Add multi-device support for training status endpoints

[matin]

Description

The current training status implementation only supports single-device scenarios. The code assumes the first device in the API response and ignores additional devices.

Current Limitation

In the _parse_response methods for training status classes, we use:

# Get the first device's data (assumes single device for now)
for device_data in report_data.values():
    # Process only the first device and return
    return [device_data] if isinstance(device_data, dict) else result

Proposed Enhancement

Add support for multiple devices by:

1. Processing all devices in the API response instead of just the first one
2. Adding device identification to the returned data structures
3. Updating the data model to include device-specific information
4. Maintaining backward compatibility with existing single-device usage

Implementation Ideas

• Add a device_id field to clearly identify which device each training status entry belongs to
• Modify _parse_response to iterate through all devices in reportData
• Consider adding a class method like list_by_device() for device-specific queries
• Update tests to cover multi-device scenarios

Files Affected

• src/garth/stats/training_status/daily.py
• src/garth/stats/training_status/weekly.py
• src/garth/stats/training_status/monthly.py
• tests/stats/test_training_status.py

Priority

Low - Current single-device support covers most use cases, but this would be valuable for users with multiple Garmin devices.