Rapid response harm scenario

doc/code/scenarios/rapid_response_harm_dataset_loading.md

@@ -0,0 +1,101 @@
+# Preloading Datasets for AI RT Scenarios
+
+## Overview
+
+The scenarios in the ai_rt folder class that test datasets be preloaded into PyRIT's `CentralMemory` before running any scenarios. This design ensures that:
+
+1. **Test data is centralized**: All prompts and objectives are stored in a consistent location
+2. **Scenarios are reusable**: Multiple scenarios can share the same datasets
+3. **Data is retrievable**: The scenario can automatically find and load the correct prompts based on strategy names
+4. **Memory is isolated**: Different test runs can use different memory instances (e.g., IN_MEMORY vs SQLite)
+
+## Dataset Naming Schema
+
+The naming schema is **critical** for these scenarios to automatically retrieve the correct datasets. The schema follows this pattern:
+
+```
+<dataset_path_prefix><strategy_name>
+```
+
+### Components
+
+1. **Dataset Path Prefix** (default: <scenario_name>):
+   - Can be customized via the `objective_dataset_path` parameter in the scenario constructor
+   - Helps organize datasets in memory when multiple scenario types are being used
+
+2. **Strategy Name** (required):
+   - Derived from the strategy enum value
+   - Converted to lowercase with underscores (e.g., `HateFictionalStory` → `hate_fictional_story`)
+   - Must match exactly for the scenario to find the dataset
+
+### Default Naming Examples for Rapid Response Harm Scenario
+
+| Strategy Enum | Dataset Name |
+|--------------|--------------|
+| `RapidResponseHarmStrategy.HateFictionalStory` | `rapid_response_harm_hate_fictional_story` |
+| `RapidResponseHarmStrategy.FairnessEthnicityInference` | `rapid_response_harm_fairness_ethnicity_inference` |
+| `RapidResponseHarmStrategy.ViolenceCivic` | `rapid_response_harm_violence_civic` |
+| `RapidResponseHarmStrategy.ViolenceProtestDisruption` | `rapid_response_harm_violence_protest_disruption` |
+| `RapidResponseHarmStrategy.SexualContent` | `rapid_response_harm_sexual_content` |
+| `RapidResponseHarmStrategy.HarassmentBullying` | `rapid_response_harm_harassment_bullying` |
+| `RapidResponseHarmStrategy.MisinformationElection` | `rapid_response_harm_misinformation_election` |
+| `RapidResponseHarmStrategy.LeakagePersonalData` | `rapid_response_harm_leakage_personal_data` |
+
+### Custom Dataset Path Prefix
+
+You can customize the prefix when creating a scenario:
+
+```python
+scenario = RapidResponseHarmScenario(
+    objective_target=my_target,
+    adversarial_chat=adversarial_target,
+    objective_dataset_path="custom_test_",  # Custom prefix
+    scenario_strategies=[RapidResponseHarmStrategy.HateFictionalStory]
+)
+
+# Now the dataset name must be: "custom_test_hate_fictional_story"
+```
+
+
+## Common Errors and Solutions
+
+### Error: "No objectives found in the dataset"
+
+**Cause**: The dataset wasn't loaded into memory or the naming doesn't match.
+
+**Solution**:
+1. Verify the dataset name matches the strategy name exactly
+2. Ensure you called `add_seed_groups_to_memory()` before running the scenario
+3. Check that the dataset includes a `SeedObjective` object
+
+```python
+# Correct naming
+dataset_name = "rapid_response_harm_" + strategy.value  # e.g., "hate_fictional_story"
+```
+
+### Error: Dataset not found for custom prefix
+
+**Cause**: The scenario's `objective_dataset_path` doesn't match the dataset names in memory.
+
+**Solution**: Ensure consistency between the scenario configuration and dataset names:
+
+```python
+# Scenario configuration
+scenario = RapidResponseHarmScenario(
+    objective_target=target,
+    adversarial_chat=adversarial,
+    objective_dataset_path="my_custom_prefix_"  # Must match dataset names
+)
+
+# Dataset must be named: "my_custom_prefix_hate_fictional_story"
+await create_seed_dataset(
+    name="my_custom_prefix_hate_fictional_story",
+    prompts=[...],
+    objective="..."
+)
+```
+
+## Additional Resources
+
+- See `rapid_response_harm_scenario.ipynb` for a complete working example
+- Check the `RapidResponseHarmStrategy` enum for all available strategies