fix: error

(cherry picked from commit 2a63cd6bea)

apps/application/flow/workflow_manage.py

@@ -10,6 +10,7 @@ import concurrent
 import json
 import threading
 import traceback
+import uuid
 from concurrent.futures import ThreadPoolExecutor
 from functools import reduce
 from typing import List, Dict
@@ -341,15 +342,19 @@ class WorkflowManage:
         self.run_chain_async(current_node, node_result_future)
         return tools.to_stream_response_simple(self.await_result())
 
-    def is_run(self, timeout=0.1):
-        self.lock.acquire()
+    def is_run(self, timeout=0.5):
+        future_list_len = len(self.future_list)
         try:
             r = concurrent.futures.wait(self.future_list, timeout)
-            return len(r.not_done) > 0
+            if len(r.not_done) > 0:
+                return True
+            else:
+                if future_list_len == len(self.future_list):
+                    return False
+                else:
+                    return True
         except Exception as e:
             return True
-        finally:
-            self.lock.release()
 
     def await_result(self):
         try:
@@ -402,12 +407,8 @@ class WorkflowManage:
             # 获取到可执行的子节点
             result_list = [{'node': node, 'future': executor.submit(self.run_chain_manage, node, None)} for node in
                            sorted_node_run_list]
-            try:
-                self.lock.acquire()
-                for r in result_list:
-                    self.future_list.append(r.get('future'))
-            finally:
-                self.lock.release()
+            for r in result_list:
+                self.future_list.append(r.get('future'))
 
     def run_chain(self, current_node, node_result_future=None):
         if node_result_future is None:
@@ -575,7 +576,7 @@ class WorkflowManage:
             details['node_id'] = node.id
             details['up_node_id_list'] = node.up_node_id_list
             details['runtime_node_id'] = node.runtime_node_id
-            details_result[node.runtime_node_id] = details
+            details_result[str(uuid.uuid1())] = details
         return details_result
 
     def get_answer_text_list(self):
@@ -664,9 +665,18 @@ class WorkflowManage:
             for edge in self.flow.edges:
                 if (edge.sourceNodeId == current_node.id and
                         f"{edge.sourceNodeId}_{current_node_result.node_variable.get('branch_id')}_right" == edge.sourceAnchorId):
-                    if self.dependent_node_been_executed(edge.targetNodeId):
+                    next_node = [node for node in self.flow.nodes if node.id == edge.targetNodeId]
+                    if len(next_node) == 0:
+                        continue
+                    if next_node[0].properties.get('condition', "AND") == 'AND':
+                        if self.dependent_node_been_executed(edge.targetNodeId):
+                            node_list.append(
+                                self.get_node_cls_by_id(edge.targetNodeId,
+                                                        [*current_node.up_node_id_list, current_node.node.id]))
+                    else:
                         node_list.append(
-                            self.get_node_cls_by_id(edge.targetNodeId, self.get_up_node_id_list(edge.targetNodeId)))
+                            self.get_node_cls_by_id(edge.targetNodeId,
+                                                    [*current_node.up_node_id_list, current_node.node.id]))
         else:
             for edge in self.flow.edges:
                 if edge.sourceNodeId == current_node.id:
@@ -676,10 +686,12 @@ class WorkflowManage:
                     if next_node[0].properties.get('condition', "AND") == 'AND':
                         if self.dependent_node_been_executed(edge.targetNodeId):
                             node_list.append(
-                                self.get_node_cls_by_id(edge.targetNodeId, self.get_up_node_id_list(edge.targetNodeId)))
+                                self.get_node_cls_by_id(edge.targetNodeId,
+                                                        [*current_node.up_node_id_list, current_node.node.id]))
                     else:
                         node_list.append(
-                            self.get_node_cls_by_id(edge.targetNodeId, [current_node.node.id]))
+                            self.get_node_cls_by_id(edge.targetNodeId,
+                                                    [*current_node.up_node_id_list, current_node.node.id]))
         return node_list
 
     def get_reference_field(self, node_id: str, fields: List[str]):

apps/application/serializers/application_serializers.py

@@ -35,7 +35,7 @@ from common.config.embedding_config import VectorStore
 from common.constants.authentication_type import AuthenticationType
 from common.db.search import get_dynamics_model, native_search, native_page_search
 from common.db.sql_execute import select_list
-from common.exception.app_exception import AppApiException, NotFound404, AppUnauthorizedFailed
+from common.exception.app_exception import AppApiException, NotFound404, AppUnauthorizedFailed, ChatException
 from common.field.common import UploadedImageField, UploadedFileField
 from common.models.db_model_manage import DBModelManage
 from common.response import result
@@ -268,7 +268,8 @@ class ApplicationSerializer(serializers.Serializer):
                         float_location = application_setting.float_location
                     if application_setting.custom_theme is not None and len(
                             application_setting.custom_theme.get('header_font_color', 'rgb(100, 106, 115)')) > 0:
-                        header_font_color = application_setting.custom_theme.get('header_font_color', 'rgb(100, 106, 115)')
+                        header_font_color = application_setting.custom_theme.get('header_font_color',
+                                                                                 'rgb(100, 106, 115)')
 
             is_auth = 'true' if application_access_token is not None and application_access_token.is_active else 'false'
             t = Template(content)
@@ -916,6 +917,12 @@ class ApplicationSerializer(serializers.Serializer):
             if application_access_token is None:
                 raise AppUnauthorizedFailed(500, "非法用户")
             application_setting_model = DBModelManage.get_model('application_setting')
+            if application.type == ApplicationTypeChoices.WORK_FLOW:
+                work_flow_version = QuerySet(WorkFlowVersion).filter(application_id=application.id).order_by(
+                    '-create_time')[0:1].first()
+                if work_flow_version is not None:
+                    application.work_flow = work_flow_version.work_flow
+
             xpack_cache = DBModelManage.get_model('xpack_cache')
             X_PACK_LICENSE_IS_VALID = False if xpack_cache is None else xpack_cache.get('XPACK_LICENSE_IS_VALID', False)
             application_setting_dict = {}
@@ -1146,9 +1153,23 @@ class ApplicationSerializer(serializers.Serializer):
         def get_application(self, app_id, with_valid=True):
             if with_valid:
                 self.is_valid(raise_exception=True)
-            application = QuerySet(Application).filter(id=self.data.get("application_id")).first()
-            return ApplicationSerializer.Operate(data={'user_id': application.user_id, 'application_id': app_id}).one(
-                with_valid=True)
+            if with_valid:
+                self.is_valid()
+            embed_application = QuerySet(Application).get(id=app_id)
+            if embed_application.type == ApplicationTypeChoices.WORK_FLOW:
+                work_flow_version = QuerySet(WorkFlowVersion).filter(application_id=embed_application.id).order_by(
+                    '-create_time')[0:1].first()
+                if work_flow_version is not None:
+                    embed_application.work_flow = work_flow_version.work_flow
+            dataset_list = self.list_dataset(with_valid=False)
+            mapping_dataset_id_list = [adm.dataset_id for adm in
+                                       QuerySet(ApplicationDatasetMapping).filter(application_id=app_id)]
+            dataset_id_list = [d.get('id') for d in
+                               list(filter(lambda row: mapping_dataset_id_list.__contains__(row.get('id')),
+                                           dataset_list))]
+            self.update_search_node(embed_application.work_flow, [str(dataset.get('id')) for dataset in dataset_list])
+            return {**ApplicationSerializer.Query.reset_application(ApplicationSerializerModel(embed_application).data),
+                    'dataset_id_list': dataset_id_list}
 
     class ApplicationKeySerializerModel(serializers.ModelSerializer):
         class Meta:

apps/common/event/__init__.py

@@ -9,6 +9,7 @@
 import setting.models
 from setting.models import Model
 from .listener_manage import *
+from common.db.sql_execute import update_execute
 
 update_document_status_sql = """
 UPDATE "public"."document" 

apps/common/event/listener_manage.py

@@ -6,26 +6,22 @@
     @date：2023/10/20 14:01
     @desc:
 """
-import datetime
 import logging
 import os
 import threading
-import time
 import traceback
 from typing import List
 
 import django.db.models
-from django.db import models, transaction
 from django.db.models import QuerySet
 from django.db.models.functions import Substr, Reverse
 from langchain_core.embeddings import Embeddings
 
 from common.config.embedding_config import VectorStore
 from common.db.search import native_search, get_dynamics_model, native_update
-from common.db.sql_execute import sql_execute, update_execute
 from common.util.file_util import get_file_content
 from common.util.lock import try_lock, un_lock
-from common.util.page_utils import page
+from common.util.page_utils import page_desc
 from dataset.models import Paragraph, Status, Document, ProblemParagraphMapping, TaskType, State
 from embedding.models import SourceType, SearchMode
 from smartdoc.conf import PROJECT_DIR
@@ -162,7 +158,7 @@ class ListenerManagement:
                 if is_the_task_interrupted():
                     break
                 ListenerManagement.embedding_by_paragraph(str(paragraph.get('id')), embedding_model)
-                post_apply()
+            post_apply()
 
         return embedding_paragraph_apply
 
@@ -241,13 +237,16 @@ class ListenerManagement:
             lock.release()
 
     @staticmethod
-    def embedding_by_document(document_id, embedding_model: Embeddings):
+    def embedding_by_document(document_id, embedding_model: Embeddings, state_list=None):
         """
         向量化文档
+        @param state_list:
         @param document_id: 文档id
         @param embedding_model 向量模型
         :return: None
         """
+        if state_list is None:
+            state_list = [State.PENDING, State.SUCCESS, State.FAILURE, State.REVOKE, State.REVOKED]
         if not try_lock('embedding' + str(document_id)):
             return
         try:
@@ -268,11 +267,17 @@ class ListenerManagement:
             VectorStore.get_embedding_vector().delete_by_document_id(document_id)
 
             # 根据段落进行向量化处理
-            page(QuerySet(Paragraph).filter(document_id=document_id).values('id'), 5,
-                 ListenerManagement.get_embedding_paragraph_apply(embedding_model, is_the_task_interrupted,
-                                                                  ListenerManagement.get_aggregation_document_status(
-                                                                      document_id)),
-                 is_the_task_interrupted)
+            page_desc(QuerySet(Paragraph)
+                      .annotate(
+                reversed_status=Reverse('status'),
+                task_type_status=Substr('reversed_status', TaskType.EMBEDDING.value,
+                                        1),
+            ).filter(task_type_status__in=state_list, document_id=document_id)
+                      .values('id'), 5,
+                      ListenerManagement.get_embedding_paragraph_apply(embedding_model, is_the_task_interrupted,
+                                                                       ListenerManagement.get_aggregation_document_status(
+                                                                           document_id)),
+                      is_the_task_interrupted)
         except Exception as e:
             max_kb_error.error(f'向量化文档:{document_id}出现错误{str(e)}{traceback.format_exc()}')
         finally:

apps/common/handle/impl/doc_split_handle.py

@@ -113,8 +113,10 @@ class DocSplitHandle(BaseSplitHandle):
     def paragraph_to_md(paragraph: Paragraph, doc: Document, images_list, get_image_id):
         try:
             psn = paragraph.style.name
-            if psn.startswith('Heading'):
-                title = "".join(["#" for i in range(int(psn.replace("Heading ", '')))]) + " " + paragraph.text
+            if psn.startswith('Heading') or psn.startswith('TOC 标题') or psn.startswith('标题'):
+                title = "".join(["#" for i in range(
+                    int(psn.replace("Heading ", '').replace('TOC 标题', '').replace('标题',
+                                                                                    '')))]) + " " + paragraph.text
                 images = reduce(lambda x, y: [*x, *y],
                                 [get_paragraph_element_images(e, doc, images_list, get_image_id) for e in
                                  paragraph._element],
@@ -202,4 +204,4 @@ class DocSplitHandle(BaseSplitHandle):
             return content
         except BaseException as e:
             traceback.print_exception(e)
-            return f'{e}'
+            return f'{e}'

apps/common/util/page_utils.py

@@ -26,3 +26,22 @@ def page(query_set, page_size, handler, is_the_task_interrupted=lambda: False):
         offset = i * page_size
         paragraph_list = query.all()[offset: offset + page_size]
         handler(paragraph_list)
+
+
+def page_desc(query_set, page_size, handler, is_the_task_interrupted=lambda: False):
+    """
+
+    @param query_set: 查询query_set
+    @param page_size: 每次查询大小
+    @param handler:   数据处理器
+    @param is_the_task_interrupted: 任务是否被中断
+    @return:
+    """
+    query = query_set.order_by("id")
+    count = query_set.count()
+    for i in sorted(range(0, ceil(count / page_size)), reverse=True):
+        if is_the_task_interrupted():
+            return
+        offset = i * page_size
+        paragraph_list = query.all()[offset: offset + page_size]
+        handler(paragraph_list)

apps/dataset/serializers/document_serializers.py

@@ -700,20 +700,28 @@ class DocumentSerializers(ApiMixin, serializers.Serializer):
             _document.save()
             return self.one()
 
-        @transaction.atomic
-        def refresh(self, with_valid=True):
+        def refresh(self, state_list=None, with_valid=True):
+            if state_list is None:
+                state_list = [State.PENDING.value, State.STARTED.value, State.SUCCESS.value, State.FAILURE.value,
+                              State.REVOKE.value,
+                              State.REVOKED.value, State.IGNORED.value]
             if with_valid:
                 self.is_valid(raise_exception=True)
             document_id = self.data.get("document_id")
             ListenerManagement.update_status(QuerySet(Document).filter(id=document_id), TaskType.EMBEDDING,
                                              State.PENDING)
-            ListenerManagement.update_status(QuerySet(Paragraph).filter(document_id=document_id),
+            ListenerManagement.update_status(QuerySet(Paragraph).annotate(
+                reversed_status=Reverse('status'),
+                task_type_status=Substr('reversed_status', TaskType.EMBEDDING.value,
+                                        1),
+            ).filter(task_type_status__in=state_list, document_id=document_id)
+                                             .values('id'),
                                              TaskType.EMBEDDING,
                                              State.PENDING)
             ListenerManagement.get_aggregation_document_status(document_id)()
             embedding_model_id = get_embedding_model_id_by_dataset_id(dataset_id=self.data.get('dataset_id'))
             try:
-                embedding_by_document.delay(document_id, embedding_model_id)
+                embedding_by_document.delay(document_id, embedding_model_id, state_list)
             except AlreadyQueued as e:
                 raise AppApiException(500, "任务正在执行中,请勿重复下发")
 
@@ -1122,14 +1130,14 @@ class DocumentSerializers(ApiMixin, serializers.Serializer):
             if with_valid:
                 self.is_valid(raise_exception=True)
             document_id_list = instance.get("id_list")
-            with transaction.atomic():
-                dataset_id = self.data.get('dataset_id')
-                for document_id in document_id_list:
-                    try:
-                        DocumentSerializers.Operate(
-                            data={'dataset_id': dataset_id, 'document_id': document_id}).refresh()
-                    except AlreadyQueued as e:
-                        pass
+            state_list = instance.get("state_list")
+            dataset_id = self.data.get('dataset_id')
+            for document_id in document_id_list:
+                try:
+                    DocumentSerializers.Operate(
+                        data={'dataset_id': dataset_id, 'document_id': document_id}).refresh(state_list)
+                except AlreadyQueued as e:
+                    pass
 
     class GenerateRelated(ApiMixin, serializers.Serializer):
         document_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid("文档id"))

apps/dataset/swagger_api/document_api.py

@@ -51,3 +51,16 @@ class DocumentApi(ApiMixin):
                                            description="1|2|3 1:向量化|2:生成问题|3:同步文档", default=1)
                 }
             )
+
+    class EmbeddingState(ApiMixin):
+        @staticmethod
+        def get_request_body_api():
+            return openapi.Schema(
+                type=openapi.TYPE_OBJECT,
+                properties={
+                    'state_list': openapi.Schema(type=openapi.TYPE_ARRAY,
+                                                 items=openapi.Schema(type=openapi.TYPE_STRING),
+                                                 title="状态列表",
+                                                 description="状态列表")
+                }
+            )

apps/dataset/views/document.py

@@ -262,6 +262,7 @@ class Document(APIView):
         @action(methods=['PUT'], detail=False)
         @swagger_auto_schema(operation_summary="刷新文档向量库",
                              operation_id="刷新文档向量库",
+                             request_body=DocumentApi.EmbeddingState.get_request_body_api(),
                              manual_parameters=DocumentSerializers.Operate.get_request_params_api(),
                              responses=result.get_default_response(),
                              tags=["知识库/文档"]
@@ -272,6 +273,7 @@ class Document(APIView):
         def put(self, request: Request, dataset_id: str, document_id: str):
             return result.success(
                 DocumentSerializers.Operate(data={'document_id': document_id, 'dataset_id': dataset_id}).refresh(
+                    request.data.get('state_list')
                 ))
 
     class BatchRefresh(APIView):

apps/embedding/task/embedding.py

@@ -56,14 +56,20 @@ def embedding_by_paragraph_list(paragraph_id_list, model_id):
 
 
 @celery_app.task(base=QueueOnce, once={'keys': ['document_id']}, name='celery:embedding_by_document')
-def embedding_by_document(document_id, model_id):
+def embedding_by_document(document_id, model_id, state_list=None):
     """
     向量化文档
+    @param state_list:
     @param document_id: 文档id
     @param model_id 向量模型
     :return: None
     """
 
+    if state_list is None:
+        state_list = [State.PENDING.value, State.STARTED.value, State.SUCCESS.value, State.FAILURE.value,
+                      State.REVOKE.value,
+                      State.REVOKED.value, State.IGNORED.value]
+
     def exception_handler(e):
         ListenerManagement.update_status(QuerySet(Document).filter(id=document_id), TaskType.EMBEDDING,
                                          State.FAILURE)
@@ -71,7 +77,7 @@ def embedding_by_document(document_id, model_id):
             f'获取向量模型失败：{str(e)}{traceback.format_exc()}')
 
     embedding_model = get_embedding_model(model_id, exception_handler)
-    ListenerManagement.embedding_by_document(document_id, embedding_model)
+    ListenerManagement.embedding_by_document(document_id, embedding_model, state_list)
 
 
 @celery_app.task(name='celery:embedding_by_document_list')

apps/setting/models_provider/impl/aliyun_bai_lian_model_provider/aliyun_bai_lian_model_provider.py

@@ -73,7 +73,9 @@ model_info_manage = (
     ModelInfoManage.builder()
     .append_model_info_list(model_info_list)
     .append_model_info_list(module_info_vl_list)
+    .append_default_model_info(module_info_vl_list[0])
     .append_model_info_list(module_info_tti_list)
+    .append_default_model_info(module_info_tti_list[0])
     .append_default_model_info(model_info_list[1])
     .append_default_model_info(model_info_list[2])
     .append_default_model_info(model_info_list[3])

apps/setting/models_provider/impl/aliyun_bai_lian_model_provider/model/image.py

@@ -2,12 +2,11 @@
 
 from typing import Dict
 
-from langchain_community.chat_models import ChatOpenAI
-
 from setting.models_provider.base_model_provider import MaxKBBaseModel
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
-class QwenVLChatModel(MaxKBBaseModel, ChatOpenAI):
+class QwenVLChatModel(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -18,6 +17,7 @@ class QwenVLChatModel(MaxKBBaseModel, ChatOpenAI):
             openai_api_base='https://dashscope.aliyuncs.com/compatible-mode/v1',
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )
         return chat_tong_yi

apps/setting/models_provider/impl/azure_model_provider/model/image.py

@@ -1,7 +1,7 @@
-from typing import Dict
+from typing import Dict, List
 
+from langchain_core.messages import BaseMessage, get_buffer_string
 from langchain_openai import AzureChatOpenAI
-from langchain_openai.chat_models import ChatOpenAI
 
 from common.config.tokenizer_manage_config import TokenizerManage
 from setting.models_provider.base_model_provider import MaxKBBaseModel
@@ -26,3 +26,17 @@ class AzureOpenAIImage(MaxKBBaseModel, AzureChatOpenAI):
             streaming=True,
             **optional_params,
         )
+
+    def get_num_tokens_from_messages(self, messages: List[BaseMessage]) -> int:
+        try:
+            return super().get_num_tokens_from_messages(messages)
+        except Exception as e:
+            tokenizer = TokenizerManage.get_tokenizer()
+            return sum([len(tokenizer.encode(get_buffer_string([m]))) for m in messages])
+
+    def get_num_tokens(self, text: str) -> int:
+        try:
+            return super().get_num_tokens(text)
+        except Exception as e:
+            tokenizer = TokenizerManage.get_tokenizer()
+            return len(tokenizer.encode(text))

apps/setting/models_provider/impl/ollama_model_provider/model/image.py

@@ -1,15 +1,8 @@
 from typing import Dict
 from urllib.parse import urlparse, ParseResult
 
-from langchain_openai.chat_models import ChatOpenAI
-
-from common.config.tokenizer_manage_config import TokenizerManage
 from setting.models_provider.base_model_provider import MaxKBBaseModel
-
-
-def custom_get_token_ids(text: str):
-    tokenizer = TokenizerManage.get_tokenizer()
-    return tokenizer.encode(text)
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
 def get_base_url(url: str):
@@ -20,7 +13,7 @@ def get_base_url(url: str):
     return result_url[:-1] if result_url.endswith("/") else result_url
 
 
-class OllamaImage(MaxKBBaseModel, ChatOpenAI):
+class OllamaImage(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -34,5 +27,6 @@ class OllamaImage(MaxKBBaseModel, ChatOpenAI):
             openai_api_key=model_credential.get('api_key'),
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )

apps/setting/models_provider/impl/ollama_model_provider/ollama_model_provider.py

@@ -171,6 +171,7 @@ model_info_manage = (
         '一个具有大令牌上下文窗口的高性能开放嵌入模型。',
         ModelTypeConst.EMBEDDING, ollama_embedding_model_credential, OllamaEmbedding), )
     .append_model_info_list(image_model_info)
+    .append_default_model_info(image_model_info[0])
     .build()
 )
 

apps/setting/models_provider/impl/openai_model_provider/model/image.py

@@ -1,17 +1,10 @@
 from typing import Dict
 
-from langchain_openai.chat_models import ChatOpenAI
-
-from common.config.tokenizer_manage_config import TokenizerManage
 from setting.models_provider.base_model_provider import MaxKBBaseModel
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
-def custom_get_token_ids(text: str):
-    tokenizer = TokenizerManage.get_tokenizer()
-    return tokenizer.encode(text)
-
-
-class OpenAIImage(MaxKBBaseModel, ChatOpenAI):
+class OpenAIImage(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -22,5 +15,6 @@ class OpenAIImage(MaxKBBaseModel, ChatOpenAI):
             openai_api_key=model_credential.get('api_key'),
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )

apps/setting/models_provider/impl/openai_model_provider/openai_model_provider.py

@@ -121,7 +121,16 @@ model_info_manage = (
     .append_model_info_list(model_info_embedding_list)
     .append_default_model_info(model_info_embedding_list[0])
     .append_model_info_list(model_info_image_list)
+    .append_default_model_info(model_info_image_list[0])
     .append_model_info_list(model_info_tti_list)
+    .append_default_model_info(model_info_tti_list[0])
+    .append_default_model_info(ModelInfo('whisper-1', '',
+              ModelTypeConst.STT, openai_stt_model_credential,
+              OpenAISpeechToText)
+    )
+    .append_default_model_info(ModelInfo('tts-1', '',
+              ModelTypeConst.TTS, openai_tts_model_credential,
+              OpenAITextToSpeech))
     .build()
 )
 

apps/setting/models_provider/impl/qwen_model_provider/model/image.py

@@ -2,12 +2,11 @@
 
 from typing import Dict
 
-from langchain_community.chat_models import ChatOpenAI
-
 from setting.models_provider.base_model_provider import MaxKBBaseModel
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
-class QwenVLChatModel(MaxKBBaseModel, ChatOpenAI):
+class QwenVLChatModel(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -18,6 +17,7 @@ class QwenVLChatModel(MaxKBBaseModel, ChatOpenAI):
             openai_api_base='https://dashscope.aliyuncs.com/compatible-mode/v1',
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )
         return chat_tong_yi

apps/setting/models_provider/impl/qwen_model_provider/qwen_model_provider.py

@@ -46,7 +46,9 @@ model_info_manage = (
     .append_default_model_info(
         ModelInfo('qwen-turbo', '', ModelTypeConst.LLM, qwen_model_credential, QwenChatModel))
     .append_model_info_list(module_info_vl_list)
+    .append_default_model_info(module_info_vl_list[0])
     .append_model_info_list(module_info_tti_list)
+    .append_default_model_info(module_info_tti_list[0])
     .build()
 )
 

apps/setting/models_provider/impl/tencent_model_provider/model/image.py

@@ -1,17 +1,10 @@
 from typing import Dict
 
-from langchain_openai.chat_models import ChatOpenAI
-
-from common.config.tokenizer_manage_config import TokenizerManage
 from setting.models_provider.base_model_provider import MaxKBBaseModel
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
-def custom_get_token_ids(text: str):
-    tokenizer = TokenizerManage.get_tokenizer()
-    return tokenizer.encode(text)
-
-
-class TencentVision(MaxKBBaseModel, ChatOpenAI):
+class TencentVision(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -22,5 +15,6 @@ class TencentVision(MaxKBBaseModel, ChatOpenAI):
             openai_api_key=model_credential.get('api_key'),
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )

apps/setting/models_provider/impl/tencent_model_provider/tencent_model_provider.py

@@ -96,13 +96,15 @@ def _initialize_model_info():
         TencentTTIModelCredential,
         TencentTextToImageModel)]
 
-
     model_info_manage = ModelInfoManage.builder() \
         .append_model_info_list(model_info_list) \
         .append_model_info_list(model_info_embedding_list) \
         .append_model_info_list(model_info_vision_list) \
+        .append_default_model_info(model_info_vision_list[0]) \
         .append_model_info_list(model_info_tti_list) \
+        .append_default_model_info(model_info_tti_list[0]) \
         .append_default_model_info(model_info_list[0]) \
+        .append_default_model_info(tencent_embedding_model_info) \
         .build()
 
     return model_info_manage

apps/setting/models_provider/impl/volcanic_engine_model_provider/model/image.py

@@ -1,17 +1,10 @@
 from typing import Dict
 
-from langchain_openai.chat_models import ChatOpenAI
-
-from common.config.tokenizer_manage_config import TokenizerManage
 from setting.models_provider.base_model_provider import MaxKBBaseModel
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
-def custom_get_token_ids(text: str):
-    tokenizer = TokenizerManage.get_tokenizer()
-    return tokenizer.encode(text)
-
-
-class VolcanicEngineImage(MaxKBBaseModel, ChatOpenAI):
+class VolcanicEngineImage(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -22,5 +15,6 @@ class VolcanicEngineImage(MaxKBBaseModel, ChatOpenAI):
             openai_api_base=model_credential.get('api_base'),
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )

apps/setting/models_provider/impl/volcanic_engine_model_provider/volcanic_engine_model_provider.py

@@ -93,6 +93,9 @@ model_info_manage = (
     .append_model_info_list(model_info_list)
     .append_default_model_info(model_info_list[0])
     .append_default_model_info(model_info_list[1])
+    .append_default_model_info(model_info_list[2])
+    .append_default_model_info(model_info_list[3])
+    .append_default_model_info(model_info_list[4])
     .build()
 )
 

apps/setting/models_provider/impl/xf_model_provider/xf_model_provider.py

@@ -40,8 +40,20 @@ model_info_list = [
     ModelInfo('embedding', '', ModelTypeConst.EMBEDDING, embedding_model_credential, XFEmbedding)
 ]
 
-model_info_manage = ModelInfoManage.builder().append_model_info_list(model_info_list).append_default_model_info(
-    ModelInfo('generalv3.5', '', ModelTypeConst.LLM, qwen_model_credential, XFChatSparkLLM)).build()
+model_info_manage = (
+    ModelInfoManage.builder()
+    .append_model_info_list(model_info_list)
+    .append_default_model_info(
+        ModelInfo('generalv3.5', '', ModelTypeConst.LLM, qwen_model_credential, XFChatSparkLLM))
+    .append_default_model_info(
+        ModelInfo('iat', '中英文识别', ModelTypeConst.STT, stt_model_credential, XFSparkSpeechToText),
+    )
+    .append_default_model_info(
+        ModelInfo('tts', '', ModelTypeConst.TTS, tts_model_credential, XFSparkTextToSpeech))
+    .append_default_model_info(
+        ModelInfo('embedding', '', ModelTypeConst.EMBEDDING, embedding_model_credential, XFEmbedding))
+    .build()
+)
 
 
 class XunFeiModelProvider(IModelProvider):

apps/setting/models_provider/impl/xinference_model_provider/model/image.py

@@ -1,17 +1,10 @@
 from typing import Dict
 
-from langchain_openai.chat_models import ChatOpenAI
-
-from common.config.tokenizer_manage_config import TokenizerManage
 from setting.models_provider.base_model_provider import MaxKBBaseModel
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
-def custom_get_token_ids(text: str):
-    tokenizer = TokenizerManage.get_tokenizer()
-    return tokenizer.encode(text)
-
-
-class XinferenceImage(MaxKBBaseModel, ChatOpenAI):
+class XinferenceImage(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -22,5 +15,6 @@ class XinferenceImage(MaxKBBaseModel, ChatOpenAI):
             openai_api_key=model_credential.get('api_key'),
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )

apps/setting/models_provider/impl/xinference_model_provider/xinference_model_provider.py

@@ -521,25 +521,29 @@ embedding_model_info = [
 rerank_list = [ModelInfo('bce-reranker-base_v1',
                          '发布新的重新排名器，建立在强大的 M3 和LLM （GEMMA 和 MiniCPM，实际上没那么大）骨干上，支持多语言处理和更大的输入，大幅提高 BEIR、C-MTEB/Retrieval 的排名性能、MIRACL、LlamaIndex 评估',
                          ModelTypeConst.RERANKER, XInferenceRerankerModelCredential(), XInferenceReranker)]
-model_info_manage = (ModelInfoManage.builder()
-                     .append_model_info_list(model_info_list)
-                     .append_model_info_list(voice_model_info)
-                     .append_default_model_info(voice_model_info[0])
-                     .append_default_model_info(voice_model_info[1])
-                     .append_default_model_info(ModelInfo('phi3',
-                                                          'Phi-3 Mini是Microsoft的3.8B参数，轻量级，最先进的开放模型。',
-                                                          ModelTypeConst.LLM, xinference_llm_model_credential,
-                                                          XinferenceChatModel))
-                     .append_model_info_list(embedding_model_info)
-                     .append_default_model_info(ModelInfo('',
-                                                          '',
-                                                          ModelTypeConst.EMBEDDING,
-                                                          xinference_embedding_model_credential, XinferenceEmbedding))
-                     .append_model_info_list(rerank_list)
-                     .append_model_info_list(image_model_info)
-                     .append_model_info_list(tti_model_info)
-                     .append_default_model_info(rerank_list[0])
-                     .build())
+model_info_manage = (
+    ModelInfoManage.builder()
+    .append_model_info_list(model_info_list)
+    .append_model_info_list(voice_model_info)
+    .append_default_model_info(voice_model_info[0])
+    .append_default_model_info(voice_model_info[1])
+    .append_default_model_info(ModelInfo('phi3',
+                                         'Phi-3 Mini是Microsoft的3.8B参数，轻量级，最先进的开放模型。',
+                                         ModelTypeConst.LLM, xinference_llm_model_credential,
+                                         XinferenceChatModel))
+    .append_model_info_list(embedding_model_info)
+    .append_default_model_info(ModelInfo('',
+                                         '',
+                                         ModelTypeConst.EMBEDDING,
+                                         xinference_embedding_model_credential, XinferenceEmbedding))
+    .append_model_info_list(rerank_list)
+    .append_model_info_list(image_model_info)
+    .append_default_model_info(image_model_info[0])
+    .append_model_info_list(tti_model_info)
+    .append_default_model_info(tti_model_info[0])
+    .append_default_model_info(rerank_list[0])
+    .build()
+)
 
 
 def get_base_url(url: str):

apps/setting/models_provider/impl/zhipu_model_provider/model/image.py

@@ -1,17 +1,10 @@
 from typing import Dict
 
-from langchain_openai.chat_models import ChatOpenAI
-
-from common.config.tokenizer_manage_config import TokenizerManage
 from setting.models_provider.base_model_provider import MaxKBBaseModel
+from setting.models_provider.impl.base_chat_open_ai import BaseChatOpenAI
 
 
-def custom_get_token_ids(text: str):
-    tokenizer = TokenizerManage.get_tokenizer()
-    return tokenizer.encode(text)
-
-
-class ZhiPuImage(MaxKBBaseModel, ChatOpenAI):
+class ZhiPuImage(MaxKBBaseModel, BaseChatOpenAI):
 
     @staticmethod
     def new_instance(model_type, model_name, model_credential: Dict[str, object], **model_kwargs):
@@ -22,5 +15,6 @@ class ZhiPuImage(MaxKBBaseModel, ChatOpenAI):
             openai_api_base='https://open.bigmodel.cn/api/paas/v4',
             # stream_options={"include_usage": True},
             streaming=True,
+            stream_usage=True,
             **optional_params,
         )

apps/setting/models_provider/impl/zhipu_model_provider/zhipu_model_provider.py

@@ -48,6 +48,9 @@ model_info_tti_list = [
     ModelInfo('cogview-3-plus', '根据用户文字描述生成高质量图像，支持多图片尺寸',
               ModelTypeConst.TTI, zhipu_tti_model_credential,
               ZhiPuTextToImage),
+    ModelInfo('cogview-3-flash', '根据用户文字描述生成高质量图像，支持多图片尺寸(免费)',
+              ModelTypeConst.TTI, zhipu_tti_model_credential,
+              ZhiPuTextToImage),
 ]
 
 model_info_manage = (
@@ -55,7 +58,9 @@ model_info_manage = (
     .append_model_info_list(model_info_list)
     .append_default_model_info(ModelInfo('glm-4', '', ModelTypeConst.LLM, qwen_model_credential, ZhipuChatModel))
     .append_model_info_list(model_info_image_list)
+    .append_default_model_info(model_info_image_list[0])
     .append_model_info_list(model_info_tti_list)
+    .append_default_model_info(model_info_tti_list[0])
     .build()
 )
 

installer/Dockerfile

@@ -22,7 +22,7 @@ RUN mkdir -p /opt/maxkb/app /opt/maxkb/model /opt/maxkb/conf && \
 COPY --from=web-build ui /opt/maxkb/app/ui
 WORKDIR /opt/maxkb/app
 RUN python3 -m venv /opt/py3 && \
-    pip install poetry --break-system-packages && \
+    pip install poetry==1.8.5 --break-system-packages && \
     poetry config virtualenvs.create false && \
     . /opt/py3/bin/activate && \
     if [ "$(uname -m)" = "x86_64" ]; then sed -i 's/^torch.*/torch = {version = "^2.2.1+cpu", source = "pytorch"}/g' pyproject.toml; fi && \

ui/src/api/document.ts

@@ -129,11 +129,12 @@ const delMulDocument: (
 const batchRefresh: (
   dataset_id: string,
   data: any,
+  stateList: Array<string>,
   loading?: Ref<boolean>
-) => Promise<Result<boolean>> = (dataset_id, data, loading) => {
+) => Promise<Result<boolean>> = (dataset_id, data, stateList, loading) => {
   return put(
     `${prefix}/${dataset_id}/document/batch_refresh`,
-    { id_list: data },
+    { id_list: data, state_list: stateList },
     undefined,
     loading
   )
@@ -157,11 +158,12 @@ const getDocumentDetail: (dataset_id: string, document_id: string) => Promise<Re
 const putDocumentRefresh: (
   dataset_id: string,
   document_id: string,
+  state_list: Array<string>,
   loading?: Ref<boolean>
-) => Promise<Result<any>> = (dataset_id, document_id, loading) => {
+) => Promise<Result<any>> = (dataset_id, document_id, state_list, loading) => {
   return put(
     `${prefix}/${dataset_id}/document/${document_id}/refresh`,
-    undefined,
+    { state_list },
     undefined,
     loading
   )

ui/src/api/type/application.ts

@@ -121,12 +121,11 @@ export class ChatRecordManage {
 
     this.chat.answer_text = this.chat.answer_text + chunk_answer
   }
-  get_current_up_node() {
-    for (let i = this.node_list.length - 2; i >= 0; i--) {
-      const n = this.node_list[i]
-      if (n.content.length > 0) {
-        return n
-      }
+  get_current_up_node(run_node: any) {
+    const index = this.node_list.findIndex((item) => item == run_node)
+    if (index > 0) {
+      const n = this.node_list[index - 1]
+      return n
     }
     return undefined
   }
@@ -144,14 +143,13 @@ export class ChatRecordManage {
       const index = this.node_list.indexOf(run_node)
       let current_up_node = undefined
       if (index > 0) {
-        current_up_node = this.get_current_up_node()
+        current_up_node = this.get_current_up_node(run_node)
       }
       let answer_text_list_index = 0
-
       if (
         current_up_node == undefined ||
         run_node.view_type == 'single_view' ||
-        (run_node.view_type == 'many_view' && current_up_node.view_type == 'single_view')
+        current_up_node.view_type == 'single_view'
       ) {
         const none_index = this.findIndex(
           this.chat.answer_text_list,

ui/src/components/ai-chat/KnowledgeSource.vue

@@ -15,7 +15,7 @@
             <div class="flex-between">
               <div class="flex">
                 <img :src="getImgUrl(item && item?.document_name)" alt="" width="20" />
-                <div class="ml-4" v-if="!item.source_url">
+                <div class="ml-4 ellipsis-1" :title="item?.document_name" v-if="!item.source_url">
                   <p>{{ item && item?.document_name }}</p>
                 </div>
                 <div class="ml-8" v-else>

ui/src/components/ai-chat/component/ParagraphCard.vue

@@ -28,14 +28,14 @@
                   : meta?.source_url
               "
               target="_blank"
-              class="ellipsis-1"
+              class="ellipsis-1 break-all"
               :title="data?.document_name?.trim()"
             >
               {{ data?.document_name?.trim() }}
             </a>
           </template>
           <template v-else>
-            <span class="ellipsis-1" :title="data?.document_name?.trim()">
+            <span class="ellipsis-1 break-all" :title="data?.document_name?.trim()">
               {{ data?.document_name?.trim() }}
             </span>
           </template>
@@ -45,7 +45,9 @@
             <img src="@/assets/icon_document.svg" style="width: 58%" alt="" />
           </AppAvatar>
 
-          <span class="ellipsis-1" :title="data?.dataset_name"> {{ data?.dataset_name }}</span>
+          <span class="ellipsis-1 break-all" :title="data?.dataset_name">
+            {{ data?.dataset_name }}</span
+          >
         </div>
       </div>
     </template>
@@ -80,7 +82,7 @@ const meta = computed(() => (isMetaObject.value ? props.data.meta : parsedMeta.v
 .paragraph-source-card {
   .footer-content {
     .item {
-      width: 50%;
+      max-width: 50%;
     }
   }
 }
@@ -96,7 +98,7 @@ const meta = computed(() => (isMetaObject.value ? props.data.meta : parsedMeta.v
     .footer-content {
       display: block;
       .item {
-        width: 100%;
+        max-width: 100%;
       }
     }
   }

ui/src/components/ai-chat/component/chat-input-operate/index.vue

@@ -401,7 +401,7 @@ const startRecording = async () => {
 2、若无 https 配置则需要修改浏览器安全配置，Chrome 设置如下：<br/>
 (1) 地址栏输入chrome://flags/#unsafely-treat-insecure-origin-as-secure；<br/>
 (2) 将 http 站点配置在文本框中，例如: http://127.0.0.1:8080。</p>
-    <img src="${new URL(`../../assets/tipIMG.jpg`, import.meta.url).href}" style="width: 100%;" />`,
+    <img src="${new URL(`@/assets/tipIMG.jpg`, import.meta.url).href}" style="width: 100%;" />`,
           {
             confirmButtonText: '我知道了',
             dangerouslyUseHTMLString: true,
@@ -418,7 +418,7 @@ const startRecording = async () => {
 2、若无 https 配置则需要修改浏览器安全配置，Chrome 设置如下：<br/>
 (1) 地址栏输入chrome://flags/#unsafely-treat-insecure-origin-as-secure；<br/>
 (2) 将 http 站点配置在文本框中，例如: http://127.0.0.1:8080。</p>
-    <img src="${new URL(`../../assets/tipIMG.jpg`, import.meta.url).href}" style="width: 100%;" />`,
+    <img src="${new URL(`@/assets/tipIMG.jpg`, import.meta.url).href}" style="width: 100%;" />`,
       {
         confirmButtonText: '我知道了',
         dangerouslyUseHTMLString: true,

ui/src/views/application-overview/component/DisplaySettingDialog.vue

@@ -8,99 +8,15 @@
     <el-form label-position="top" ref="displayFormRef" :model="form">
       <el-form-item>
         <el-space direction="vertical" alignment="start">
-          <el-checkbox v-model="form.show_source" label="显示知识来源" />
           <el-checkbox
-            v-model="form.show_history"
-            label="显示历史记录"
-            v-if="user.isEnterprise()"
-          />
-          <el-checkbox
-            v-model="form.draggable"
-            label="可拖拽位置（浮窗模式）"
-            v-if="user.isEnterprise()"
-          />
-          <el-checkbox
-            v-model="form.show_guide"
-            label="显示引导图（浮窗模式）"
-            v-if="user.isEnterprise()"
+            v-model="form.show_source"
+            :label="isWorkFlow(detail.type) ? '显示执行详情' : '显示知识来源'"
           />
         </el-space>
       </el-form-item>
-      <el-form-item label="对话头像" v-if="user.isEnterprise()">
-        <div class="flex mt-8">
-          <div class="border border-r-4 mr-16" style="padding: 8px">
-            <el-image
-              v-if="imgUrl.avatar"
-              :src="imgUrl.avatar"
-              alt=""
-              fit="cover"
-              style="width: 50px; height: 50px; display: block"
-            />
-            <LogoIcon v-else height="50px" style="width: 50px; height: 50px; display: block" />
-          </div>
-
-          <el-upload
-            ref="uploadRef"
-            action="#"
-            :auto-upload="false"
-            :show-file-list="false"
-            accept="image/jpeg, image/png, image/gif"
-            :on-change="(file: any, fileList: any) => onChange(file, fileList, 'avatar')"
-          >
-            <el-button icon="Upload">{{
-              $t('views.applicationOverview.appInfo.EditAvatarDialog.upload')
-            }}</el-button>
-            <template #tip>
-              <div class="el-upload__tip info" style="margin-top: 0">
-                建议尺寸 32*32，支持 JPG、PNG、GIF，大小不超过 10 MB
-              </div>
-            </template>
-          </el-upload>
-        </div>
-      </el-form-item>
-      <el-form-item label="浮窗入口图标" v-if="user.isEnterprise()">
-        <div class="flex mt-8">
-          <div class="border border-r-4 mr-16" style="padding: 8px">
-            <el-image
-              v-if="imgUrl.float_icon"
-              :src="imgUrl.float_icon"
-              alt=""
-              fit="cover"
-              style="width: 50px; height: 50px; display: block"
-            />
-            <img
-              v-else
-              src="@/assets/logo/logo.svg"
-              height="50px"
-              style="width: 50px; height: 50px; display: block"
-            />
-          </div>
-
-          <el-upload
-            ref="uploadRef"
-            action="#"
-            :auto-upload="false"
-            :show-file-list="false"
-            accept="image/jpeg, image/png, image/gif"
-            :on-change="(file: any, fileList: any) => onChange(file, fileList, 'float_icon')"
-          >
-            <el-button icon="Upload">{{
-              $t('views.applicationOverview.appInfo.EditAvatarDialog.upload')
-            }}</el-button>
-            <template #tip>
-              <div class="el-upload__tip info" style="margin-top: 0">
-                建议尺寸 32*32，支持 JPG、PNG、GIF，大小不超过 10 MB
-              </div>
-            </template>
-          </el-upload>
-        </div>
-      </el-form-item>
     </el-form>
     <template #footer>
       <span class="dialog-footer">
-        <el-button v-if="user.isEnterprise()" type="primary" @click.prevent="resetForm" link
-          >恢复默认
-        </el-button>
         <el-button @click.prevent="dialogVisible = false"
           >{{ $t('views.applicationOverview.appInfo.LimitDialog.cancelButtonText') }}
         </el-button>
@@ -116,11 +32,9 @@ import { ref, watch } from 'vue'
 import { useRoute } from 'vue-router'
 import type { FormInstance, FormRules, UploadFiles } from 'element-plus'
 import applicationApi from '@/api/application'
-import applicationXpackApi from '@/api/application-xpack'
+import { isWorkFlow } from '@/utils/application'
 import { MsgSuccess, MsgError } from '@/utils/message'
 import { t } from '@/locales'
-import useStore from '@/stores'
-const { user } = useStore()
 
 const route = useRoute()
 const {
@@ -129,33 +43,13 @@ const {
 
 const emit = defineEmits(['refresh'])
 
-const defaultSetting = {
-  show_source: false,
-  show_history: true,
-  draggable: true,
-  show_guide: true,
-  avatar: '',
-  float_icon: ''
-}
 
 const displayFormRef = ref()
 const form = ref<any>({
   show_source: false
 })
 
-const xpackForm = ref<any>({
-  show_source: false,
-  show_history: false,
-  draggable: false,
-  show_guide: false,
-  avatar: '',
-  float_icon: ''
-})
-
-const imgUrl = ref<any>({
-  avatar: '',
-  float_icon: ''
-})
+const detail = ref<any>(null)
 
 const dialogVisible = ref<boolean>(false)
 const loading = ref(false)
@@ -165,50 +59,12 @@ watch(dialogVisible, (bool) => {
     form.value = {
       show_source: false
     }
-    imgUrl.value = {
-      avatar: '',
-      float_icon: ''
-    }
   }
 })
 
-function resetForm() {
-  form.value = {
-    ...defaultSetting
-  }
-  imgUrl.value = {
-    avatar: '',
-    float_icon: ''
-  }
-}
-
-const onChange = (file: any, fileList: UploadFiles, attr: string) => {
-  //1、判断文件大小是否合法，文件限制不能大于 10 MB
-  const isLimit = file?.size / 1024 / 1024 < 10
-  if (!isLimit) {
-    // @ts-ignore
-    MsgError(t('views.applicationOverview.appInfo.EditAvatarDialog.fileSizeExceeded'))
-    return false
-  } else {
-    xpackForm.value[attr] = file.raw
-    imgUrl.value[attr] = URL.createObjectURL(file.raw)
-  }
-}
-
-const open = (data: any) => {
-  if (user.isEnterprise()) {
-    xpackForm.value.show_source = data.show_source
-    xpackForm.value.show_history = data.show_history
-    xpackForm.value.draggable = data.draggable
-    xpackForm.value.show_guide = data.show_guide
-    xpackForm.value.avatar = data.avatar
-    xpackForm.value.float_icon = data.float_icon
-    imgUrl.value.avatar = data.avatar
-    imgUrl.value.float_icon = data.float_icon
-    form.value = xpackForm.value
-  } else {
-    form.value.show_source = data.show_source
-  }
+const open = (data: any, content: any) => {
+  detail.value = content
+  form.value.show_source = data.show_source
 
   dialogVisible.value = true
 }
@@ -217,28 +73,15 @@ const submit = async (formEl: FormInstance | undefined) => {
   if (!formEl) return
   await formEl.validate((valid, fields) => {
     if (valid) {
-      if (user.isEnterprise()) {
-        let fd = new FormData()
-        Object.keys(form.value).map((item) => {
-          fd.append(item, form.value[item])
-        })
-        applicationXpackApi.putAccessToken(id as string, fd, loading).then((res) => {
-          emit('refresh')
-          // @ts-ignore
-          MsgSuccess(t('views.applicationOverview.appInfo.LimitDialog.settingSuccessMessage'))
-          dialogVisible.value = false
-        })
-      } else {
-        const obj = {
-          show_source: form.value.show_source
-        }
-        applicationApi.putAccessToken(id as string, obj, loading).then((res) => {
-          emit('refresh')
-          // @ts-ignore
-          MsgSuccess(t('views.applicationOverview.appInfo.LimitDialog.settingSuccessMessage'))
-          dialogVisible.value = false
-        })
+      const obj = {
+        show_source: form.value.show_source
       }
+      applicationApi.putAccessToken(id as string, obj, loading).then((res) => {
+        emit('refresh')
+        // @ts-ignore
+        MsgSuccess(t('views.applicationOverview.appInfo.LimitDialog.settingSuccessMessage'))
+        dialogVisible.value = false
+      })
     }
   })
 }

ui/src/views/application-overview/component/XPackDisplaySettingDialog.vue

@@ -264,7 +264,10 @@
           </el-card>
 
           <el-space direction="vertical" alignment="start" :size="2">
-            <el-checkbox v-model="form.show_source" label="显示知识来源" />
+            <el-checkbox
+              v-model="form.show_source"
+              :label="isWorkFlow(detail.type) ? '显示执行详情' : '显示知识来源'"
+            />
             <el-checkbox v-model="form.show_history" label="显示历史记录" />
             <el-checkbox v-model="form.show_guide" label="显示引导图（浮窗模式）" />
             <el-checkbox v-model="form.disclaimer" label="免责声明" @change="changeDisclaimer" />
@@ -298,7 +301,7 @@
 import { computed, ref, watch } from 'vue'
 import { useRoute } from 'vue-router'
 import type { FormInstance, FormRules, UploadFiles } from 'element-plus'
-import { isAppIcon } from '@/utils/application'
+import { isAppIcon, isWorkFlow } from '@/utils/application'
 import applicationXpackApi from '@/api/application-xpack'
 import { MsgSuccess, MsgError } from '@/utils/message'
 import { t } from '@/locales'

ui/src/views/application-overview/index.vue

@@ -274,7 +274,7 @@ function openDisplaySettingDialog() {
   if (user.isEnterprise()) {
     XPackDisplaySettingDialogRef.value?.open(accessToken.value, detail.value)
   } else {
-    DisplaySettingDialogRef.value?.open(accessToken.value)
+    DisplaySettingDialogRef.value?.open(accessToken.value, detail.value)
   }
 }
 function openEditAvatar() {

ui/src/views/application-workflow/component/DropdownMenu.vue

@@ -243,7 +243,6 @@ function onmousedown(item: any, data?: any, type?: string) {
       }
     }
   }
-  console.log('xx', item)
   props.workflowRef?.onmousedown(item)
   emit('onmousedown', item)
 }

ui/src/views/document/component/EmbeddingContentDialog.vue

@@ -0,0 +1,41 @@
+<template>
+  <el-dialog v-model="dialogVisible" title="选择向量化内容" width="500" :before-close="close">
+    <el-radio-group v-model="state">
+      <el-radio value="error" size="large">向量化未成功的分段</el-radio>
+      <el-radio value="all" size="large">全部分段</el-radio>
+    </el-radio-group>
+    <template #footer>
+      <div class="dialog-footer">
+        <el-button @click="close">取消</el-button>
+        <el-button type="primary" @click="submit"> 提交 </el-button>
+      </div>
+    </template>
+  </el-dialog>
+</template>
+<script setup lang="ts">
+import { ref } from 'vue'
+const dialogVisible = ref<boolean>(false)
+const state = ref<'all' | 'error'>('error')
+const stateMap = {
+  all: ['0', '1', '2', '3', '4', '5', 'n'],
+  error: ['0', '1', '3', '4', '5', 'n']
+}
+const submit_handle = ref<(stateList: Array<string>) => void>()
+const submit = () => {
+  if (submit_handle.value) {
+    submit_handle.value(stateMap[state.value])
+  }
+  close()
+}
+
+const open = (handle: (stateList: Array<string>) => void) => {
+  submit_handle.value = handle
+  dialogVisible.value = true
+}
+const close = () => {
+  submit_handle.value = undefined
+  dialogVisible.value = false
+}
+defineExpose({ open, close })
+</script>
+<style lang="scss" scoped></style>

ui/src/views/document/index.vue

@@ -422,6 +422,7 @@
       </el-text>
       <el-button class="ml-16" type="primary" link @click="clearSelection"> 清空 </el-button>
     </div>
+    <EmbeddingContentDialog ref="embeddingContentDialogRef"></EmbeddingContentDialog>
   </LayoutContainer>
 </template>
 <script setup lang="ts">
@@ -439,6 +440,7 @@ import { MsgSuccess, MsgConfirm, MsgError } from '@/utils/message'
 import useStore from '@/stores'
 import StatusVlue from '@/views/document/component/Status.vue'
 import GenerateRelatedDialog from '@/components/generate-related-dialog/index.vue'
+import EmbeddingContentDialog from '@/views/document/component/EmbeddingContentDialog.vue'
 import { TaskType, State } from '@/utils/status'
 const router = useRouter()
 const route = useRoute()
@@ -469,7 +471,7 @@ onBeforeRouteLeave((to: any) => {
 })
 const beforePagination = computed(() => common.paginationConfig[storeKey])
 const beforeSearch = computed(() => common.search[storeKey])
-
+const embeddingContentDialogRef = ref<InstanceType<typeof EmbeddingContentDialog>>()
 const SyncWebDialogRef = ref()
 const loading = ref(false)
 let interval: any
@@ -621,10 +623,14 @@ function syncDocument(row: any) {
       .catch(() => {})
   }
 }
+
 function refreshDocument(row: any) {
-  documentApi.putDocumentRefresh(row.dataset_id, row.id).then(() => {
-    getList()
-  })
+  const embeddingDocument = (stateList: Array<string>) => {
+    return documentApi.putDocumentRefresh(row.dataset_id, row.id, stateList).then(() => {
+      getList()
+    })
+  }
+  embeddingContentDialogRef.value?.open(embeddingDocument)
 }
 
 function rowClickHandle(row: any, column: any) {
@@ -691,19 +697,16 @@ function deleteMulDocument() {
 }
 
 function batchRefresh() {
-  const arr: string[] = []
-  multipleSelection.value.map((v) => {
-    if (v) {
-      arr.push(v.id)
-    }
-  })
-  documentApi.batchRefresh(id, arr, loading).then(() => {
-    MsgSuccess('批量向量化成功')
-    multipleTableRef.value?.clearSelection()
-  })
+  const arr: string[] = multipleSelection.value.map((v) => v.id)
+  const embeddingBatchDocument = (stateList: Array<string>) => {
+    documentApi.batchRefresh(id, arr, stateList, loading).then(() => {
+      MsgSuccess('批量向量化成功')
+      multipleTableRef.value?.clearSelection()
+    })
+  }
+  embeddingContentDialogRef.value?.open(embeddingBatchDocument)
 }
 
-
 function deleteDocument(row: any) {
   MsgConfirm(
     `是否删除文档：${row.name} ?`,

ui/src/workflow/nodes/application-node/index.vue

@@ -20,7 +20,7 @@
           }"
         >
           <NodeCascader
-            ref="nodeCascaderRef"
+            ref="applicationNodeFormRef"
             :nodeModel="nodeModel"
             class="w-full"
             placeholder="请选择检索问题"
@@ -142,10 +142,10 @@
 </template>
 
 <script setup lang="ts">
-import { set, groupBy } from 'lodash'
+import { set, groupBy, create } from 'lodash'
 import { app } from '@/main'
 import NodeContainer from '@/workflow/common/NodeContainer.vue'
-import { ref, computed, onMounted } from 'vue'
+import { ref, computed, onMounted, onActivated } from 'vue'
 import NodeCascader from '@/workflow/common/NodeCascader.vue'
 import type { FormInstance } from 'element-plus'
 import applicationApi from '@/api/application'
@@ -202,7 +202,6 @@ const update_field = () => {
     .then((ok) => {
       const old_api_input_field_list = props.nodeModel.properties.node_data.api_input_field_list
       const old_user_input_field_list = props.nodeModel.properties.node_data.user_input_field_list
-
       if (isWorkFlow(ok.data.type)) {
         const nodeData = ok.data.work_flow.nodes[0].properties.node_data
         const new_api_input_field_list = ok.data.work_flow.nodes[0].properties.api_input_field_list
@@ -213,11 +212,17 @@ const update_field = () => {
             (old_item: any) => old_item.variable == item.variable
           )
           if (find_field) {
-            return { ...item, default_value: JSON.parse(JSON.stringify(find_field.default_value)) }
+            return {
+              ...item,
+              value: find_field.value,
+              label:
+                typeof item.label === 'object' && item.label != null ? item.label.label : item.label
+            }
           } else {
             return item
           }
         })
+        console.log(merge_api_input_field_list)
         set(
           props.nodeModel.properties.node_data,
           'api_input_field_list',
@@ -228,12 +233,16 @@ const update_field = () => {
             (old_item: any) => old_item.field == item.field
           )
           if (find_field) {
-            return { ...item, default_value: JSON.parse(JSON.stringify(find_field.default_value)) }
+            return {
+              ...item,
+              value: find_field.value,
+              label:
+                typeof item.label === 'object' && item.label != null ? item.label.label : item.label
+            }
           } else {
             return item
           }
         })
-        console.log(merge_user_input_field_list)
         set(
           props.nodeModel.properties.node_data,
           'user_input_field_list',
@@ -255,7 +264,7 @@ const update_field = () => {
       }
     })
     .catch((err) => {
-      set(props.nodeModel.properties, 'status', 500)
+      // set(props.nodeModel.properties, 'status', 500)
     })
 }