feat: add bedrock api

source/infrastructure/cli/magic-config.ts

@@ -36,7 +36,19 @@ const embeddingModels = [
     name: "bce-embedding-and-bge-reranker",
     commitId: "43972580a35ceacacd31b95b9f430f695d07dde9",
     dimensions: 768,
-  }
+  },
+  {
+    provider: "OpenAI API",
+    name: "text-embedding-3-small",
+    commitId: "",
+    dimensions: 1536,
+  },
+  {
+    provider: "OpenAI API",
+    name: "text-embedding-3-large",
+    commitId: "",
+    dimensions: 3072,
+  },
 ];
 
 const apiInferenceProviders = [

source/lambda/etl/chatbot_management.py

@@ -1,7 +1,6 @@
 import json
 import logging
 import os
-import time
 from datetime import datetime, timezone
 
 import boto3
@@ -13,6 +12,8 @@
     initiate_index,
     initiate_model
 )
+from constant import ModelProvider
+
 
 logger = logging.getLogger()
 logger.setLevel(logging.INFO)
@@ -52,6 +53,9 @@
 def create_chatbot(event, group_name):
     request_body = json.loads(event["body"])
     chatbot_id = request_body.get("chatbotId", group_name.lower())
+    model_provider = request_body.get("modelProvider", ModelProvider.BEDROCK.value)
+    base_url = request_body.get("baseUrl", "")
+    api_key_arn = request_body.get("apiKeyArn", "")
     chatbot_description = request_body.get(
         "chatbotDescription", "Answer question based on search result"
     )
@@ -61,7 +65,15 @@ def create_chatbot(event, group_name):
     create_time = str(datetime.now(timezone.utc))
 
     model_type = initiate_model(
-        model_table, group_name, model_id, chatbot_embedding, create_time)
+        model_table,
+        group_name,
+        model_id,
+        chatbot_embedding,
+        model_provider,
+        base_url,
+        api_key_arn,
+        create_time
+    )
     index = request_body.get("index", {"qq":{"admin-qq-default": "Answer question based on search result"},"qd":{"admin-qd-default": "Answer question based on search result"},"intention":{"admin-intention-default": "Answer question based on search result"}})
     for index_type in index:
         index_ids = list(index[index_type].keys())
@@ -142,6 +154,7 @@ def __list_chatbot(event, group_name):
             item_json["ModelId"] = chatbot_model_item.get("modelId", "")
             item_json["LastModifiedTime"] = item.get(
                 "updateTime", {"S": ""})["S"]
+            item_json["ModelProvider"] = chatbot_model_item["parameter"].get("ModelProvider", "")
             page_json.append(item_json)
         page_json.sort(key=lambda x: x["LastModifiedTime"], reverse=True)
         output["Items"] = page_json
@@ -180,6 +193,8 @@ def __get_chatbot(event, group_name):
         model = model_item.get("parameter", {})
         model_endpoint = model.get("ModelEndpoint", {})
         model_name = model.get("ModelName", {})
+        model_provider = model.get("ModelProvider", "")
+        base_url = model.get("BaseUrl", "")
         chatbot_index = []
         for key, value in chatbot_index_ids.items():
             v = value.get('value',{})
@@ -201,7 +216,9 @@ def __get_chatbot(event, group_name):
             "updateTime": update_time,
             "model": {
                 "model_endpoint": model_endpoint,
-                "model_name": model_name
+                "model_name": model_name,
+                "model_provider": model_provider,
+                "base_url": base_url
             },
             "index": chatbot_index,
         }

source/lambda/etl/constant.py

@@ -52,3 +52,11 @@ class OperationType(Enum):
     UPDATE = "update"
     DELETE = "delete"
     EXTRACT_ONLY = "extract_only"
+
+
+@unique
+class ModelProvider(Enum):
+    DMAA = "dmaa"
+    BEDROCK = "Bedrock"
+    BRCONNECTOR_BEDROCK = "Bedrock API"
+    OPENAI = "OpenAI API"

source/lambda/etl/utils/ddb_utils.py

@@ -31,14 +31,21 @@ def update_model(model_table, item_key, model_parameter):
     )
 
 
-def initiate_model(model_table, group_name, model_id, embedding_endpoint, create_time=None):
+def initiate_model(model_table, group_name, model_id, embedding_endpoint, model_provider, base_url, api_key_arn, create_time=None):
     existing_item = item_exist(model_table, {"groupName": group_name, "modelId": model_id})
     embedding_info = get_embedding_info(embedding_endpoint)
     embedding_info["ModelEndpoint"] = embedding_endpoint
+    embedding_info["ModelProvider"] = model_provider
+    embedding_info["BaseUrl"] = base_url
+    embedding_info["ApiKeyArn"] = api_key_arn
+
     if existing_item:
         existing_embedding_endpoint = existing_item["parameter"]["ModelEndpoint"]
         if existing_embedding_endpoint != embedding_endpoint:
             embedding_info = get_embedding_info(embedding_endpoint)
+            embedding_info["ModelProvider"] = model_provider
+            embedding_info["BaseUrl"] = base_url
+            embedding_info["ApiKeyArn"] = api_key_arn
             update_model(
                 model_table,
                 {"groupName": group_name, "modelId": model_id},

source/lambda/etl/utils/embeddings.py

@@ -20,13 +20,6 @@ def get_embedding_info(embedding_endpoint_name):
         embeddings_model_name = "bge-m3"
         embeddings_model_dimensions = 1024
         embeddings_model_type = "m3"
-
-    elif "embedding" in embedding_endpoint_name:
-        embeddings_model_provider = "Netease"
-        embeddings_model_name = "bce_embedding_model.tar.gz"
-        embeddings_model_dimensions = 768
-        embeddings_model_type = "bce"
-
     elif "cohere" in embedding_endpoint_name:
         embeddings_model_provider = "Cohere"
         embeddings_model_name = "cohere.embed-english-v3"
@@ -42,6 +35,21 @@ def get_embedding_info(embedding_endpoint_name):
         embeddings_model_name = "amazon.titan-embed-text-v2:0"
         embeddings_model_dimensions = 1024
         embeddings_model_type = "bedrock"
+    elif "text-embedding-3-small" in embedding_endpoint_name:
+        embeddings_model_provider = "OpenAI API"
+        embeddings_model_name = "text-embedding-3-small"
+        embeddings_model_dimensions = 1536
+        embeddings_model_type = "OpenAI API"
+    elif "text-embedding-3-large" in embedding_endpoint_name:
+        embeddings_model_provider = "OpenAI API"
+        embeddings_model_name = "text-embedding-3-large"
+        embeddings_model_dimensions = 3072
+        embeddings_model_type = "OpenAI API"  
+    elif "embedding" in embedding_endpoint_name:
+        embeddings_model_provider = "Netease"
+        embeddings_model_name = "bce_embedding_model.tar.gz"
+        embeddings_model_dimensions = 768
+        embeddings_model_type = "bce"
     else:
         embeddings_model_provider = "Not Found"
         embeddings_model_name = "Not Found"

source/lambda/job/dep/llm_bot_dep/sm_utils.py

@@ -25,10 +25,9 @@
 
 logger = logging.getLogger()
 logger.setLevel(logging.INFO)
-region_name = os.environ["AWS_REGION"]
 session = boto3.session.Session()
 secret_manager_client = session.client(
-    service_name="secretsmanager", region_name=region_name
+    service_name="secretsmanager"
 )
 
 

source/portal/src/locale/en.json

@@ -1,8 +1,11 @@
 {
   "translation": {
-    "name": "AI Customer Service",
+    "name": "Multi-modal Enterprise Knowledge Base in Manufacturing",
+    "welcome": "Welcome to Multi-modal Enterprise Knowledge Base in Manufacturing",
+    "mead": "Manufacturing",
+    "projectDescription": "Multi-modal Enterprise Knowledge Base in Manufacturing offers a streamlined workflow for developing scalable, production-grade agent-based applications",
+    "solutionName": "Multi-modal Enterprise Knowledge Base in Manufacturing",
     "desc": "Description",
-    "welcome": "Welcome to AI Customer Service",
     "you": "Y o u",
     "welcomeMessage": "Hello, how can I help you today?",
     "server": "Server",
@@ -98,6 +101,7 @@
     "noChatbotFound": "No chatbot found",
     "scenario": "Scenario",
     "modelType": "Model provider",
+    "modelProvider": "Model provider",
     "maxTokens": "Max Tokens",
     "temperature": "Temperature",
     "sessionHistory": "Recent chats",
@@ -175,7 +179,7 @@
     "referenceDocuments": "Reference Documents",
     "selectADocumentToPreview": "Select a document to preview its content",
     "button": {
-      "login": "Log in to AI Customer Service",
+      "login": "Log in to Multi-modal Enterprise Knowledge Base in Manufacturing",
       "reload": "Reload",
       "send": "Send",
       "modelSettings": "Model Settings",
@@ -247,11 +251,8 @@
       "next": "Next",
       "skip": "Skip tour"
     },
-    "solutionName": "AI Customer Service",
     "subTitle": "Streamlined Workflow for Building Agent-Based Applications",
-    "projectDescription": "AI Customer Service offers a streamlined workflow for developing scalable, production-grade agent-based applications",
     "awsSolutionGuidance": "Amazon Web Services Solution Guidance",
-    "mead": "Media & Advertisement",
     "homeSidebar": "Home",
     "chatbotDetail": "ChatBot Detail",
     "featuresAndBenefits": {
@@ -266,18 +267,18 @@
       },
       "feat3": {
         "name": "Intention detection",
-        "desc": "desc"
+        "desc": "Desc"
       },
       "feat4": {
         "name": "Management and governance",
-        "desc": "desc"
+        "desc": "Desc"
       }
     },
     "useCases": {
       "name": "Use cases",
       "case1": {
-        "name": "MEAD use case",
-        "desc": "desc"
+        "name": "Use case",
+        "desc": "Desc"
       }
     },
     "gettingStarted": {

source/portal/src/locale/zh.json

@@ -1,8 +1,11 @@
 {
   "translation": {
-    "name": "AI Customer Service",
+    "name": "Multi-modal Enterprise Knowledge Base in Manufacturing",
+    "welcome": "欢迎使用 Multi-modal Enterprise Knowledge Base in Manufacturing",
+    "mead": "制造",
+    "projectDescription": "Multi-modal Enterprise Knowledge Base in Manufacturing 提供一个高效简洁的工作流程，用于开发可扩展的、生产级别的、基于 agent（代理）的应用",
+    "solutionName": "Multi-modal Enterprise Knowledge Base in Manufacturing",
     "desc": "描述",
-    "welcome": "欢迎使用 AI Customer Service",
     "you": "你",
     "welcomeMessage": "您好，我今天有什么可以帮助您？",
     "server": "服务器",
@@ -108,6 +111,7 @@
     "embeddingModelDesc": "用于此机器人的向量化操作",
     "scenario": "场景",
     "modelType": "模型供应商",
+    "modelProvider": "模型供应商",
     "maxTokens": "最大 Token 数",
     "temperature": "温度",
     "sessionHistory": "历史记录",
@@ -240,11 +244,8 @@
       "next": "下一步",
       "skip": "跳过向导"
     },
-    "solutionName": "AI Customer Service",
     "subTitle": "构建基于代理的应用程序的优化工作流",
-    "projectDescription": "AI Customer Service提供一个高效简洁的工作流程，用于开发可扩展的、生产级别的、基于 agent（代理）的应用",
     "awsSolutionGuidance": "Amazon Web Services 解决方案指南",
-    "mead": "媒体与广告",
     "homeSidebar": "首页",
     "chatbotDetail": "机器人详情",
     "featuresAndBenefits": {
@@ -259,18 +260,18 @@
       },
       "feat3": {
         "name": "Intention detection",
-        "desc": "desc"
+        "desc": "Desc"
       },
       "feat4": {
         "name": "Management and governance",
-        "desc": "desc"
+        "desc": "Desc"
       }
     },
     "useCases": {
       "name": "使用案例",
       "case1": {
-        "name": "MEAD use case",
-        "desc": "desc"
+        "name": "Use case",
+        "desc": "Desc"
       }
     },
     "gettingStarted": {

source/portal/src/pages/chatbot/ChatBot.tsx

@@ -119,9 +119,7 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
   const [currentAIMessageId, setCurrentAIMessageId] = useState('');
   const [aiSpeaking, setAiSpeaking] = useState(false);
   const [modelOption, setModelOption] = useState('');
-  // const [brModelOption, setBrApiModelOption] = useState('');
   const [modelList, setModelList] = useState<SelectProps.Option[]>([]);
-  // const [brApiModelList, setBrApiModelList] = useState<SelectProps.Option[]>([]);
   const [chatbotList, setChatbotList] = useState<SelectProps.Option[]>([]);
   const [chatbotOption, setChatbotOption] = useState<SelectProps.Option>(null as any);
   const [useChatHistory, setUseChatHistory] = useState(localStorage.getItem(USE_CHAT_HISTORY) == null || localStorage.getItem(USE_CHAT_HISTORY) == "true" ? true : false);
@@ -774,7 +772,7 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
                 <div style={{fontSize: 16, fontWeight: 700, marginBottom: 15, marginTop: 15}}>{t('common')}</div>
                 <SpaceBetween size="xs" direction="vertical">
                 <Grid gridDefinition={[{colspan: 5},{colspan: 6}]}>
-                  <FormField label={t('modelType')} stretch={true} description={t('scenarioDesc')}>
+                  <FormField label={t('modelProvider')} stretch={true} description={t('scenarioDesc')}>
                     <Select
                       options={MODEL_TYPE_LIST}
                       selectedOption={modelType}

source/portal/src/pages/chatbotManagement/ChatbotDetail.tsx

@@ -224,7 +224,9 @@ const ChatbotDetail: React.FC = () => {
         chatbotId: data.chatbotId,
         model: data.model?.model_name,
         index: data.index,
-        updateTime: data.updateTime
+        updateTime: data.updateTime,
+        modelProvider: data.model?.model_provider,
+        baseUrl: data.model?.base_url
       };
       setChatbotDetail(chatbotDetail)
       // setIndexList(chatbotDetail.index)
@@ -277,15 +279,35 @@ const ChatbotDetail: React.FC = () => {
           <SpaceBetween direction="vertical" size="xs">
 
             <div className="mt-10">
-            <Grid
-      gridDefinition={[{ colspan: 4 }, { colspan: 4 }, { colspan: 4 }]}
+    <Grid
+      gridDefinition={[
+        { colspan: 3 },
+        { colspan: 3 },
+        { colspan: 3 },
+        { colspan: 3 },
+        { colspan: 3 }  
+      ]}
     >
       <div>
       <FormField
       description={id}
 
       label={t("chatbotName")}
     >
+    </FormField>
+      </div>
+      <div>
+      <FormField
+      description={chatbotDetail?.modelProvider}
+      label={t("modelProvider")}
+    >
+    </FormField>
+      </div>
+      <div>
+      <FormField
+      description={chatbotDetail?.baseUrl}
+      label={t("apiEndpoint")}
+    >
     </FormField>
       </div>
       <div>