fix(ingest): Fix JSON serialization for reports with tuple/enum keys

[daha]

What

Fixes JSON serialization error when IngestionStageReport (and other reports) contain tuple or enum dictionary keys.

Why

The DataHub GC source and other sources using stage tracking fail with:

TypeError: keys must be str, int, float, bool or None, not tuple

This occurs because:

1. IngestionStageReport.ingestion_stage_durations uses tuple keys: (IngestionHighStage, str)
2. IngestionStageReport.ingestion_high_stage_seconds uses enum keys: IngestionHighStage
3. Report.to_pure_python_obj() doesn't properly convert these to JSON-compatible strings

JSON specification requires object keys to be strings (or numbers converted to strings). Python tuple and enum keys must be explicitly converted.

How

Updated Report.to_pure_python_obj() method in metadata-ingestion/src/datahub/ingestion/api/report.py:

• Changed dict comprehension to explicit loop for better control
• Added explicit handling for tuple keys: convert to string representation
• Added explicit handling for enum keys: use .value if available, fallback to string
• Maintained backward compatibility for existing key types (str, int, etc.)
• Recursively processes nested structures

Testing

Unit Tests Added

• Test tuple key conversion (simple and nested)
• Test enum key conversion
• Test TopKDict with tuple keys
• Test actual IngestionStageReport serialization
• Test backward compatibility with existing key types

Related Issues

Fixes #15445

Before you submit your PR, please go through the checklist below:

• The PR conforms to DataHub's Contributing Guideline (particularly PR Title Format)
• Links to related issues (if applicable)
• Tests for the changes have been added/updated (if applicable)
• Docs related to the changes have been added/updated (if applicable). If a new feature has been added a Usage Guide has been added for the same.
• For any breaking change/potential downtime/deprecation/big changes an entry has been made in Updating DataHub