merge

README.md

@@ -1,4 +1,4 @@
-English | [**中文**](https://github.com/modelscope/agentscope/blob/main/README_ZH.md)
+English | [**中文**](https://github.com/modelscope/agentscope/blob/main/README_ZH.md) | [**日本語**](https://github.com/modelscope/agentscope/blob/main/README_JA.md)
 
 # AgentScope
 

README_ZH.md

@@ -1,4 +1,4 @@
-[English](https://github.com/modelscope/agentscope/blob/main/README.md) | 中文
+[English](https://github.com/modelscope/agentscope/blob/main/README.md) | 中文 | [日本語](https://github.com/modelscope/agentscope/blob/main/README_JA.md)
 
 # AgentScope
 

docs/sphinx_doc/zh_CN/source/index.rst

@@ -58,6 +58,7 @@ AgentScope 文档
    agentscope.service
    agentscope.rpc
    agentscope.server
+   agentscope.environment
    agentscope.web
    agentscope.prompt
    agentscope.utils

docs/sphinx_doc/zh_CN/source/tutorial/206-prompt.md

@@ -11,7 +11,7 @@ AgentScope内置策略的目标是**使初学者能够顺利调用模型API ，
 
 ## 构建提示面临的挑战
 
-在多智能体应用中，LLM通常在对话中扮演不同的角色。当使用模型的Chat API时，时长会面临以下挑战：
+在多智能体应用中，LLM通常在对话中扮演不同的角色。当使用模型的Chat API时，时常会面临以下挑战：
 
 1. 大多数Chat类型的模型API是为聊天机器人场景设计的，`role`字段只支持`"user"`和`"assistant"`，不支持`name`字段，即API本身不支持角色扮演。
 

examples/environments/chatroom/chatroom_example.py

@@ -59,7 +59,12 @@ def main(args: argparse.Namespace) -> None:
         ),
         role="system",
     )
-    r = ChatRoom(name="chat", announcement=ann, model_config_name=YOUR_MODEL_CONFIGURATION_NAME, to_dist=args.use_dist)
+    r = ChatRoom(
+        name="chat",
+        announcement=ann,
+        model_config_name=YOUR_MODEL_CONFIGURATION_NAME,
+        to_dist=args.use_dist,
+    )
 
     # Setup the persona of Alice, Bob and Carol
     alice = ChatRoomAgent(  # Game Art Designer

examples/environments/chatroom/chatroom_with_assistant_example.py

@@ -70,12 +70,12 @@ def main(args: argparse.Namespace) -> None:
 
     bob = ChatRoomAgentWithAssistant(
         name="Bob",
-        sys_prompt=r"""You are Bob's chat room assistant and he is """
+        sys_prompt=r"""You are Bob's chat room assistant and Bob is """
         r"""currently unable to reply to messages. Please generate a """
-        r"""suitable response based on the following chat history. """
-        r"""The content you reply to must be based on the chat history. """
-        r"""Please refuse to reply to questions that are beyond the scope """
-        r"""of the chat history.""",
+        r"""suitable response based on the following chat history without """
+        r"""reasoning. The content you reply to must be based on the chat """
+        r"""history. Please refuse to reply to questions that are beyond """
+        r"""the scope of the chat history.""",
         model_config_name=YOUR_MODEL_CONFIGURATION_NAME,
         to_dist=args.use_dist,
         timeout=args.timeout,
@@ -176,11 +176,10 @@ def main(args: argparse.Namespace) -> None:
     # Setup the persona of Carol
     carol = ChatRoomAgent(
         name="Carol",
-        sys_prompt=r"""You are Carol, and now you need to interview Bob. """
-        r"""Just ask him where he is from, which school he graduated from, """
-        r"""his profession, and his hobbies. At the end of the interview, """
-        r"""please output a reply containing Goodbye to indicate the end """
-        r"""of the conversation.""",
+        sys_prompt="""You are Carol, and now you need to interview Bob. """
+        """Just ask him where he is from, which school he graduated from, """
+        """his profession, and his hobbies. You'd better only ask one """
+        """question at a time.""",
         model_config_name=YOUR_MODEL_CONFIGURATION_NAME,
         to_dist=args.use_dist,
     )