pengjunjie/ragflow
1
0
Fork 1
mirror of https://github.com/infiniflow/ragflow.git synced 2025-12-08 20:42:30 +08:00
Code Issues Packages Projects Releases Wiki Activity

Refa: improve flow of GraphRAG and RAPTOR (#10709)

Browse Source 
### What problem does this PR solve?

Improve flow of GraphRAG and RAPTOR.

### Type of change

- [x] Refactoring
This commit is contained in:
Yongteng Lei
2025-10-22 09:29:20 +08:00
committed by GitHub
parent acc0f7396e
commit 2d491188b8
3 changed files with 22 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
api/db/services/document_service.py
View File

@ -671,9 +671,11 @@ class DocumentService(CommonService):
@classmethod
@DB.connection_context()
def _sync_progress(cls, docs:list[dict]):
from api.db.services.task_service import TaskService
for d in docs:
try:
tsks = Task.query(doc_id=d["id"], order_by=Task.create_time)
tsks = TaskService.query(doc_id=d["id"], order_by=Task.create_time)
if not tsks:
continue
msg = []
@ -791,21 +793,23 @@ class DocumentService(CommonService):
"cancelled": int(cancelled),
}
def queue_raptor_o_graphrag_tasks(doc, ty, priority, fake_doc_id="", doc_ids=[]):
def queue_raptor_o_graphrag_tasks(sample_doc_id, ty, priority, fake_doc_id="", doc_ids=[]):
"""
You can provide a fake_doc_id to bypass the restriction of tasks at the knowledgebase level.
Optionally, specify a list of doc_ids to determine which documents participate in the task.
"""
chunking_config = DocumentService.get_chunking_config(doc["id"])
assert ty in ["graphrag", "raptor", "mindmap"], "type should be graphrag, raptor or mindmap"
chunking_config = DocumentService.get_chunking_config(sample_doc_id["id"])
hasher = xxhash.xxh64()
for field in sorted(chunking_config.keys()):
hasher.update(str(chunking_config[field]).encode("utf-8"))
def new_task():
nonlocal doc
nonlocal sample_doc_id
return {
"id": get_uuid(),
"doc_id": fake_doc_id if fake_doc_id else doc["id"],
"doc_id": sample_doc_id["id"],
"from_page": 100000000,
"to_page": 100000000,
"task_type": ty,
@ -820,9 +824,9 @@ def queue_raptor_o_graphrag_tasks(doc, ty, priority, fake_doc_id="", doc_ids=[])
task["digest"] = hasher.hexdigest()
bulk_insert_into_db(Task, [task], True)
if ty in ["graphrag", "raptor", "mindmap"]:
task["doc_ids"] = doc_ids
DocumentService.begin2parse(doc["id"])
task["doc_id"] = fake_doc_id
task["doc_ids"] = doc_ids
DocumentService.begin2parse(sample_doc_id["id"])
assert REDIS_CONN.queue_product(get_svr_queue_name(priority), message=task), "Can't access Redis. Please check the Redis' status."
return task["id"]