kb/create and /datasets

kb/create and /datasets

The kb/create used by the frontend and /datasets for creating datasets are essentially the same — they both call the same core service method to create datasets, but differ in interface design, parameter handling, and response format. kb_app.py:57-78 dataset.py:56-173

---

Detailed Analysis

Core Implementation is the Same

Both endpoints use the same core service method:

1. kb/create (Legacy endpoint):
   • Calls KnowledgebaseService.create_with_name() to create the dataset configuration kb_app.py:63-68
   • Uses KnowledgebaseService.save() to persist to the database kb_app.py:74-76
2. /datasets (SDK endpoint):
   • Also calls KnowledgebaseService.create_with_name() dataset.py:141-146
   • Also uses KnowledgebaseService.save() dataset.py:164-165

Key Differences

Feature	kb/create	/datasets
Route	/v1/kb/create kb_app.py:57	/api/v1/datasets dataset.py:56
Auth Method	@login_required kb_app.py:58	@token_required dataset.py:57
Param Validation	@validate_request("name") kb_app.py:59	validate_and_parse_json_request(request, CreateDatasetReq) dataset.py:120
Response Format	Returns {"kb_id": id} kb_app.py:76	Returns full dataset object processed by remap_dictionary_keys() dataset.py:169

Frontend Usage

The frontend uses these two endpoints in different scenarios:

1. Legacy pages use kb/create:
   • Calls kbService.createKb() inside the useCreateKnowledge hook use-knowledge-request.ts:187-188
   • Corresponds to the backend’s /v1/kb/create endpoint
2. SDK/API calls use /datasets:
   • Directly calls /api/v1/datasets in test code common.py:36-38

Unified Underlying Service

Both endpoints rely on the KnowledgebaseService.create_with_name() method, which is responsible for:

• Validating the dataset name
• Handling duplicate names
• Building the dataset configuration
• Setting the parser configuration knowledgebase_service.py:376-430

Conclusion

Although the two endpoints differ in interface design and usage scenarios, they are functionally equivalent — both serve as different entry points for creating datasets. /datasets is the more modern SDK interface, offering more complete parameter validation and response format, while kb/create is the legacy web interface primarily used for frontend page interaction.

Notes

• Both endpoints use the same database model and service layer, ensuring data consistency
• The frontend selects the appropriate endpoint based on the scenario, but the dataset object created is ultimately the same
• The /datasets endpoint supports more parameter configurations, such as advanced features like automatic metadata extraction dataset.py:123-140

References

• https://deepwiki.com/search/kbcreatedatasets_ae360aa1-ac58-402b-a7ad-1c4e31f5869f?mode=fast

Opens

1. current_user mean current login user, if multiple user login at same time, how to make sure the current user is desired user?