Add comprehensive test suite with 97% code coverage

Implemented full test coverage for core bot components:

Test Coverage:
- MentionDetector: 98% (28 tests) - mention detection, nickname handling, content extraction
- Logger: 98% (15 tests) - file logging, log levels, unicode support
- ConversationMemory: 91% (25 tests) - history management, time filtering, JSON persistence
- PerplexityProvider: 100% (22 tests) - API calls, error handling, statistics
- Config: 100% (26 tests) - env loading, validation, JSON config

Infrastructure:
- Added pytest, pytest-asyncio, pytest-mock, pytest-cov to requirements.txt
- Created pytest.ini with coverage configuration
- Created .coveragerc to exclude non-production files
- Added conftest.py with shared fixtures and test isolation

Test Features:
- 116 total tests, all passing
- Isolated test environment (clean env vars, logging handlers)
- Async testing support for PerplexityProvider
- Mocked HTTP requests to avoid real API calls
- Comprehensive edge case coverage
- Unicode/German character support testing

Total: 97% code coverage across 273 statements

tests/README.md

@@ -0,0 +1,111 @@
+# Test Suite for Eugen Twitch Bot
+
+Comprehensive test coverage for the Eugen bot components.
+
+## Test Coverage
+
+**Overall: 97% code coverage**
+
+### Components Tested
+
+- **MentionDetector** (utils.py) - 98% coverage
+  - Explicit mentions (@Eugen, Eugen:, etc.)
+  - Nickname detection and generation
+  - Ambiguous greeting detection
+  - Content extraction
+  - Edge cases and unicode support
+
+- **Logger** (utils.py) - 98% coverage
+  - File-based logging
+  - Log levels (INFO, DEBUG, ERROR, WARNING)
+  - API call logging
+  - Unicode character support
+
+- **ConversationMemory** (memory.py) - 91% coverage
+  - User history storage and retrieval
+  - Time-based message filtering
+  - Message limits enforcement
+  - JSON persistence
+  - Error handling
+
+- **PerplexityProvider** (ai_provider.py) - 100% coverage
+  - API request/response handling
+  - Error handling (timeouts, network errors, HTTP errors)
+  - Statistics tracking
+  - API key validation
+
+- **Config** (config.py) - 100% coverage
+  - Environment variable loading
+  - JSON configuration
+  - Validation
+  - Default values
+
+## Running Tests
+
+### Run all tests
+```bash
+pytest tests/
+```
+
+### Run with coverage report
+```bash
+pytest tests/ --cov=. --cov-report=term-missing
+```
+
+### Run specific test file
+```bash
+pytest tests/test_utils.py
+pytest tests/test_memory.py
+pytest tests/test_ai_provider.py
+pytest tests/test_config.py
+```
+
+### Run specific test
+```bash
+pytest tests/test_utils.py::TestMentionDetector::test_mention_with_at_symbol
+```
+
+## Test Structure
+
+```
+tests/
+├── __init__.py              # Package marker
+├── conftest.py              # Shared fixtures and test configuration
+├── test_utils.py            # Tests for MentionDetector and Logger
+├── test_memory.py           # Tests for ConversationMemory
+├── test_ai_provider.py      # Tests for PerplexityProvider
+└── test_config.py           # Tests for Config
+```
+
+## Key Features
+
+- **Test Isolation**: Each test runs in a clean environment with:
+  - Isolated environment variables
+  - Clean logging handlers
+  - Temporary directories for file operations
+
+- **Async Testing**: Full support for async/await with pytest-asyncio
+
+- **Mocking**: HTTP requests mocked using pytest-mock to avoid real API calls
+
+- **Fixtures**: Reusable test data in conftest.py:
+  - `temp_dir`: Temporary directory with automatic cleanup
+  - `mock_env_file`: Sample .env file
+  - `mock_config_json`: Sample config.json
+  - `sample_conversation_history`: Test chat history
+  - `mock_perplexity_response`: Sample API response
+
+## Test Statistics
+
+- **Total Tests**: 116
+- **Passing**: 116 (100%)
+- **Code Coverage**: 97%
+- **Test Execution Time**: ~1.2 seconds
+
+## Coverage Gaps
+
+The 3% uncovered code consists of:
+- Error handling edge cases in memory.py (lines 92-94, 111-112, 143-144)
+- Rarely-used utility functions in utils.py (lines 101, 165)
+
+These are defensive code paths that are difficult to trigger in tests.