[Question]: Turn on the Reasoning function and an error will be reported during the chat.

[zhouyangit]

Describe your problem

I used the latest v0.17.0 version. In the assistant, after turning on Reasoning, an error is reported every time during the chat.

2025-03-03 17:52:43,762 INFO     19 [THINK]Query: 0. 帮我介绍一下苹果
Traceback (most recent call last):
  File "/ragflow/api/apps/conversation_app.py", line 231, in stream
    for ans in chat(dia, msg, True, **req):
  File "/ragflow/api/db/services/dialog_service.py", line 193, in chat
    for think in reasoner.thinking(kbinfos, " ".join(questions)):
  File "/ragflow/agentic_reasoning/deep_research.py", line 117, in thinking
    kbinfos = self._kb_retrieve(question=search_query) if self._kb_retrieve else {"chunks": [], "doc_aggs": []}
  File "<@beartype(rag.nlp.search.Dealer.retrieval) at 0x75ce1e9d6200>", line 35, in retrieval
  File "/ragflow/rag/nlp/search.py", line 371, in retrieval
    sim, tsim, vsim = self.rerank(
  File "<@beartype(rag.nlp.search.Dealer.rerank) at 0x75ce1e9d6050>", line 35, in rerank
  File "/ragflow/rag/nlp/search.py", line 301, in rerank
    rank_fea = self._rank_feature_scores(rank_feature, sres)
  File "/ragflow/rag/nlp/search.py", line 261, in _rank_feature_scores
    for t, sc in eval(search_res.field[i].get(TAG_FLD, "{}")).items():
  File "<string>", line 0

If Reasoning is not turned on, there will be no problem. What should I do to solve it?

[KevinHuSh]

It seemed you utilizing a Tag KB to put some labels to the chunks. But something went wrong.
How to reproduce it?
Or, you could add a print statement in the code like bellow:

1. Go into container.
2. Add this line of code.
3. kill -9 PID of ragflow_server.py
4. Chat again and see what print out from the backend logs.

[zhouyangit]

Good. Give it a try.

[tyonia]

v0.17.0，same here

[zhouyangit]

I am not using a Tag KB to label some chunks.

[liaoshuren]

I also meet the same question and I think may be the result is not a correct python dict string. so I change the code alittle bit ,blow is my change

        tag_data = search_res.field[i].get(TAG_FLD, "{}")
        print("Raw TAG_FLD data:", repr(tag_data))  # 使~Trepr~X示转~I~W符
        parsed_data={}
        try:
            parsed_data = eval(tag_data)
        except SyntaxError as e:
             print(f"Syntax error in string: {e}")
        for t, sc in parsed_data.items():