feedback_forensics.data.operations.merge

Merge operation for AnnotatedPairs datasets.

Module Contents

Functions

merge_ap	Merge two AnnotatedPairs datasets.
_categorize_comparisons	Categorize comparisons between datasets by ID overlap.
_merge_single_comparison	Merge two comparisons, first takes precedence in conflicts.
_merge_value	Generic value merging with conflict checking.
_merge_dict	Generic dictionary merging with optional strict conflict checking.
_merge_prompt	Merge prompt strings from two comparisons, raising on conflicts.
_merge_response	Merge response objects from two comparisons, raising on conflicts.
_merge_annotations	Merge annotation dictionaries from two comparisons, warning on conflicts.
_merge_comparison_metadata	Merge comparison-level metadata from two comparisons, detecting conflicts.
_merge_annotators	Merge annotator definitions, first takes precedence.
_merge_name_or_description	Merge name or description fields, reusing if identical or creating combined name.
_merge_default_annotator	Merge default annotator fields with first precedence and warning on disagreement.
_merge_metadata	Create merged metadata.

API

feedback_forensics.data.operations.merge.merge_ap(first_data: Dict[str, Any], second_data: Dict[str, Any], dataset_name: Optional[str] = None, description: Optional[str] = None) → Dict[str, Any]

Merge two AnnotatedPairs datasets.

Args: first_data: First AnnotatedPairs dataset (takes precedence in conflicts) second_data: Second AnnotatedPairs dataset dataset_name: Override dataset name for merged result description: Override description for merged result

Returns: Merged AnnotatedPairs data structure

Raises: ValueError: If inputs are invalid

feedback_forensics.data.operations.merge._categorize_comparisons(comps1: List[Dict], comps2: List[Dict]) → Tuple[List[Tuple[int, int]], List[int], List[int]]

Categorize comparisons between datasets by ID overlap.

Returns: Tuple of: - paired: List of (first_idx, second_idx) tuples for matches - first_only: List of indices only in first dataset - second_only: List of indices only in second dataset

feedback_forensics.data.operations.merge._merge_single_comparison(comp1: Dict, comp2: Dict) → Dict

Merge two comparisons, first takes precedence in conflicts.

feedback_forensics.data.operations.merge._merge_value(value1: Any, value2: Any, context: str, strict: bool) → Any

Generic value merging with conflict checking.

feedback_forensics.data.operations.merge._merge_dict(dict1: Dict, dict2: Dict, context: str, strict: bool) → Dict

Generic dictionary merging with optional strict conflict checking.

feedback_forensics.data.operations.merge._merge_prompt(comp1: Dict, comp2: Dict) → Optional[str]

Merge prompt strings from two comparisons, raising on conflicts.

feedback_forensics.data.operations.merge._merge_response(comp1: Dict, comp2: Dict, field_name: str) → Optional[Dict]

Merge response objects from two comparisons, raising on conflicts.

feedback_forensics.data.operations.merge._merge_annotations(comp1: Dict, comp2: Dict) → Dict

Merge annotation dictionaries from two comparisons, warning on conflicts.

feedback_forensics.data.operations.merge._merge_comparison_metadata(comp1: Dict, comp2: Dict) → Dict

Merge comparison-level metadata from two comparisons, detecting conflicts.

feedback_forensics.data.operations.merge._merge_annotators(ann1: Dict, ann2: Dict) → Dict

Merge annotator definitions, first takes precedence.

feedback_forensics.data.operations.merge._merge_name_or_description(value1: Optional[str], value2: Optional[str]) → Optional[str]

Merge name or description fields, reusing if identical or creating combined name.

feedback_forensics.data.operations.merge._merge_default_annotator(annotator1: Optional[str], annotator2: Optional[str]) → Optional[str]

Merge default annotator fields with first precedence and warning on disagreement.

feedback_forensics.data.operations.merge._merge_metadata(meta1: Dict, meta2: Dict) → Dict

Create merged metadata.