Add CRC error count endpoint with persistent SQLite storage

[dmduran12]

Summary

Adds a new GET /api/crc_count?hours=24 endpoint that tracks and counts CRC errors detected by the SX1262 radio, with persistent SQLite storage.

Problem

CRC errors ([RX] CRC error detected - discarding packet) are logged by the radio driver but not tracked or queryable via the API. This makes it difficult to monitor radio health over time.

Changes

• sqlite_handler.py — New crc_errors table with timestamp index, store_crc_error() and get_crc_error_count(hours) methods, cleanup in cleanup_old_data()
• storage_collector.py — record_crc_error() and get_crc_error_count() pass-throughs (same pattern as noise_floor)
• http_server.py — CRCErrorTracker logging handler that intercepts CRC error log messages from SX1262_wrapper and persists them to SQLite
• main.py — Register _crc_tracker on root logger alongside _log_buffer
• api_endpoints.py — New GET /api/crc_count?hours=24 endpoint
• web/__init__.py — Export new symbols

Design

• No pymc_core changes needed — intercepts existing log output via a logging.Handler
• Follows the exact same pattern as the existing noise_floor feature (table, handler methods, storage pass-through, API endpoint)
• Data automatically cleaned up by existing cleanup_old_data() retention policy
• Uses the standard ?hours=24 time period parameter consistent with all other endpoints

Response Format

{
  "success": true,
  "data": {
    "crc_error_count": 42,
    "hours": 24,
    "oldest_event": 1707350239.044,
    "newest_event": 1707436639.044
  }
}

Co-Authored-By: Warp agent@warp.dev

[dmduran12]

this endpoint will allow reporting CRC errors in the dashboard @rightup