro (recommended for chat bots)
2. sonar (lightweight, fast)
3. sonar-reasoning (real-time reasoning)
4. sonar-reasoning-pro (advanced reasoning)
5. sonar-deep-research (research reports)
Updated info popup with:
- Accurate model descriptions from official docs
- Built on Llama 3.3 70B (sonar/sonar-pro)
- DeepSeek-R1 for sonar-reasoning-pro
- Clear use case recommendations
Updated MODEL_CHANGE.md with correct information.
Source: Perplexity API official documentation (2026)
Setup wizard improvements:
- Expanded model dropdown with 6 Perplexity models:
* sonar-pro (recommended default)
* sonar
* sonar-reasoning
* llama-3.1-sonar-small-128k-online
* llama-3.1-sonar-large-128k-online
* llama-3.1-sonar-huge-128k-online
- Added info button (?) next to model selector
- Info popup explains each model's characteristics:
* Performance
* Use cases
* Speed/quality tradeoffs
* Cost considerations
Created MODEL_CHANGE.md documentation:
- How to change model in .env file
- How to re-run setup wizard
- Detailed model descriptions
- Example configuration
Users can now easily choose the right model for their needs
and change it anytime without reconfiguring everything.
Enhanced bot to respond to partial names and ambiguous greetings:
MentionDetector improvements (utils.py):
- Auto-generate nicknames from bot name (kenearosmd → Kene, Kenearos)
- Add patterns for all nicknames (full name + partial names)
- New is_ambiguous_greeting() method for greetings without clear mention
- Updated extract_content() to remove all name variants
Bot logic improvements (chatbot.py):
- New check_if_addressed() method uses AI to determine if ambiguous
messages are directed at the bot
- AI analyzes greetings like "Hi wie gehts" and decides if bot should respond
- If AI confirms, message is processed like a normal mention
Recognition examples:
- kenearosmd: wie gehts → responds ✅
- Kene was meinst du → responds ✅
- Kenearos! → responds ✅
- Hi wie gehts → AI checks context → responds if appropriate ✅
This prevents the bot from spamming on every greeting while still
being responsive when addressed indirectly.
Enhanced MentionDetector to recognize bot name in more contexts:
Detection patterns now include:
- @botname anywhere (with word boundary)
- botname with punctuation (: ! ? , .)
- botname at start of message
- botname anywhere as whole word (new catchall pattern)
Content extraction improved to remove bot name from:
- Start of message (with @mention or plain)
- End of message (with optional punctuation)
- Middle of message (with punctuation)
Now responds to all these variations:
- @kenearosmd Hi
- kenearosmd: wie gehts?
- kenearosmd was meinst du?
- Hallo kenearosmd
- kenearosmd!
- Frage an kenearosmd hier
The bot was detecting mentions correctly but not responding because
the async event loop was never started.
Changes:
- Created separate _run_event_loop() method to run loop.run_forever()
- Start event loop in its own daemon thread before IRC bot
- IRC bot thread no longer sets event loop (handled by loop thread)
- Now asyncio.run_coroutine_threadsafe() can properly execute async tasks
This fixes the issue where mentions were detected but handle_mention()
was never executed because there was no running event loop to process
the coroutine.
Improved error handling in bot shutdown process:
- Check connection state before attempting to quit IRC
- Use thread-safe event loop shutdown
- Filter expected shutdown errors from error logs
- Changed shutdown errors to debug level logging
This prevents the "file descriptor cannot be a negative integer (-1)"
error from appearing as an ERROR when cleanly shutting down the bot.
Fixed inconsistent error handling in ai_provider.py and memory.py by:
- Adding optional logger parameter to PerplexityProvider and ConversationMemory classes
- Replacing all print() statements with proper logger.error() calls
- Updating chatbot.py to pass logger instance to both components
- Ensuring consistent logging architecture across the codebase
This improves error visibility and integrates with the existing logging infrastructure.
- Log extracted content to help debug empty content issue
- Add warning when content is empty after mention extraction
- Display extraction info in dashboard for better visibility
Helps diagnose why bot detects mentions but doesn't respond
- Replace PySimpleGUI 5.0.8.3 with FreeSimpleGUI 5.1.1
- FreeSimpleGUI is a community fork without license requirements
- PySimpleGUI 5.x requires paid license after trial period
- Remove compatibility fallback code (no longer needed)
- Update documentation to reflect FreeSimpleGUI usage
Fixes: AttributeError 'module PySimpleGUI has no attribute theme'
- Fix CI/CD failure due to PySimpleGUI 4.60.0 being removed from PyPI
- Upgrade to PySimpleGUI 5.0.8.3 (latest stable version)
- GUI code uses basic features that should remain compatible
Changes:
- Create setup_wizard.py with interactive configuration flow
- Validates Twitch OAuth token via IRC test
- Validates Perplexity API key with test request
- Guides user through all required settings
- Creates .env file and necessary directories
- Update config.py system prompt to reflect Kene's personality
- Bot now responds AS Kene (first person)
- Includes tech expertise (Bambu X1C, H2C, coding/bots)
- Reflects political stance (left-leaning, anti-right)
- Maintains authentic conversational style
- Update README.md with setup wizard instructions
- Clear step-by-step for Option 1 (wizard)
- Lists wizard features and benefits
This commit implements the full Eugen bot based on specifications in CLAUDE.md and eugen_claude.md.
Features implemented:
- Smart name recognition (@Eugen, Eugen:, etc.)
- Persistent conversation memory per user (max 25 messages, 1 hour retention)
- Perplexity Sonar API integration for AI responses
- Live monitoring dashboard with PySimpleGUI
- Setup wizard for first-time configuration
- Comprehensive logging (main log + API debug log)
Files added:
- config.py: Configuration management from .env and config.json
- utils.py: MentionDetector and Logger utility classes
- memory.py: ConversationMemory for persistent chat history
- ai_provider.py: PerplexityProvider for API integration
- gui.py: Dashboard and SetupWizard GUI components
- chatbot.py: Main EugenBot orchestrator with IRC handling
- requirements.txt: Python dependencies
- .env.example: Template for environment variables
- .gitignore: Renamed from gitignore for proper Git usage
Updated:
- README.md: Complete usage instructions and documentation
The bot is ready to use - users just need to add their API keys and run python chatbot.py
