feat: captain agent

api/agentok_api/services/codegen.py

@@ -93,7 +93,9 @@ def generate_project(self, project: Project) -> str:
         conversable_nodes = [
             node for node in flow.nodes if node["type"] == "conversable"
         ]
-        assistant_nodes = [node for node in flow.nodes if node["type"] == "assistant"]
+        ASSISTANT_NODE_TYPES = ["assistant", "captain"]
+        assistant_nodes = [node for node in flow.nodes if node["type"] in ASSISTANT_NODE_TYPES]
+        print("assistant_nodes", assistant_nodes)
         gpt_assistant_nodes = [
             node for node in flow.nodes if node["type"] == "gpt_assistant"
         ]

api/agentok_api/templates/config.j2

@@ -34,4 +34,4 @@ llm_config = {
     {%- endif %}
 }
 
-{%- endmacro %}
+{% endmacro %}

api/agentok_api/templates/import_mapping.j2

@@ -5,6 +5,7 @@
         'UserProxyAgent': 'from autogen import UserProxyAgent',
         'AssistantAgent': 'from autogen import AssistantAgent',
         'ConversableAgent': 'from autogen import ConversableAgent',
+        'CaptainAgent': 'from autogen.agentchat.contrib.captainagent import CaptainAgent',
         'CompressibleAgent': 'from autogen.agentchat.contrib.compressible_agent import CompressibleAgent',
         'GroupChat': 'from autogen import GroupChat\nfrom autogen import GroupChatManager',
         'GPTAssistantAgent': 'from autogen.agentchat.contrib.gpt_assistant_agent import GPTAssistantAgent',

api/agentok_api/templates/user.j2

@@ -1,5 +1,6 @@
 # user.j2
 {% macro generate_users(user_nodes) %}
+# User Proxy Agents
 {%- for node in user_nodes %}
 {%- set class_name = node['data']['class_type'] %}
 {%- set name = node['data']['name'] %}

api/examples/sequential-chat.py

@@ -0,0 +1,143 @@
+# This file is generated with Agentok Studio.
+# Last generated: 2025-01-10 21:58:16
+#
+# Project Name: Sequential Chat
+# Author: Unknown ([email protected])
+# Last Updated: 2025-01-10T12:55:56.49+00:00
+# Description:
+"""
+A sample flow that talks to multiple assistant, which is named Sequential Chats.
+"""
+
+
+
+from dotenv import load_dotenv, dotenv_values
+load_dotenv()  # This will load all environment variables from .env
+
+import argparse
+import os
+import time
+from termcolor import colored
+from typing import Annotated
+
+# Parse command line arguments
+parser = argparse.ArgumentParser(description='Start a chat with agents.')
+parser.add_argument('message', type=str, help='The message to send to agent.')
+args = parser.parse_args()
+
+import autogen
+
+# openai, whisper are optional dependencies
+# However, we beleive they are useful for other future examples, so we include them here as part of standard imports
+from openai import OpenAI
+from autogen import AssistantAgent
+from autogen import UserProxyAgent
+
+# Replace the default get_human_input function for status control
+def custom_get_human_input(self, prompt: str) -> str:
+    # Set wait_for_human_input to True
+    print('__STATUS_WAIT_FOR_HUMAN_INPUT__', prompt, flush=True)
+    reply = input(prompt)
+    # Restore the status to running
+    print('__STATUS_RECEIVED_HUMAN_INPUT__', prompt, flush=True)
+    return reply
+
+autogen.ConversableAgent.get_human_input = custom_get_human_input
+
+# Get the directory of the current script
+current_dir = os.path.dirname(os.path.abspath(__file__))
+
+config_list = autogen.config_list_from_json(
+    env_or_file="OAI_CONFIG_LIST",
+)
+
+llm_config = {
+    "config_list": config_list,
+}
+
+import tempfile
+temp_dir = tempfile.gettempdir()
+
+# Conversable Agents
+
+
+# Assistant Agents
+node_assistant_aycOo1y7 = AssistantAgent(
+    name="Adder",
+    system_message="""You add 1 to each number I give you and return me the new numbers, one number each line.""",
+    human_input_mode="NEVER",
+    llm_config=llm_config,
+)
+node_assistant_WOaXJRCW = AssistantAgent(
+    name="Multiplier",
+    system_message="""You multiply each number I give you by 2 and return me the new numbers, one number each line.""",
+    human_input_mode="NEVER",
+    llm_config=llm_config,
+)
+node_assistant_UZj4G_LO = AssistantAgent(
+    name="Divider",
+    system_message="""You divide each number I give you by 2 and return me the new numbers, one number each line.""",
+    human_input_mode="NEVER",
+    llm_config=llm_config,
+)
+node_assistant_Vk4rDF6b = AssistantAgent(
+    name="Subtractor",
+    system_message="""You subtract 1 from each number I give you and return me the new numbers, one number each line.""",
+    human_input_mode="NEVER",
+    llm_config=llm_config,
+)
+node_user_7j_1RHfz = UserProxyAgent(
+    name="Number",
+    system_message="""You return me the numbers I give you, one number each line.""",
+    human_input_mode="NEVER",
+    code_execution_config={ # Make code excution always available for user_proxy nodes
+        "executor": autogen.coding.LocalCommandLineCodeExecutor(
+            work_dir=os.path.join(temp_dir, "user_code"),
+        )
+    },
+)
+# Group Chats
+# Nested Chats
+# Tools
+# Start the conversation
+
+
+# Sequential Chats
+chat_results = node_user_7j_1RHfz.initiate_chats(
+    [
+      {
+        "recipient": node_assistant_aycOo1y7,
+        "message": args.message,
+        "max_turns": 2,
+        "summary_method": "last_msg",
+      },
+      {
+        "recipient": node_assistant_WOaXJRCW,
+        "message": "These are my numbers.",
+        "max_turns": 2,
+        "summary_method": "last_msg",
+      },
+      {
+        "recipient": node_assistant_UZj4G_LO,
+        "message": "These are my numbers.",
+        "max_turns": 2,
+        "summary_method": "last_msg",
+      },
+      {
+        "recipient": node_assistant_Vk4rDF6b,
+        "message": "These are my numbers.",
+        "max_turns": 2,
+        "summary_method": "last_msg",
+      }
+    ]
+)
+
+# Output the sequential chat results
+import json
+print("__CHAT_RESULTS__", json.dumps([{
+    "chat_id": result.chat_id,
+    "chat_history": result.chat_history,
+    "summary": result.summary,
+    "cost": result.cost,
+    "human_input": result.human_input
+} for result in chat_results]))