feat: add GET /projects/{id}/ontology/index-status endpoint#76
Merged
Conversation
The frontend already calls this endpoint to display index status in project settings and poll after reindex, but it returned 404 because only the POST /reindex route existed. Closes #48. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
Warning Rate limit exceeded
Your organization is not enrolled in usage-based pricing. Contact your admin to enable usage-based pricing to continue reviews beyond the rate limit, or try again in 4 minutes and 29 seconds. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (3)
📝 WalkthroughWalkthroughAdded DI factories for index/change/embedding services; implemented GET /{project_id}/ontology/index-status; added an ontology index WebSocket /{project_id}/ontology/index-ws with token auth, Redis pub/sub, and per-project connection manager; introduced get_git_ontology_path helper; updated embedding/worker fallbacks; and added unit tests for DI, index endpoints, websockets, embedding, and helpers. Changes
Sequence Diagram(s)sequenceDiagram
participant Client as Frontend Client
participant API as Projects API
participant ProjectSvc as ProjectService/GitService
participant IndexSvc as OntologyIndexService
participant DB as Database
Client->>API: GET /projects/{id}/ontology/index-status?branch=...
API->>ProjectSvc: get(project_id)
ProjectSvc->>DB: SELECT Project
DB-->>ProjectSvc: Project
alt branch not provided
API->>ProjectSvc: get_default_branch(project_id)
ProjectSvc-->>API: branch
end
API->>IndexSvc: get_index_status(project_id, branch)
IndexSvc->>DB: SELECT OntologyIndexStatus
DB-->>IndexSvc: record | null
IndexSvc-->>API: record | None
alt record found
API-->>Client: 200 {status, entity_count, indexed_at, commit_hash, error_message}
else no record
API-->>Client: 404 {"detail":"No index status found"}
end
sequenceDiagram
participant Client as Frontend (WS)
participant WS as API WebSocket Handler
participant Manager as IndexConnectionManager
participant Redis as Redis Pub/Sub
participant Pub as Redis Pubsub Loop
participant DB as Database
Client->>WS: Open WS /projects/{id}/ontology/index-ws?token=...
WS->>WS: authenticate_ws(websocket, project_id, token)
alt auth & access ok
WS->>Manager: connect(project_id, websocket)
Manager-->>WS: registered
WS->>Redis: subscribe "ontology_index:updates"
loop poll pubsub
Pub->>WS: message JSON
alt message.project_id == connected project_id
WS->>Client: send_json(payload)
else
Note right of WS: ignore message
end
end
else auth failed
WS-->>Client: close(code=4001/4003/4004/1011)
end
Note right of WS: on disconnect/error -> unsubscribe, Manager.disconnect()
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related issues
Possibly related PRs
Suggested labels
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 3 | ❌ 2❌ Failed checks (2 warnings)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Codecov Report❌ Patch coverage is
📢 Thoughts on this report? Let us know! |
…erage Rewrite tests to use authed_client fixture and FastAPI dependency overrides, asserting actual response status codes, JSON fields, and service call arguments. Covers success, 404, branch param, default branch fallback, and failed status scenarios. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The worker and embedding service used getattr(project, "git_ontology_path", None) which always returned None since that field only exists on ProjectResponse, not the Project ORM model. This caused indexing/embedding to look for "ontology.ttl" instead of the actual file path (e.g. "sources/ontology-semantic-canon.ttl"). Extract get_git_ontology_path() into ontokit/models/project.py and eagerly load the github_integration relationship so the correct file path is resolved. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@ontokit/api/routes/projects.py`:
- Around line 1423-1432: The route currently instantiates OntologyIndexService
directly; change it to accept the service via FastAPI DI by adding a parameter
like index_service: OntologyIndexService = Depends(get_ontology_index_service)
(or a factory that itself depends on AsyncSession) and remove the direct
construction in the handler; then call
index_service.get_index_status(project.id, resolved_branch) as before. Ensure
the get_ontology_index_service provider (or factory) injects the AsyncSession
(get_db) so overrides/mock wiring will work in tests.
In `@ontokit/services/embedding_service.py`:
- Around line 260-264: The code rejects projects early by checking
Project.source_file_path before attempting git-backed resolution; update the
logic around the proj_result handling so that if Project.source_file_path is
falsy you call get_git_ontology_path(project) and/or attempt load_from_git(...)
to derive a path and only raise the error if both the explicit source_file_path
and the git-derived path are missing; apply the same change to the second
occurrence noted (around the later source_file_path check) and reference
Project.source_file_path, get_git_ontology_path, and load_from_git in the
updated flow.
In `@ontokit/worker.py`:
- Around line 65-70: The current eager project load in worker.py checks Project
and its github_integration but still requires source_file_path before attempting
git-based index loading; update the logic in the project-loading/indexing path
(look for select(Project).options(selectinload(Project.github_integration)) and
any subsequent checks of source_file_path) to only enforce source_file_path when
performing file-system or import-specific indexing, and instead branch to use
Project.github_integration metadata when present to drive git-based indexing;
ensure the code path that handles github_integration is reached even if
source_file_path is None and add a clear conditional that prefers
github_integration when available.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 2dd413b5-80cb-4b80-bc71-0688f55d4287
📒 Files selected for processing (5)
ontokit/api/routes/projects.pyontokit/models/project.pyontokit/services/embedding_service.pyontokit/worker.pytests/unit/test_index_status_route.py
Add WS /projects/{id}/ontology/index-ws that forwards index_started,
index_complete, and index_failed events from the background worker via
Redis pubsub, mirroring the existing lint WebSocket pattern.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Use FastAPI DI for OntologyIndexService in index-status route via get_index dependency, replacing direct construction. Tests updated to use dependency overrides instead of @patch. - Relax source_file_path guard in worker and embedding service: allow projects with github_integration but no source_file_path to proceed with git-based loading via get_git_ontology_path(). - Storage fallback in embedding service now checks source_file_path before attempting load_from_storage. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace direct Service(db) construction with Depends() factories in all route handlers for consistency and testability: - EmbeddingService: 7 instances across embeddings.py and semantic_search.py - ChangeEventService: 5 instances across analytics.py and projects.py - OntologyIndexService: 1 remaining instance in delete_branch handler All route-level services now use the same DI pattern. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- 6 tests for get_git_ontology_path() covering all branches - 7 tests for IndexConnectionManager and WebSocket endpoint (4004 close) - 6 tests for DI factory functions across routes - 1 test for worker no-git-no-storage error path - 4 tests for embedding service relaxed source_file_path guard 24 new tests, bringing total to 1366. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add async unit tests for ontology_index_websocket function covering: - Project not found (closes with 4004) - Successful connection, message forwarding for matching project - Skipping messages for other projects - Malformed JSON handling (skip and continue) - Cleanup on unexpected exceptions (disconnect + unsubscribe) Brings projects.py coverage from 88% to 94%. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@ontokit/api/routes/projects.py`:
- Around line 1533-1553: The WebSocket handler ontology_index_websocket
currently only checks project existence and allows unauthenticated access;
change it to require and validate the authenticated user via FastAPI dependency
injection and ProjectService.get before accepting the socket: add a dependency
parameter for the current user (e.g., user: User = Depends(get_current_user))
and replace the manual db lookup with a call to await
ProjectService.get(project_id, user) (or call it after fetching to enforce
access), and if ProjectService.get raises NotFound/Unauthorized or returns None
then close the websocket with an appropriate code/reason instead of calling
index_ws_manager.connect; finally only call index_ws_manager.connect(websocket,
project_id_str) after ProjectService.get confirms access.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: b3d71484-6e53-47ab-ac74-2db6d19d0afd
📒 Files selected for processing (12)
ontokit/api/routes/analytics.pyontokit/api/routes/embeddings.pyontokit/api/routes/projects.pyontokit/api/routes/semantic_search.pyontokit/services/embedding_service.pyontokit/worker.pytests/unit/test_di_factories.pytests/unit/test_embedding_service.pytests/unit/test_get_git_ontology_path.pytests/unit/test_index_status_route.pytests/unit/test_index_ws.pytests/unit/test_worker.py
✅ Files skipped from review due to trivial changes (1)
- tests/unit/test_index_status_route.py
🚧 Files skipped from review as they are similar to previous changes (1)
- ontokit/services/embedding_service.py
The WebSocket endpoint now requires a `token` query parameter for authentication. The token is validated via Zitadel JWT, and project access is verified through ProjectService.get before accepting the connection. Unauthenticated requests are closed with 4001, and unauthorized/not-found with 4004. Tests updated with auth mocking for all WebSocket test cases. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Apply the same authentication pattern as the index WebSocket: require a token query parameter, validate via Zitadel JWT, and verify project access through ProjectService.get before accepting the connection. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
7 tests covering: no token (4001), invalid token (4001), project not found (4004), message forwarding, project filtering, malformed JSON, and cleanup on exception. Brings lint.py coverage from 72% to 96%. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Member
Author
|
@CodeRabbit review |
✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (2)
ontokit/api/routes/lint.py (1)
550-583: Extract the WebSocket auth/access gate into a shared helper.This block is now duplicated almost verbatim here and in
ontokit/api/routes/projects.py, and it opens its own session / constructsProjectServiceinside the route. Centralizing it behind a reusable helper or dependency will keep the two sockets from drifting and restore the API-layer DI pattern.As per coding guidelines
ontokit/api/**/*.py: "Use dependency injection via FastAPI'sDepends()for service access".🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@ontokit/api/routes/lint.py` around lines 550 - 583, The WebSocket auth and project-access logic in this route is duplicated and should be extracted into a shared helper/dependency; create a reusable async helper (e.g., websocket_auth_and_authorize) that encapsulates token validation (validate_token), userinfo fetching (fetch_userinfo), CurrentUser construction, opening a DB session via async_session_maker, and project access check using ProjectService.get, and have the route call/await that helper instead of duplicating the block and constructing ProjectService inline; ensure the helper returns the CurrentUser and the opened DB/session or raises/returns appropriate errors so the route can close the websocket with the same codes/reasons used here (4001 for auth failures) and keep existing semantics.ontokit/api/routes/projects.py (1)
1506-1506: Import the channel name fromontokit.workerinstead of re-declaring it.The worker already publishes on
ONTOLOGY_INDEX_UPDATES_CHANNEL. Duplicating the raw string here means the publisher and subscriber can silently drift apart on the next rename.Suggested fix
+from ontokit.worker import ONTOLOGY_INDEX_UPDATES_CHANNEL ... -ONTOLOGY_INDEX_UPDATES_CHANNEL = "ontology_index:updates"🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@ontokit/api/routes/projects.py` at line 1506, The constant ONTOLOGY_INDEX_UPDATES_CHANNEL is being redeclared in projects.py which can drift from the publisher; remove the local declaration and import ONTOLOGY_INDEX_UPDATES_CHANNEL from ontokit.worker (where the worker publishes) and update any references in this file to use the imported symbol so publisher and subscriber use the same canonical channel name.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@ontokit/api/routes/lint.py`:
- Around line 560-585: The current blanket except Exception around
validate_token/fetch_userinfo/CurrentUser construction and around
ProjectService.get collapses server-side failures into client errors; change
these to explicitly catch the framework's HTTPException for expected auth/access
failures (from validate_token and ProjectService.get) and handle those by
closing the websocket with the existing client-facing codes (e.g., keep the 4001
for auth failures and 4004 for access failures), but let other exceptions
propagate or catch them separately and close the websocket with an
internal-error code (e.g., 1011) and/or log the full exception; update the
try/except blocks around validate_token/fetch_userinfo/CurrentUser and the
async_session_maker()/ProjectService.get call to reflect this separation.
In `@ontokit/api/routes/projects.py`:
- Around line 1432-1445: The route handler get_ontology_index_status is using
RequiredUser which prevents anonymous viewers from accessing public project
index status; change the user parameter type to OptionalUser so visibility rules
in ProjectService.get(project_id, user) can apply and allow read-only access for
viewers/anonymous users. Update the function signature to accept user:
OptionalUser (and add or adjust the import for OptionalUser if missing), keep
calling service.get(project_id, user) and leave all other logic
(resolved_branch, index_service.get_index_status) unchanged.
---
Nitpick comments:
In `@ontokit/api/routes/lint.py`:
- Around line 550-583: The WebSocket auth and project-access logic in this route
is duplicated and should be extracted into a shared helper/dependency; create a
reusable async helper (e.g., websocket_auth_and_authorize) that encapsulates
token validation (validate_token), userinfo fetching (fetch_userinfo),
CurrentUser construction, opening a DB session via async_session_maker, and
project access check using ProjectService.get, and have the route call/await
that helper instead of duplicating the block and constructing ProjectService
inline; ensure the helper returns the CurrentUser and the opened DB/session or
raises/returns appropriate errors so the route can close the websocket with the
same codes/reasons used here (4001 for auth failures) and keep existing
semantics.
In `@ontokit/api/routes/projects.py`:
- Line 1506: The constant ONTOLOGY_INDEX_UPDATES_CHANNEL is being redeclared in
projects.py which can drift from the publisher; remove the local declaration and
import ONTOLOGY_INDEX_UPDATES_CHANNEL from ontokit.worker (where the worker
publishes) and update any references in this file to use the imported symbol so
publisher and subscriber use the same canonical channel name.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 68053c22-15c7-4475-b8db-91a810937a83
📒 Files selected for processing (4)
ontokit/api/routes/lint.pyontokit/api/routes/projects.pytests/unit/test_index_ws.pytests/unit/test_lint_ws.py
✅ Files skipped from review due to trivial changes (1)
- tests/unit/test_index_ws.py
…nonymous index-status Address CodeRabbit review findings: - Extract duplicated WebSocket auth logic into shared authenticate_ws() helper in ontokit/api/utils/ws_auth.py. Distinguishes HTTPException (4001/4003/4004) from unexpected errors (1011 with logging). - Move Redis pubsub channel constants from worker.py to core/constants.py. Both publisher (worker) and subscribers (lint.py, projects.py) now import from the same canonical source. - Change index-status route from RequiredUser to OptionalUser so public projects are accessible to anonymous/viewer users. - Add 7 tests for authenticate_ws covering all error paths and success. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Starlette returns HTTP 403 if websocket.close() is called before websocket.accept(). Move accept() into authenticate_ws() so error close frames are delivered properly to clients. Remove duplicate accept() from connection managers. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Member
Author
|
@coderabbitai review |
✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (2)
ontokit/api/routes/projects.py (1)
1259-1410: InjectEmbeddingServiceviaDepends()for consistency.Line 1388 instantiates
EmbeddingService(db)directly instead of via FastAPI'sDepends(). The route already injectsChangeEventServicethis way, so the auto-embed path should follow the same pattern for consistency and testability.<summary>Line 1388 example:</summary> embed_config = await EmbeddingService(db).get_config(project_id)Per coding guideline
ontokit/api/**/*.py: "Use dependency injection via FastAPI'sDepends()for service access".🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@ontokit/api/routes/projects.py` around lines 1259 - 1410, The route directly constructs EmbeddingService(db) (see embed_config = await EmbeddingService(db).get_config(project_id)); change this to inject EmbeddingService via FastAPI Depends by adding a parameter like embedding_service: Annotated[EmbeddingService, Depends(get_embedding_service)] to save_source_content and replace the direct instantiation with embed_config = await embedding_service.get_config(project_id); ensure you import or use the existing provider function (get_embedding_service) consistent with other injected services and update any references to EmbeddingService(db) in this function to use the injected embedding_service instead.ontokit/api/utils/ws_auth.py (1)
72-75: InjectProjectServicevia dependency injection instead of constructing it directly.Opening a fresh session and building
ProjectServiceinside the helper bypasses the API-layer DI pattern established elsewhere. Have the websocket route inject the service viaDepends()and pass it intoauthenticate_ws(...)instead, matching the HTTP auth pattern in the codebase. This improves testability and consistency with the guideline forontokit/api/**/*.py: "Use dependency injection via FastAPI'sDepends()for service access".🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@ontokit/api/utils/ws_auth.py` around lines 72 - 75, The helper authenticate_ws currently opens a fresh session and constructs ProjectService via async_session_maker/ProjectService; change authenticate_ws to accept a ProjectService instance injected from the FastAPI route (use Depends(ProjectService) in the websocket route) and remove the local async_session_maker/ProjectService construction; inside authenticate_ws call the injected svc.get(project_id, user) instead. Update the websocket route to declare a parameter svc: ProjectService = Depends(...) and pass that svc into authenticate_ws so the function follows the API-layer DI pattern and becomes testable.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@ontokit/api/routes/projects.py`:
- Around line 1553-1584: The realtime setup can fail
(get_arq_pool()/pubsub/subscribe) after authenticate_ws has accepted the
connection, so ensure the WebSocket is deterministically closed with a close
code and reason whenever setup or the pubsub loop raises before normal
operation: in the try/except around get_arq_pool(), pool.pubsub(), and await
pubsub.subscribe(...) (and in the top-level except Exception e block) call await
websocket.close(code=1011, reason=f"realtime setup failed: {e}" or a static
reason when pool is None) before letting the function proceed to finally; keep
index_ws_manager.disconnect(websocket, project_id_str) and the
pubsub.unsubscribe logic in finally so disconnection is still tracked and
resources cleaned up.
In `@tests/unit/test_ws_auth.py`:
- Around line 29-152: Add assertions that the WebSocket was accepted before
being closed (or left open) in each test that calls authenticate_ws: for tests
test_no_token_closes_4001, test_invalid_token_http_exception_closes_4001,
test_unexpected_auth_error_closes_1011, test_project_not_found_closes_4004,
test_access_denied_closes_4003, test_unexpected_project_error_closes_1011, and
test_success_returns_true, assert that the AsyncMock WebSocket's accept() was
awaited (e.g., ws.accept.assert_awaited_once() or assert_awaited) prior to
checking close behavior so the test verifies the helper calls ws.accept() before
closing or leaving the socket open.
---
Nitpick comments:
In `@ontokit/api/routes/projects.py`:
- Around line 1259-1410: The route directly constructs EmbeddingService(db) (see
embed_config = await EmbeddingService(db).get_config(project_id)); change this
to inject EmbeddingService via FastAPI Depends by adding a parameter like
embedding_service: Annotated[EmbeddingService, Depends(get_embedding_service)]
to save_source_content and replace the direct instantiation with embed_config =
await embedding_service.get_config(project_id); ensure you import or use the
existing provider function (get_embedding_service) consistent with other
injected services and update any references to EmbeddingService(db) in this
function to use the injected embedding_service instead.
In `@ontokit/api/utils/ws_auth.py`:
- Around line 72-75: The helper authenticate_ws currently opens a fresh session
and constructs ProjectService via async_session_maker/ProjectService; change
authenticate_ws to accept a ProjectService instance injected from the FastAPI
route (use Depends(ProjectService) in the websocket route) and remove the local
async_session_maker/ProjectService construction; inside authenticate_ws call the
injected svc.get(project_id, user) instead. Update the websocket route to
declare a parameter svc: ProjectService = Depends(...) and pass that svc into
authenticate_ws so the function follows the API-layer DI pattern and becomes
testable.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: e380fada-93a8-42ee-ae8c-43bf260c3242
📒 Files selected for processing (9)
ontokit/api/routes/lint.pyontokit/api/routes/projects.pyontokit/api/utils/ws_auth.pyontokit/core/constants.pyontokit/worker.pytests/unit/test_index_ws.pytests/unit/test_lint_routes_extended.pytests/unit/test_lint_ws.pytests/unit/test_ws_auth.py
✅ Files skipped from review due to trivial changes (2)
- ontokit/core/constants.py
- tests/unit/test_lint_ws.py
🚧 Files skipped from review as they are similar to previous changes (1)
- ontokit/api/routes/lint.py
… EmbeddingService Address CodeRabbit findings: - Close WebSocket with 1011 when pubsub setup fails (both lint and index endpoints), preventing clients from hanging on setup errors. - Add ws.accept.assert_awaited_once() to all authenticate_ws tests, verifying the WebSocket is accepted before any close. - Replace last direct EmbeddingService(db) in save_source_content with DI via Depends(get_embeddings). Remove now-unused db parameter. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
6 tasks
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
1 participant
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.
Summary
This PR started as a fix for issue #48 (missing GET index-status endpoint) but expanded to address several related improvements discovered during implementation and code review.
New endpoints
/projects/{id}/ontology/index-status— returns index status fields (status,entity_count,indexed_at,commit_hash,error_message); 404 if no record exists/projects/{id}/ontology/index-ws— real-time WebSocket forindex_started,index_complete,index_failedevents via Redis pubsub (mirrors lint WS pattern)Bug fixes
getattr(project, "git_ontology_path", None)which always returnedNone(field only exists onProjectResponse, not the ORM model). Extractedget_git_ontology_path()intomodels/project.pyand eagerly loadgithub_integrationrelationshipsource_file_pathguard — worker and embedding service now allow projects withgithub_integrationbut nosource_file_pathto proceed with git-based loadingtokenquery parameter, validated via Zitadel JWT +ProjectService.getaccess check (previously unauthenticated)Refactoring
Service(db)construction withDepends()factories:OntologyIndexService(2 instances inprojects.py)EmbeddingService(7 instances acrossembeddings.py,semantic_search.py)ChangeEventService(5 instances acrossanalytics.py,projects.py)Tests
get_git_ontology_path, worker/embedding edge casesCloses #48
Test plan
🤖 Generated with Claude Code
Summary by CodeRabbit
New Features
Refactoring
Tests