## Major Changes

### OpenAI SDK Migration (0.28 → 1.x+)
- Created LLMClient abstraction layer (quantcli/llm_client.py)
- Migrated all openai.ChatCompletion.create() calls to new SDK
- Removed global openai.api_key configuration
- Added LLMResponse dataclass for standardized responses
- Improved error handling and logging for LLM calls

### Dependency Updates
- Bumped Python requirement to >= 3.9
- Updated OpenAI SDK to >= 1.0.0
- Added modern dependencies: rich, pytest, mypy, ruff
- Created clean requirements.txt (replaced legacy freeze)
- Updated setup.py to v1.0.0 with enhanced metadata

### Testing Infrastructure
- Added pytest configuration (pytest.ini)
- Created test suite for LLMClient (tests/test_llm_client.py)
- Configured coverage reporting
- Added test markers (unit/integration/slow)

### Documentation
- Completely rewrote README.md (removed "legacy" branding)
- Added CHANGELOG.md with migration guide
- Updated installation instructions
- Added badges and modern formatting
- Highlighted v1.0.0 improvements

### Code Quality
- Added LLMClient with proper type hints
- Improved logging in processor.py
- Token usage tracking for cost monitoring
- Structured error handling

## Breaking Changes
- Requires OpenAI SDK >= 1.0.0 (incompatible with 0.28)
- Python >= 3.9 required (was 3.8)

## Migration Path
Users can upgrade seamlessly:
```bash
pip install --upgrade openai>=1.0.0
pip install -e .
```

No code changes required - LLMClient handles SDK differences internally.

---

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <[email protected]>

- Move tkinter imports from module level to function level in processor.py
- Make GUI imports conditional in cli.py
- CLI commands now work without python3-tk system package
- Interactive mode shows helpful error message if tkinter unavailable
- All other CLI commands (search, list, download, etc.) fully functional

PROBLEM: Generated algorithms had 3 runtime errors on first backtest
- NoneType comparison in max/min operations
- Division by zero in position sizing
- Missing null guards before trading logic

SOLUTION: Multi-layered quality improvements

1. Enhanced Code Generation Prompt
   - Added explicit defensive programming requirements
   - Included safety pattern examples (None checks, division guards)
   - Emphasized PRODUCTION-READY code generation

2. Enhanced Refinement Prompt
   - Added safety checklist to refinement process
   - Ensures all defensive patterns are present

3. New QuantConnectValidator Class
   - Runtime safety validation layer
   - Detects: division by zero, None comparisons, missing IsReady checks
   - Provides actionable fix suggestions
   - File: quantcli/qc_validator.py

4. CLI Improvements
   - Fixed None return value handling in generate-code command
   - Better error messages for failed code generation

IMPACT:
- Improves generated code reliability
- Reduces manual fixes required
- Moves toward production-ready algorithms

TODO (v1.1.0):
- Integrate validator into generation pipeline
- Add test suite
- Document in CHANGELOG

Issue: See /tmp/github_issue_quality.md for full details

## Security Enhancements
- Add missing `requests` import in utils.py (CRITICAL BUG FIX)
- Implement URL validation before webbrowser.open() calls
- Add validate_url() function to check URL safety
- Replace hardcoded email with UNPAYWALL_EMAIL environment variable

## Configuration
- Create .env.example documenting all environment variables
- Add support for UNPAYWALL_EMAIL configuration

## Changes
- quantcli/utils.py: Add requests import, urlparse, validate_url()
- quantcli/cli.py: Add URL validation to browser operations
- quantcli/gui.py: Add URL validation to browser operations
- .env.example: Document all configurable environment variables

These changes improve security by validating all URLs before opening
them in the browser and fix the critical missing requests import bug.

- Installation verification steps
- Security testing procedures
- Example workflows
- Troubleshooting guide
- Complete CLI command reference

…V1sadPRrxj5sPHjWp7Wa

Claude/refactor modernize 2025 011 cv1sad p rrxj5s p hj wp7 wa

Claude/add enhanced help beta jwrs m