results

results

Classes:

Name	Description
AnonymizerResult	Result returned by full anonymization runs.
PreviewResult	Result returned by preview runs.

AnonymizerResult(dataframe, trace_dataframe, resolved_text_column, failed_records, replace_method=None, rewrite_config=None, entity_labels=None, data_summary=None, excluded_entity_labels=None) dataclass

Bases: _DisplayMixin

Result returned by full anonymization runs.

Attributes:

Name	Type	Description
dataframe	DataFrame	User-facing columns only (text, replaced/rewritten text, scores).
trace_dataframe	DataFrame	Full pipeline trace including all internal columns.
resolved_text_column	str	Name of the user-facing text column. Equals the user's requested text_column unless the reader had to rename it to avoid colliding with an Anonymizer output column, in which case it is the post-rename identifier (e.g. "final_entities__input").
failed_records	list[FailedRecord]	Records that failed during pipeline processing.
replace_method	ReplaceMethod | None	The replace strategy that produced this result. Set by run() / preview(); consumed by evaluate() to dispatch the right judges. None on results that were constructed by hand or loaded from a pre-strategy-tracking format.
rewrite_config	PrivacyGoal | None	The privacy goal that produced this result when rewrite mode was used. Set by run() / preview(); consumed by evaluate() to dispatch the rewrite judges. Mutually exclusive with replace_method.
entity_labels	list[str] | None	Allowlist of entity labels that were in scope during detection. Preserved for evaluate() so the coverage judge scopes its evaluation to the same label set. None means all default labels were in scope.
data_summary	str | None	Optional dataset context supplied with the original input. Preserved for evaluate() so entity-coverage judging uses the same context as detection.
excluded_entity_labels	list[str] | None	Labels that were explicitly excluded from detection. Preserved for evaluate() so the coverage judge does not penalise the output for not anonymizing excluded labels.

PreviewResult(dataframe, trace_dataframe, resolved_text_column, failed_records, preview_num_records, replace_method=None, rewrite_config=None, entity_labels=None, data_summary=None, excluded_entity_labels=None) dataclass

Bases: _DisplayMixin

Result returned by preview runs.

Attributes:

Name	Type	Description
dataframe	DataFrame	User-facing columns only (text, replaced/rewritten text, scores).
trace_dataframe	DataFrame	Full pipeline trace including all internal columns.
resolved_text_column	str	Name of the user-facing text column. Equals the user's requested text_column unless the reader had to rename it to avoid colliding with an Anonymizer output column, in which case it is the post-rename identifier (e.g. "final_entities__input").
failed_records	list[FailedRecord]	Records that failed during pipeline processing.
preview_num_records	int	Number of records requested for the preview.
replace_method	ReplaceMethod | None	The replace strategy that produced this preview. Set by preview(); consumed by evaluate() to dispatch the right judges. None on results that were constructed by hand or loaded from a pre-strategy-tracking format.
rewrite_config	PrivacyGoal | None	The privacy goal that produced this preview when rewrite mode was used. Set by preview(); consumed by evaluate() to dispatch the rewrite judges. Mutually exclusive with replace_method.
entity_labels	list[str] | None	Allowlist of entity labels that were in scope during detection. Preserved for evaluate() so the coverage judge scopes its evaluation to the same label set. None means all default labels were in scope.
data_summary	str | None	Optional dataset context supplied with the original input. Preserved for evaluate() so entity-coverage judging uses the same context as detection.
excluded_entity_labels	list[str] | None	Labels that were explicitly excluded from detection. Preserved for evaluate() so the coverage judge does not penalise the output for not anonymizing excluded labels.