Collapse the openapi-namespace per-app reads into one canonical endpoint
GET /openapi/v1/apps/<id>/describe[?fields=info,parameters,input_schema]
returning a single AppDescribeResponse with all blocks Optional and a new
JSON-Schema input_schema block derived server-side from user_input_form +
app mode.
- AppDescribeQuery (Pydantic, extra=forbid) parses the ?fields allow-list;
unknown member -> 422.
- _input_schema.build_input_schema(app) derives Draft 2020-12 JSON Schema:
chat-family modes carry top-level query (string, minLength=1, required);
workflow / completion only carry inputs. AppUnavailableError -> empty
sentinel (EMPTY_INPUT_SCHEMA).
- Drop AppByIdApi (/apps/<id>) and AppParametersApi (/apps/<id>/parameters)
route classes; delete app_info.py module + app_info_payload helper.
- AppDescribeResponse.{info,parameters,input_schema} now Optional[None].
Lock-step deploy with difyctl Phase B (/describe consumer migration).
- Shared conftest at tests/integration_tests/controllers/openapi/:
workspace_account, app_in_workspace, mint_token (factory + tracked
cleanup of OAuthAccessToken rows), account_token convenience fixture,
autouse disable_enterprise (default ENTERPRISE_ENABLED=False; tests
needing the EE branch override in-test), autouse _flush_auth_redis.
- test_auth.py covers Layer 0 + per-token rate limit + scope on /info.
other_workspace_app fixture is a generator that cleans up the second
tenant + app on teardown.
- test_apps.py covers the read surface: /apps list with pagination
envelope, /apps/<id>, /info, /parameters, /describe, /account/sessions
envelope migration, plus dfoe_ scope rejection on apps:read routes.
The `ChatMessageApi` (`POST /console/api/apps/{app_id}/chat-messages`) and
`ModelConfigResource` (`POST /console/api/apps/{app_id}/model-config`)
endpoints do not properly validate user permissions, allowing users without `editor`
permission to access restricted functionality.
This PR addresses this issue by adding proper permission check.
This pull request introduces a feature aimed at improving the debugging experience during workflow editing. With the addition of variable persistence, the system will automatically retain the output variables from previously executed nodes. These persisted variables can then be reused when debugging subsequent nodes, eliminating the need for repetitive manual input.
By streamlining this aspect of the workflow, the feature minimizes user errors and significantly reduces debugging effort, offering a smoother and more efficient experience.
Key highlights of this change:
- Automatic persistence of output variables for executed nodes.
- Reuse of persisted variables to simplify input steps for nodes requiring them (e.g., `code`, `template`, `variable_assigner`).
- Enhanced debugging experience with reduced friction.
Closes#19735.
