[Bug]: SDK API DELETE /datasets/{id}/chunks does not actually stop parsing task

[tedhappy]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (Language Policy).
• Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• Please do not modify this template :) and fill in all the required fields.

RAGFlow workspace code commit ID

cfdcceb

RAGFlow image version

v0.22.1

Other environment information

- Hardware parameters: Windows 11
- OS type: Windows
- Others: Using SDK API via HTTP

Actual behavior

Bug Description

When calling the SDK API DELETE /api/v1/datasets/{dataset_id}/chunks to stop document parsing, the background task continues and eventually completes, overwriting the CANCEL status with DONE.

Root Cause Analysis

Looking at the source code:

SDK API ([api/apps/sdk/doc.py]:

info = {"run": "2", "progress": 0, "chunk_num": 0}
DocumentService.update_by_id(id, info)
settings.docStoreConn.delete({"doc_id": doc[0].id}, search.index_name(tenant_id), dataset_id)

The SDK API only updates the database status but does NOT call cancel_all_task_of() to send the cancellation signal via Redis.

### Expected behavior

The SDK API should call cancel_all_task_of(doc_id) to properly stop the background parsing task.

### Steps to reproduce

```Markdown
1. Upload a document to a dataset
2. Start parsing the document (status becomes RUNNING)
3. Call SDK API `DELETE /api/v1/datasets/{dataset_id}/chunks` with document_ids to stop parsing
4. Observe that the document parsing continues and eventually completes (status becomes DONE instead of CANCEL)

Additional information

Root Cause Analysis

The SDK API endpoint DELETE /datasets/{dataset_id}/chunks only updates database status but does NOT call cancel_all_task_of() to stop the background task.

SDK API (api/apps/sdk/doc.py line 842-844):

info = {"run": "2", "progress": 0, "chunk_num": 0}
DocumentService.update_by_id(id, info)
settings.docStoreConn.delete({"doc_id": doc[0].id}, search.index_name(tenant_id), dataset_id)

[dosubot]

Hi @tedhappy! I'm Dosu and I’m helping the ragflow team.

You’re correct about the root cause: the SDK API endpoint for DELETE /datasets/{id}/chunks only updates the document status and deletes chunks, but does not call cancel_all_task_of(), so background parsing tasks keep running and eventually overwrite the CANCEL status with DONE. Other endpoints, like /run in document_app.py, do call cancel_all_task_of() to send a cancellation signal via Redis, which properly stops background tasks source.

To fix this, the SDK API should invoke cancel_all_task_of(doc_id) for each document being stopped, just like the main document management API does. This will ensure the cancellation signal is sent and background workers can halt processing source.

There are no merged PRs or tracked tickets for this bug yet, so adding the cancellation call in the SDK endpoint is the recommended solution.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}