feat: application i18n (#2020)

Author: shaohuzhang1

apps/application/chat_pipeline/step/chat_step/i_chat_step.py

@@ -9,6 +9,7 @@
 from abc import abstractmethod
 from typing import Type, List
 
+from django.utils.translation import gettext_lazy as _
 from langchain.chat_models.base import BaseChatModel
 from langchain.schema import BaseMessage
 from rest_framework import serializers
@@ -23,7 +24,7 @@
 class ModelField(serializers.Field):
     def to_internal_value(self, data):
         if not isinstance(data, BaseChatModel):
-            self.fail('模型类型错误', value=data)
+            self.fail(_('Model type error'), value=data)
         return data
 
     def to_representation(self, value):
@@ -33,7 +34,7 @@ def to_representation(self, value):
 class MessageField(serializers.Field):
     def to_internal_value(self, data):
         if not isinstance(data, BaseMessage):
-            self.fail('message类型错误', value=data)
+            self.fail(_('Message type error'), value=data)
         return data
 
     def to_representation(self, value):
@@ -52,37 +53,37 @@ class IChatStep(IBaseChatPipelineStep):
     class InstanceSerializer(serializers.Serializer):
         # 对话列表
         message_list = serializers.ListField(required=True, child=MessageField(required=True),
-                                             error_messages=ErrMessage.list("对话列表"))
-        model_id = serializers.UUIDField(required=False, allow_null=True, error_messages=ErrMessage.uuid("模型id"))
+                                             error_messages=ErrMessage.list(_("Conversation list")))
+        model_id = serializers.UUIDField(required=False, allow_null=True, error_messages=ErrMessage.uuid(_("Model id")))
         # 段落列表
-        paragraph_list = serializers.ListField(error_messages=ErrMessage.list("段落列表"))
+        paragraph_list = serializers.ListField(error_messages=ErrMessage.list(_("Paragraph List")))
         # 对话id
-        chat_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid("对话id"))
+        chat_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid(_("Conversation ID")))
         # 用户问题
-        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.uuid("用户问题"))
+        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.uuid(_("User Questions")))
         # 后置处理器
         post_response_handler = InstanceField(model_type=PostResponseHandler,
-                                              error_messages=ErrMessage.base("用户问题"))
+                                              error_messages=ErrMessage.base(_("Post-processor")))
         # 补全问题
-        padding_problem_text = serializers.CharField(required=False, error_messages=ErrMessage.base("补全问题"))
+        padding_problem_text = serializers.CharField(required=False, error_messages=ErrMessage.base(_("Completion Question")))
         # 是否使用流的形式输出
-        stream = serializers.BooleanField(required=False, error_messages=ErrMessage.base("流式输出"))
-        client_id = serializers.CharField(required=True, error_messages=ErrMessage.char("客户端id"))
-        client_type = serializers.CharField(required=True, error_messages=ErrMessage.char("客户端类型"))
+        stream = serializers.BooleanField(required=False, error_messages=ErrMessage.base(_("Streaming Output")))
+        client_id = serializers.CharField(required=True, error_messages=ErrMessage.char(_("Client id")))
+        client_type = serializers.CharField(required=True, error_messages=ErrMessage.char(_("Client Type")))
         # 未查询到引用分段
-        no_references_setting = NoReferencesSetting(required=True, error_messages=ErrMessage.base("无引用分段设置"))
+        no_references_setting = NoReferencesSetting(required=True, error_messages=ErrMessage.base(_("No reference segment settings")))
 
-        user_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid("用户id"))
+        user_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid(_("User ID")))
 
         model_params_setting = serializers.DictField(required=False, allow_null=True,
-                                                     error_messages=ErrMessage.dict("模型参数设置"))
+                                                     error_messages=ErrMessage.dict(_("Model parameter settings")))
 
         def is_valid(self, *, raise_exception=False):
             super().is_valid(raise_exception=True)
             message_list: List = self.initial_data.get('message_list')
             for message in message_list:
                 if not isinstance(message, BaseMessage):
-                    raise Exception("message 类型错误")
+                    raise Exception(_("message type error"))
 
     def get_step_serializer(self, manage: PipelineManage) -> Type[serializers.Serializer]:
         return self.InstanceSerializer

apps/application/chat_pipeline/step/chat_step/impl/base_chat_step.py

@@ -26,6 +26,7 @@
 from application.models.api_key_model import ApplicationPublicAccessClient
 from common.constants.authentication_type import AuthenticationType
 from setting.models_provider.tools import get_model_instance_by_model_user_id
+from django.utils.translation import gettext_lazy as _
 
 
 def add_access_num(client_id=None, client_type=None, application_id=None):
@@ -172,7 +173,8 @@ def get_stream_result(message_list: List[BaseMessage],
             return iter(
                 [AIMessageChunk(content=no_references_setting.get('value').replace('{question}', problem_text))]), False
         if chat_model is None:
-            return iter([AIMessageChunk('抱歉，没有配置 AI 模型，请先去应用中设置 AI 模型。')]), False
+            return iter([AIMessageChunk(
+                _('Sorry, the AI model is not configured. Please go to the application to set up the AI model first.'))]), False
         else:
             return chat_model.stream(message_list), True
 
@@ -216,7 +218,7 @@ def get_block_result(message_list: List[BaseMessage],
                 'status') == 'designated_answer':
             return AIMessage(no_references_setting.get('value').replace('{question}', problem_text)), False
         if chat_model is None:
-            return AIMessage('抱歉，没有配置 AI 模型，请先去应用中设置 AI 模型。'), False
+            return AIMessage(_('Sorry, the AI model is not configured. Please go to the application to set up the AI model first.')), False
         else:
             return chat_model.invoke(message_list), True
 

apps/application/chat_pipeline/step/generate_human_message_step/i_generate_human_message_step.py

@@ -9,6 +9,7 @@
 from abc import abstractmethod
 from typing import Type, List
 
+from django.utils.translation import gettext_lazy as _
 from langchain.schema import BaseMessage
 from rest_framework import serializers
 
@@ -23,26 +24,26 @@
 class IGenerateHumanMessageStep(IBaseChatPipelineStep):
     class InstanceSerializer(serializers.Serializer):
         # 问题
-        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.char("问题"))
+        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.char(_("question")))
         # 段落列表
         paragraph_list = serializers.ListField(child=InstanceField(model_type=ParagraphPipelineModel, required=True),
-                                               error_messages=ErrMessage.list("段落列表"))
+                                               error_messages=ErrMessage.list(_("Paragraph List")))
         # 历史对答
         history_chat_record = serializers.ListField(child=InstanceField(model_type=ChatRecord, required=True),
-                                                    error_messages=ErrMessage.list("历史对答"))
+                                                    error_messages=ErrMessage.list(_("History Questions")))
         # 多轮对话数量
-        dialogue_number = serializers.IntegerField(required=True, error_messages=ErrMessage.integer("多轮对话数量"))
+        dialogue_number = serializers.IntegerField(required=True, error_messages=ErrMessage.integer(_("Number of multi-round conversations")))
         # 最大携带知识库段落长度
         max_paragraph_char_number = serializers.IntegerField(required=True, error_messages=ErrMessage.integer(
-            "最大携带知识库段落长度"))
+            _("Maximum length of the knowledge base paragraph")))
         # 模板
-        prompt = serializers.CharField(required=True, error_messages=ErrMessage.char("提示词"))
+        prompt = serializers.CharField(required=True, error_messages=ErrMessage.char(_("Prompt word")))
         system = serializers.CharField(required=False, allow_null=True, allow_blank=True,
-                                       error_messages=ErrMessage.char("系统提示词(角色)"))
+                                       error_messages=ErrMessage.char(_("System prompt words (role)")))
         # 补齐问题
-        padding_problem_text = serializers.CharField(required=False, error_messages=ErrMessage.char("补齐问题"))
+        padding_problem_text = serializers.CharField(required=False, error_messages=ErrMessage.char(_("Completion problem")))
         # 未查询到引用分段
-        no_references_setting = NoReferencesSetting(required=True, error_messages=ErrMessage.base("无引用分段设置"))
+        no_references_setting = NoReferencesSetting(required=True, error_messages=ErrMessage.base(_("No reference segment settings")))
 
     def get_step_serializer(self, manage: PipelineManage) -> Type[serializers.Serializer]:
         return self.InstanceSerializer

apps/application/chat_pipeline/step/reset_problem_step/i_reset_problem_step.py

@@ -9,12 +9,11 @@
 from abc import abstractmethod
 from typing import Type, List
 
-from langchain.chat_models.base import BaseChatModel
+from django.utils.translation import gettext_lazy as _
 from rest_framework import serializers
 
 from application.chat_pipeline.I_base_chat_pipeline import IBaseChatPipelineStep
 from application.chat_pipeline.pipeline_manage import PipelineManage
-from application.chat_pipeline.step.chat_step.i_chat_step import ModelField
 from application.models import ChatRecord
 from common.field.common import InstanceField
 from common.util.field_message import ErrMessage
@@ -23,15 +22,16 @@
 class IResetProblemStep(IBaseChatPipelineStep):
     class InstanceSerializer(serializers.Serializer):
         # 问题文本
-        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.float("问题文本"))
+        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.float(_("question")))
         # 历史对答
         history_chat_record = serializers.ListField(child=InstanceField(model_type=ChatRecord, required=True),
-                                                    error_messages=ErrMessage.list("历史对答"))
+                                                    error_messages=ErrMessage.list(_("History Questions")))
         # 大语言模型
-        model_id = serializers.UUIDField(required=False, allow_null=True, error_messages=ErrMessage.uuid("模型id"))
-        user_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid("用户id"))
+        model_id = serializers.UUIDField(required=False, allow_null=True, error_messages=ErrMessage.uuid(_("Model id")))
+        user_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid(_("User ID")))
         problem_optimization_prompt = serializers.CharField(required=False, max_length=102400,
-                                                            error_messages=ErrMessage.char("问题补全提示词"))
+                                                            error_messages=ErrMessage.char(
+                                                                _("Question completion prompt")))
 
     def get_step_serializer(self, manage: PipelineManage) -> Type[serializers.Serializer]:
         return self.InstanceSerializer

apps/application/chat_pipeline/step/reset_problem_step/impl/base_reset_problem_step.py

@@ -14,9 +14,10 @@
 from application.models import ChatRecord
 from common.util.split_model import flat_map
 from setting.models_provider.tools import get_model_instance_by_model_user_id
+from django.utils.translation import gettext_lazy as _
 
-prompt = (
-    '()里面是用户问题,根据上下文回答揣测用户问题({question}) 要求: 输出一个补全问题,并且放在<data></data>标签中')
+prompt = _(
+    "() contains the user's question. Answer the guessed user's question based on the context ({question}) Requirement: Output a complete question and put it in the <data></data> tag")
 
 
 class BaseResetProblemStep(IResetProblemStep):

apps/application/chat_pipeline/step/search_dataset_step/i_search_dataset_step.py

@@ -11,6 +11,7 @@
 from typing import List, Type
 
 from django.core import validators
+from django.utils.translation import gettext_lazy as _
 from rest_framework import serializers
 
 from application.chat_pipeline.I_base_chat_pipeline import IBaseChatPipelineStep, ParagraphPipelineModel
@@ -21,29 +22,30 @@
 class ISearchDatasetStep(IBaseChatPipelineStep):
     class InstanceSerializer(serializers.Serializer):
         # 原始问题文本
-        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.char("问题"))
+        problem_text = serializers.CharField(required=True, error_messages=ErrMessage.char(_("question")))
         # 系统补全问题文本
-        padding_problem_text = serializers.CharField(required=False, error_messages=ErrMessage.char("系统补全问题文本"))
+        padding_problem_text = serializers.CharField(required=False,
+                                                     error_messages=ErrMessage.char(_("System completes question text")))
         # 需要查询的数据集id列表
         dataset_id_list = serializers.ListField(required=True, child=serializers.UUIDField(required=True),
-                                                error_messages=ErrMessage.list("数据集id列表"))
+                                                error_messages=ErrMessage.list(_("Dataset id list")))
         # 需要排除的文档id
         exclude_document_id_list = serializers.ListField(required=True, child=serializers.UUIDField(required=True),
-                                                         error_messages=ErrMessage.list("排除的文档id列表"))
+                                                         error_messages=ErrMessage.list(_("List of document ids to exclude")))
         # 需要排除向量id
         exclude_paragraph_id_list = serializers.ListField(required=True, child=serializers.UUIDField(required=True),
-                                                          error_messages=ErrMessage.list("排除向量id列表"))
+                                                          error_messages=ErrMessage.list(_("List of exclusion vector ids")))
         # 需要查询的条数
         top_n = serializers.IntegerField(required=True,
-                                         error_messages=ErrMessage.integer("引用分段数"))
+                                         error_messages=ErrMessage.integer(_("Reference segment number")))
         # 相似度 0-1之间
         similarity = serializers.FloatField(required=True, max_value=1, min_value=0,
-                                            error_messages=ErrMessage.float("引用分段数"))
+                                            error_messages=ErrMessage.float(_("Similarity")))
         search_mode = serializers.CharField(required=True, validators=[
             validators.RegexValidator(regex=re.compile("^embedding|keywords|blend$"),
-                                      message="类型只支持register|reset_password", code=500)
-        ], error_messages=ErrMessage.char("检索模式"))
-        user_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid("用户id"))
+                                      message=_("The type only supports register|reset_password"), code=500)
+        ], error_messages=ErrMessage.char(_("Retrieval Mode")))
+        user_id = serializers.UUIDField(required=True, error_messages=ErrMessage.uuid(_("User ID")))
 
     def get_step_serializer(self, manage: PipelineManage) -> Type[InstanceSerializer]:
         return self.InstanceSerializer

apps/application/chat_pipeline/step/search_dataset_step/impl/base_search_dataset_step.py

@@ -10,6 +10,8 @@
 from typing import List, Dict
 
 from django.db.models import QuerySet
+from django.utils.translation import gettext_lazy as _
+from rest_framework.utils.formatting import lazy_format
 
 from application.chat_pipeline.I_base_chat_pipeline import ParagraphPipelineModel
 from application.chat_pipeline.step.search_dataset_step.i_search_dataset_step import ISearchDatasetStep
@@ -26,18 +28,19 @@
 def get_model_by_id(_id, user_id):
     model = QuerySet(Model).filter(id=_id).first()
     if model is None:
-        raise Exception("模型不存在")
+        raise Exception(_("Model does not exist"))
     if model.permission_type == 'PRIVATE' and str(model.user_id) != str(user_id):
-        raise Exception(f"无权限使用此模型:{model.name}")
+        message = lazy_format(_('No permission to use this model {model_name}'), model_name=model.name)
+        raise Exception(message)
     return model
 
 
 def get_embedding_id(dataset_id_list):
     dataset_list = QuerySet(DataSet).filter(id__in=dataset_id_list)
     if len(set([dataset.embedding_mode_id for dataset in dataset_list])) > 1:
-        raise Exception("关联知识库的向量模型不一致，无法召回分段。")
+        raise Exception(_("The vector model of the associated knowledge base is inconsistent and the segmentation cannot be recalled."))
     if len(dataset_list) == 0:
-        raise Exception("知识库设置错误,请重新设置知识库")
+        raise Exception(_("The knowledge base setting is wrong, please reset the knowledge base"))
     return dataset_list[0].embedding_mode_id
 
 

apps/application/flow/step_node/ai_chat_step_node/i_chat_node.py

@@ -8,26 +8,30 @@
 """
 from typing import Type
 
+from django.utils.translation import gettext_lazy as _
 from rest_framework import serializers
 
 from application.flow.i_step_node import INode, NodeResult
 from common.util.field_message import ErrMessage
 
 
 class ChatNodeSerializer(serializers.Serializer):
-    model_id = serializers.CharField(required=True, error_messages=ErrMessage.char("模型id"))
+    model_id = serializers.CharField(required=True, error_messages=ErrMessage.char(_("Model id")))
     system = serializers.CharField(required=False, allow_blank=True, allow_null=True,
-                                   error_messages=ErrMessage.char("角色设定"))
-    prompt = serializers.CharField(required=True, error_messages=ErrMessage.char("提示词"))
+                                   error_messages=ErrMessage.char(_("Role Setting")))
+    prompt = serializers.CharField(required=True, error_messages=ErrMessage.char(_("Prompt word")))
     # 多轮对话数量
-    dialogue_number = serializers.IntegerField(required=True, error_messages=ErrMessage.integer("多轮对话数量"))
+    dialogue_number = serializers.IntegerField(required=True, error_messages=ErrMessage.integer(
+        _("Number of multi-round conversations")))
 
-    is_result = serializers.BooleanField(required=False, error_messages=ErrMessage.boolean('是否返回内容'))
+    is_result = serializers.BooleanField(required=False,
+                                         error_messages=ErrMessage.boolean(_('Whether to return content')))
 
-    model_params_setting = serializers.DictField(required=False, error_messages=ErrMessage.integer("模型参数相关设置"))
+    model_params_setting = serializers.DictField(required=False,
+                                                 error_messages=ErrMessage.integer(_("Model parameter settings")))
 
     dialogue_type = serializers.CharField(required=False, allow_blank=True, allow_null=True,
-                                          error_messages=ErrMessage.char("上下文类型"))
+                                          error_messages=ErrMessage.char(_("Context Type")))
 
 
 class IChatNode(INode):