MAINT: Frontend core refactor#1753
Open
rlundeen2 wants to merge 43 commits into
Open
Conversation
…5_11_scenario_gaps
…ndeen/2026_05_13_add_initializer
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…5_13_add_initializer # Conflicts: # pyrit/backend/models/initializers.py # pyrit/backend/routes/initializers.py # pyrit/backend/services/initializer_service.py # tests/unit/backend/test_initializer_service.py
Create new pyrit/printer/ module with abstract base classes that contain all formatting logic. Data-fetching operations (CentralMemory calls) are abstract methods implemented by framework subclasses. This enables thin clients to reuse all pretty-printing by subclassing the base printers and implementing data-fetching via REST endpoints. The thin client only needs pyrit.models + pyrit.identifiers + colorama. Changes: - New pyrit/printer/ module with attack_result, scenario_result, scorer subpackages - ConsoleAttackPrinterBase: all attack console formatting, abstract get_conversation/get_scores - ConsoleScenarioPrinterBase: all scenario console formatting - ConsoleScorerPrinterBase: all scorer formatting, abstract get_objective/harm_metrics - Existing framework printers refactored to thin subclasses (backward compatible) - Added to_dict()/from_dict() to AttackResult, ScenarioResult, ScenarioIdentifier, ConversationReference, Score, MessagePiece, Message for serialization round-tripping - Message.to_full_dict() added for rich serialization (to_dict() unchanged for compat) All 675 existing tests pass with no modifications. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Move framework CentralMemory implementations into pyrit/printer/ alongside their base classes. CentralMemory is imported lazily inside constructors, so thin clients importing the module never pay the SQLAlchemy cost. - ConsoleAttackResultPrinter now lives in pyrit.printer.attack_result.console - ConsoleScenarioResultPrinter now lives in pyrit.printer.scenario_result.console - ConsoleScorerPrinter now lives in pyrit.printer.scorer.console - Old locations (executor/attack/printer/, scenario/printer/, score/printer/) become pure re-exports for backward compatibility - Updated test patch paths to match new module locations Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…6.0) Old locations now use PEP 562 __getattr__ lazy re-exports with DeprecationWarning. Only concrete classes are re-exported (not bases). - pyrit.executor.attack.printer → pyrit.printer.attack_result - pyrit.scenario.printer → pyrit.printer.scenario_result - pyrit.score.printer → pyrit.printer.scorer - Updated all internal callers to new canonical paths - Old ABC files (attack_result_printer.py, scenario_result_printer.py, scorer_printer.py) kept for now but deprecated via __init__.py Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…of bases Concrete classes that use CentralMemory/scorer registry are now named to clearly indicate their data source: - ConsoleAttackResultPrinter → ConsoleAttackMemoryPrinter - ConsoleScenarioResultPrinter → ConsoleScenarioMemoryPrinter - ConsoleScorerPrinter → ConsoleScorerMemoryPrinter Moved ScorerEvaluationIdentifier (pyrit internal) from base class into the concrete ConsoleScorerMemoryPrinter. Base classes now contain only formatting logic with no pyrit-internal imports beyond models/identifiers. Deprecated re-exports at old paths still work (mapping old names to new), scheduled for removal in 0.16.0. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Created MarkdownAttackPrinterBase + MarkdownAttackMemoryPrinter in pyrit/printer/attack_result/markdown.py (same pattern as console) - Deleted dead old ABC files: - pyrit/executor/attack/printer/attack_result_printer.py - pyrit/scenario/printer/scenario_result_printer.py - pyrit/score/printer/scorer_printer.py - Old markdown_printer.py now a deprecation re-export shim - Updated all internal imports and test patches Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Changed dict[str, object] to dict[str, Any] in MessagePiece.from_dict() and Message.from_dict() to satisfy pyright (dict.get returns object otherwise) - Added Any import to message_piece.py and message.py - Wrapped get_prompt_scores return in list() for Sequence -> list coercion - Added isinstance check in display_image_async for type safety Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Added return type annotation (-> type) to all __getattr__ deprecation shims - Added noqa: B027 to display_image_async intentional no-op default - Added Returns/Raises sections to short docstrings (DOC201, DOC501) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
objective_target_identifier and objective_scorer_identifier may be None when deserializing from dicts. The printer bases already handle None. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…5_13_frontend_core_refactor # Conflicts: # .pyrit_conf_example # pyrit/backend/models/initializers.py # pyrit/backend/routes/initializers.py # pyrit/backend/services/initializer_service.py # pyrit/cli/frontend_core.py # pyrit/cli/pyrit_backend.py # pyrit/registry/class_registries/initializer_registry.py # pyrit/setup/configuration_loader.py # tests/unit/backend/test_initializer_service.py # tests/unit/cli/test_pyrit_backend.py # tests/unit/registry/test_initializer_registry.py
…rlundeen/2026_05_13_frontend_core_refactor # Conflicts: # pyrit/cli/frontend_core.py # pyrit/cli/pyrit_shell.py
…rlundeen/2026_05_13_frontend_core_refactor
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…ctor Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
rlundeen2
commented
May 18, 2026
| }, | ||
| ) | ||
| async def get_scenario_run_results(scenario_result_id: str) -> ScenarioRunDetail: | ||
| async def get_scenario_run_results(scenario_result_id: str) -> dict: |
Contributor
Author
There was a problem hiding this comment.
I have a plan for this I want to propose; but I think this is okay for this PR
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
rlundeen2
commented
May 19, 2026
| # --------------------------------------------------------------------------- | ||
|
|
||
|
|
||
| def print_scenario_list(*, items: list[dict[str, Any]]) -> None: |
Contributor
Author
There was a problem hiding this comment.
Similarly, I feel like dict is not ideal; I think a good solution is to go towards a shared model
rlundeen2
commented
May 19, 2026
| Returns: | ||
| dict[str, Any]: Serialized payload suitable for REST APIs or persistence. | ||
| """ | ||
| from pyrit.models.conversation_reference import ConversationReference |
Contributor
Author
There was a problem hiding this comment.
there is another pr for this that is a dependency; see #1738
…5_13_frontend_core_refactor # Conflicts: # pyrit/cli/pyrit_shell.py # tests/unit/cli/test_pyrit_shell.py
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
… CLI
Fixes the failing `docker_build/Test GUI (local)` CI check and brings
diff coverage above the 90% threshold for the frontend refactor.
GUI / runtime fixes
- `docker/start.sh` and `frontend/dev.py`: update stale references to
the deleted `pyrit.cli.pyrit_backend` module (now
`pyrit.backend.pyrit_backend`).
- `doc/code/scenarios/0_scenarios.{py,ipynb}` and
`2_custom_scenario_parameters.{py,ipynb}`: replace deleted
`pyrit.cli.frontend_core` imports with `get_scenario_service` plus
`print_scenario_list`.
- `doc/myst.yml`: drop dead API doc entries for the removed modules and
add entries for the new `api_client` and `cli_helpers` modules.
Test coverage
- New unit tests:
- `tests/unit/cli/test_output.py` (printer functions)
- `tests/unit/cli/test_api_client.py` (PyRITApiClient + error paths)
- `tests/unit/cli/test_server_launcher.py` (probe/start/stop)
- `tests/unit/cli/test_config_reader.py` (layered YAML loading)
- Extended existing unit tests for `pyrit_backend`, `backend/main`,
`pyrit_scan`, `pyrit_shell`, and `configuration_loader`.
Diff coverage on the PR is now 96%.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…hell) The frontend refactor regressed scenario-declared parameter handling on both CLI entry points - flags like `--max-turns 7` were rejected before the second pass could fetch metadata from the server. pyrit_scan - `parse_args` now uses `parse_known_args` (pass 1 is tolerant of scenario-specific flags) and stashes leftovers + the raw arg list on the Namespace for the second pass. - `_reparse_with_scenario_params` threads the original `args` list instead of reading `sys.argv[1:]`, so explicit-args callers work. - `main` does a strict re-parse when there are unknown args but no scenario is specified, preserving the original `exit 2` error for truly invalid flags. pyrit_shell - `do_run` now fetches `get_scenario_async` first, builds `Parameter` objects from the response's `supported_parameters`, and threads `declared_params` into `parse_run_arguments`. - Calls `extract_scenario_args` and propagates `scenario_params` on the REST request payload. Tests - New `TestScenarioParamFlow` (4 tests) and `TestShellScenarioParamFlow` (4 tests) regression cases covering forward, invalid-flag, no-params, and metadata-fetch failure paths. Diff coverage remains at 96%. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…er config Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR removes the custom logic for the front-end, using the REST APIs instead of custom CLI logic.
In future PRs: