feat: Add comprehensive error handling module with custom exceptions

[xingzihai]

Summary

This PR adds a unified error handling module (scripts/lib/errors.py) that significantly improves error handling and logging capabilities across the last30days-skill project.

Changes

1. Custom Exception Hierarchy

A structured exception hierarchy with the following types:

• Last30DaysError: Base exception for all project errors
• NetworkError: Network-level errors (DNS, connection refused)
• APIError: External API errors with status code and response body
• RateLimitError: Rate limiting errors (HTTP 429) with retry_after support
• TimeoutError: Operation timeout errors
• AuthError: Authentication/authorization errors
• ConfigError: Configuration errors
• ValidationError: Input validation errors
• ParsingError: Data parsing errors
• DependencyError: Missing dependency errors with install instructions
• SourceUnavailableError: Data source unavailability errors

2. Rich Error Context

• ErrorContext dataclass for structured debugging information
• ErrorSeverity enum (LOW, MEDIUM, HIGH, FATAL)
• ErrorCategory enum for error classification
• Automatic traceback capture in debug mode
• Serialization support via to_dict()

3. Retry Mechanism

• @with_retry decorator with exponential backoff
• Support for RateLimitError.retry_after to respect API rate limits
• Configurable retry callbacks
• Jitter to avoid thundering herd problem
• Async version available

4. Utility Functions

• ErrorAggregator for collecting multiple errors in parallel operations
• wrap_exception() to convert generic exceptions to Last30DaysError
• safe_call() for safe function execution with fallback
• http_error_to_last30days() to bridge existing http.HTTPError

5. Comprehensive Test Suite

• 43 unit tests covering all exception types
• Tests for retry decorator behavior
• Tests for error aggregation and utilities

Testing

All 43 tests pass:

python3 -m unittest tests.test_errors -v

Benefits

1. Better Debugging: Rich error context with source, operation, and extra data
2. Structured Error Handling: Categorized errors with severity levels
3. Automatic Recovery: Built-in retry mechanism for transient failures
4. API Integration: RateLimitError.retry_after is automatically used by retry decorator
5. Backward Compatible: Works alongside existing http.HTTPError and RedditRateLimitError

[mvanhorn]

See my reply on #109 covering all your PRs. Can't commit to anything right now with the v3.0 refactor underway, but I promise I'll consider each one once it lands. Appreciate the work.