promptguard.utils

promptguard.utils.summarize_results(results)[source]

Summarize a list of analysis results.

Parameters:

results (List[RiskScore | None]) – List of RiskScore objects

Returns:

Dictionary with summary statistics

Return type:

Dict[str, Any]

promptguard.utils.filter_by_risk_level(results, risk_level)[source]

Filter results by risk level.

Parameters:

  • results (List[RiskScore | None]) – List of RiskScore objects

  • risk_level (str) – Risk level to filter by (‘low’, ‘medium’, ‘high’)

Returns:

Filtered list of results

Return type:

List[RiskScore]

promptguard.utils.get_most_dangerous(results, top_n=10)[source]

Get the most dangerous prompts from results.

Parameters:

  • results (List[RiskScore | None]) – List of RiskScore objects

  • top_n (int) – Number of top results to return

Returns:

List of top N most dangerous prompts

Return type:

List[RiskScore]

promptguard.utils.export_to_csv(results, prompts, filename)[source]

Export results to CSV file.

Parameters:

  • results (List[RiskScore]) – List of RiskScore objects

  • prompts (List[str]) – Original prompts

  • filename (str) – Output CSV filename

promptguard.utils.results_to_dataframe(results, prompts=None)[source]

Convert results to pandas DataFrame.