Add comprehensive schema debugging for all Garmin data types

Now all data extraction functions have detailed logging:
- Sleep: API response + DTO keys
- HRV: API response + source format detection + keys
- Heart Rate: API response + data keys
- Stress: API response + data keys
- Weight: Already had comprehensive debugging

This will help identify schema issues across all data types,
not just weight. When sync runs, we'll see exactly what
Garmin API returns for each data type.

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

Co-Authored-By: Claude <noreply@anthropic.com>

Author: nikolay-e

pull_garmin_data.py

@@ -152,11 +152,16 @@ def extract_sleep_data(
     """Extract sleep data for a specific date and user using Pydantic validation."""
     try:
         raw_sleep_data = api.get_sleep_data(date_str)
+        logger.info(f"Sleep API response for {date_str}: {raw_sleep_data}")
         if not raw_sleep_data:
+            logger.info(f"No sleep data found for {date_str}")
             return None
 
         # Extract the actual sleep data from the response
         sleep_dto = raw_sleep_data.get("dailySleepDTO", raw_sleep_data)
+        logger.info(
+            f"Sleep DTO keys for {date_str}: {list(sleep_dto.keys()) if isinstance(sleep_dto, dict) else type(sleep_dto)}"
+        )
 
         # Use Pydantic for robust parsing
         parsed_sleep = GarminSleepData.from_garmin_response(sleep_dto)
@@ -186,15 +191,23 @@ def extract_hrv_data(
     """Extract HRV data for a specific date and user using Pydantic validation."""
     try:
         raw_hrv_data = api.get_hrv_data(date_str)
+        logger.info(f"HRV API response for {date_str}: {raw_hrv_data}")
         if not raw_hrv_data:
+            logger.info(f"No HRV data found for {date_str}")
             return None
 
         # Handle different response formats
         hrv_source = raw_hrv_data
         if "hrvStatusSummary" in raw_hrv_data:
             hrv_source = raw_hrv_data["hrvStatusSummary"]
+            logger.info(f"Using hrvStatusSummary for {date_str}")
         elif isinstance(raw_hrv_data, list) and raw_hrv_data:
             hrv_source = raw_hrv_data[0]
+            logger.info(f"Using list[0] for HRV {date_str}")
+
+        logger.info(
+            f"HRV source keys for {date_str}: {list(hrv_source.keys()) if isinstance(hrv_source, dict) else type(hrv_source)}"
+        )
 
         # Use Pydantic for robust parsing
         parsed_hrv = GarminHRVData.from_garmin_response(hrv_source)
@@ -391,9 +404,15 @@ def extract_heart_rate_data(
         # The get_heart_rates method might need the username
         # Try to get the heart rate data
         hr_data = api.get_heart_rates(date_str)
+        logger.info(f"Heart Rate API response for {date_str}: {hr_data}")
         if not hr_data:
+            logger.info(f"No heart rate data found for {date_str}")
             return None
 
+        logger.info(
+            f"Heart Rate data keys for {date_str}: {list(hr_data.keys()) if isinstance(hr_data, dict) else type(hr_data)}"
+        )
+
         return HeartRate(
             user_id=user_id,
             date=target_date,
@@ -413,9 +432,15 @@ def extract_stress_data(
     """Extract stress data for a specific date and user using Pydantic validation."""
     try:
         raw_stress_data = api.get_stress_data(date_str)
+        logger.info(f"Stress API response for {date_str}: {raw_stress_data}")
         if not raw_stress_data:
+            logger.info(f"No stress data found for {date_str}")
             return None
 
+        logger.info(
+            f"Stress data keys for {date_str}: {list(raw_stress_data.keys()) if isinstance(raw_stress_data, dict) else type(raw_stress_data)}"
+        )
+
         # Use Pydantic for robust parsing
         parsed_stress = GarminStressData.from_garmin_response(raw_stress_data)
         if not parsed_stress: