修复重建知识库界面无反应的问题，添加一些log信息 (#1258)

* close #1172: 给webui_page/utils添加一些log信息，方便定位错误

* 修复：重建知识库时页面未实时显示进度

---------

Co-authored-by: liunux4odoo <liunu@qq.com>

server/knowledge_base/kb_doc_api.py

@@ -183,7 +183,7 @@ async def recreate_vector_store(
     set allow_empty_kb to True make it applied on empty knowledge base which it not in the info.db or having no documents.
     '''
 
-    async def output():
+    def output():
         kb = KBServiceFactory.get_service(knowledge_base_name, vs_type, embed_model)
         if not kb.exists() and not allow_empty_kb:
             yield {"code": 404, "msg": f"未找到知识库 ‘{knowledge_base_name}’"}

webui_pages/knowledge_base/knowledge_base.py

@@ -244,7 +244,6 @@ def knowledge_base_page(api: ApiRequest):
 
         cols = st.columns(3)
 
-        # todo: freezed
         if cols[0].button(
                 "依据源文件重建向量库",
                 # help="无需上传文件，通过其它方式将文档拷贝到对应知识库content目录下，点击本按钮即可重建知识库。",
@@ -258,7 +257,7 @@ def knowledge_base_page(api: ApiRequest):
                     if msg := check_error_msg(d):
                         st.toast(msg)
                     else:
-                        empty.progress(d["finished"] / d["total"], f"正在处理： {d['doc']}")
+                        empty.progress(d["finished"] / d["total"], d["msg"])
                 st.experimental_rerun()
 
         if cols[2].button(

webui_pages/utils.py

@@ -25,6 +25,7 @@ from server.utils import run_async, iter_over_async
 from configs.model_config import NLTK_DATA_PATH
 import nltk
 nltk.data.path = [NLTK_DATA_PATH] + nltk.data.path
+from pprint import pprint
 
 
 def set_httpx_timeout(timeout=60.0):
@@ -224,9 +225,17 @@ class ApiRequest:
         try:
             with response as r:
                 for chunk in r.iter_text(None):
-                    if as_json and chunk:
-                        yield json.loads(chunk)
-                    elif chunk.strip():
+                    if not chunk: # fastchat api yield empty bytes on start and end
+                        continue
+                    if as_json:
+                        try:
+                            data = json.loads(chunk)
+                            pprint(data, depth=1)
+                            yield data
+                        except Exception as e:
+                            logger.error(f"接口返回json错误： ‘{chunk}’。错误信息是：{e}。")
+                    else:
+                        print(chunk, end="", flush=True)
                         yield chunk
         except httpx.ConnectError as e:
             msg = f"无法连接API服务器，请确认 ‘api.py’ 已正常启动。"
@@ -274,6 +283,9 @@ class ApiRequest:
             return self._fastapi_stream2generator(response)
         else:
             data = msg.dict(exclude_unset=True, exclude_none=True)
+            print(f"received input message:")
+            pprint(data)
+
             response = self.post(
                 "/chat/fastchat",
                 json=data,
@@ -300,6 +312,9 @@ class ApiRequest:
             "stream": stream,
         }
 
+        print(f"received input message:")
+        pprint(data)
+
         if no_remote_api:
             from server.chat.chat import chat
             response = chat(**data)
@@ -334,6 +349,9 @@ class ApiRequest:
             "local_doc_url": no_remote_api,
         }
 
+        print(f"received input message:")
+        pprint(data)
+
         if no_remote_api:
             from server.chat.knowledge_base_chat import knowledge_base_chat
             response = knowledge_base_chat(**data)
@@ -367,6 +385,9 @@ class ApiRequest:
             "stream": stream,
         }
 
+        print(f"received input message:")
+        pprint(data)
+
         if no_remote_api:
             from server.chat.search_engine_chat import search_engine_chat
             response = search_engine_chat(**data)