agent中 DB Assistant模板连不上es，第一次可能行，后面就会报错。试了好几个版本 #6125
Description
我使用的是官方的docker部署，
其中四个版本：infiniflow/ragflow:v0.17.0 17.2 和 v0.17.1-slim、v0.17.2-slim
都是报这个错：
2025-03-15 10:51:49,901 INFO 14 HTTP Request: POST http://192.168.66.59:11434/api/embeddings "HTTP/1.1 200 OK"
2025-03-15 10:51:49,908 ERROR 14 Canvas.run got exception:
Traceback (most recent call last):
File "/ragflow/agent/canvas.py", line 222, in prepare2run
ans = cpn.run(self.history, **kwargs)
File "/ragflow/agent/component/base.py", line 420, in run
raise e
File "/ragflow/agent/component/base.py", line 416, in run
res = self._run(history, **kwargs)
File "/ragflow/agent/component/retrieval.py", line 73, in _run
kbinfos = settings.retrievaler.retrieval(query, embd_mdl, kbs[0].tenant_id, self._param.kb_ids,
File "<@beartype(rag.nlp.search.Dealer.retrieval) at 0x7f9f94e56200>", line 35, in retrieval
File "/ragflow/rag/nlp/search.py", line 360, in retrieval
sres = self.search(req, [index_name(tid) for tid in tenant_ids],
File "<@beartype(rag.nlp.search.Dealer.search) at 0x7f9f94e55e10>", line 95, in search
File "/ragflow/rag/nlp/search.py", line 115, in search
res = self.dataStore.search(src, highlightFields, filters, matchExprs, orderBy, offset, limit,
File "/ragflow/rag/utils/es_conn.py", line 177, in search
assert len(matchExprs) == 3 and isinstance(matchExprs[0], MatchTextExpr) and isinstance(matchExprs[1],
AssertionError
Traceback (most recent call last):
File "/ragflow/api/apps/canvas_app.py", line 135, in sse
for ans in canvas.run(stream=True):
File "/ragflow/agent/canvas.py", line 271, in run
for m in prepare2run(downstream):
File "/ragflow/agent/canvas.py", line 227, in prepare2run
raise e
File "/ragflow/agent/canvas.py", line 222, in prepare2run
ans = cpn.run(self.history, **kwargs)
File "/ragflow/agent/component/base.py", line 420, in run
raise e
File "/ragflow/agent/component/base.py", line 416, in run
res = self._run(history, **kwargs)
File "/ragflow/agent/component/retrieval.py", line 73, in _run
kbinfos = settings.retrievaler.retrieval(query, embd_mdl, kbs[0].tenant_id, self._param.kb_ids,
File "<@beartype(rag.nlp.search.Dealer.retrieval) at 0x7f9f94e56200>", line 35, in retrieval
File "/ragflow/rag/nlp/search.py", line 360, in retrieval
sres = self.search(req, [index_name(tid) for tid in tenant_ids],
File "<@beartype(rag.nlp.search.Dealer.search) at 0x7f9f94e55e10>", line 95, in search
File "/ragflow/rag/nlp/search.py", line 115, in search
res = self.dataStore.search(src, highlightFields, filters, matchExprs, orderBy, offset, limit,
File "/ragflow/rag/utils/es_conn.py", line 177, in search
assert len(matchExprs) == 3 and isinstance(matchExprs[0], MatchTextExpr) and isinstance(matchExprs[1],
AssertionError
然后我降级到了v.17.0 也是经常不行。
说我没有：但实际我是有的，并且docker logs -f ragflow-server 也没见报错信息。
Nothing found in DDL!
Nothing found in DB-Description!
Nothing found in Q->SQL!
v.17.0 很偶尔还会报这种错：就是到LLm那里的时候
2025-03-15 16:31:05,013 WARNING 18 /ragflow/agent/component/generate.py:79: FutureWarning: A value is trying to be set on a copy of a DataFrame or Series through chained assignment using an inplace method.
The behavior will change in pandas 3.0. This inplace method will never work because the intermediate object on which we are setting values always behaves as a copy.
For example, when doing 'df[col].method(value, inplace=True)', try using 'df.method({col: value}, inplace=True)' or df[col] = df[col].method(value) instead, to perform the operation inplace on the original object.
retrieval_res["empty_response"].fillna("", inplace=True)