Merge branch 'main' into remove-paths-filter

ag2ai · Feb 4, 2025 · 6dec31b · 6dec31b
2 parents 387dc16 + cc74423
commit 6dec31b
Show file tree

Hide file tree

Showing 5 changed files with 163 additions and 30 deletions.
diff --git a/autogen/agentchat/realtime/experimental/realtime_swarm.py b/autogen/agentchat/realtime/experimental/realtime_swarm.py
@@ -107,6 +107,8 @@ def parse_oai_message(message: Union[dict[str, Any], str], role: str, adressee:
 class SwarmableAgent:
     """A class for an agent that can participate in a swarm chat."""
 
+    __exported_module__ = ""
+
     def __init__(
         self,
         name: str,

diff --git a/notebook/agents_websurfer.ipynb b/notebook/agents_websurfer.ipynb
@@ -29,6 +29,7 @@
     "   \n",
     "   ```bash\n",
     "   playwright install\n",
+    "   playwright install-deps\n",
     "   ```\n",
     "\n",
     "3. For running the code in Jupyther, use `nest_asyncio` to allow nested event loops.\n",
@@ -76,16 +77,48 @@
     "\n",
     "llm_config = {\n",
     "    \"config_list\": config_list,\n",
-    "}\n",
+    "}"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "There are two ways to start a chat session which is using only one agent with LLM configuration.\n",
     "\n",
-    "user_proxy = UserProxyAgent(name=\"user_proxy\", human_input_mode=\"NEVER\")\n",
+    "#### **Recommended:** Using the `run` Method\n",
+    "\n",
+    "The new `run` method simplifies the process by eliminating the need for manual `UserProxyAgent` creation.\n",
+    "\n",
+    "- ✅ **Easier setup** – No need to manually register tools"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# The `web_tool=\"browser_use\"` tells the agent to use the `BrowserUseTool` to surf the web.\n",
     "websurfer = WebSurferAgent(name=\"WebSurfer\", llm_config=llm_config, web_tool=\"browser_use\")\n",
     "\n",
-    "websurfer_tools = websurfer.tools\n",
-    "# WebSurferAgent has a list of tools which are registered for LLM\n",
-    "# We need to register the tools for execution with the UserProxyAgent\n",
-    "for tool in websurfer_tools:\n",
-    "    tool.register_for_execution(user_proxy)"
+    "websurfer.run(\n",
+    "    message=\"Get info from https://docs.ag2.ai/docs/Home\",\n",
+    "    tools=websurfer.tools,\n",
+    "    max_turns=2,\n",
+    "    user_input=False,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "#### **Manual Setup:** Using `initiate_chat` Method\n",
+    "This method requires manually creating a `UserProxyAgent` and registering tools for execution.\n",
+    "\n",
+    "- ⚠️ **More setup required**\n",
+    "- ⚠️ **Must manually register tools**"
    ]
   },
   {
@@ -94,6 +127,13 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "websurfer = WebSurferAgent(name=\"WebSurfer\", llm_config=llm_config, web_tool=\"browser_use\")\n",
+    "user_proxy = UserProxyAgent(name=\"user_proxy\", human_input_mode=\"NEVER\")\n",
+    "# WebSurferAgent has a list of tools which are registered for LLM\n",
+    "# We need to register the tools for execution with the UserProxyAgent\n",
+    "for tool in websurfer.tools:\n",
+    "    tool.register_for_execution(user_proxy)\n",
+    "\n",
     "user_proxy.initiate_chat(\n",
     "    recipient=websurfer,\n",
     "    message=\"Get info from https://docs.ag2.ai/docs/Home\",\n",
@@ -119,6 +159,7 @@
     "   \n",
     "   ```bash\n",
     "   playwright install\n",
+    "   playwright install-deps\n",
     "   ```\n",
     "\n",
     "3. For running the code in Jupyther, use `nest_asyncio` to allow nested event loops.\n",
@@ -153,9 +194,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Crawl4AI WebSurferAgent\n",
-    "\n",
-    "The ONLY difference is the `web_tool` parameter which needs to be set to `crawl4ai`\n"
+    "### Crawl4AI WebSurferAgent"
    ]
   },
   {
@@ -168,16 +207,48 @@
     "\n",
     "llm_config = {\n",
     "    \"config_list\": config_list,\n",
-    "}\n",
+    "}"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "There are two ways to start a chat session which is using only one agent with LLM configuration.\n",
     "\n",
-    "user_proxy = UserProxyAgent(name=\"user_proxy\", human_input_mode=\"NEVER\")\n",
+    "#### **Recommended:** Using the `run` Method\n",
+    "\n",
+    "The new `run` method simplifies the process by eliminating the need for manual `UserProxyAgent` creation.\n",
+    "\n",
+    "- ✅ **Easier setup** – No need to manually register tools\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# `web_tool` parameter must be set to `crawl4ai` in order for the `Crawl4AITool` to be used.\n",
     "websurfer = WebSurferAgent(name=\"WebSurfer\", llm_config=llm_config, web_tool=\"crawl4ai\")\n",
     "\n",
-    "websurfer_tools = websurfer.tools\n",
-    "# WebSurferAgent has a list of tools which are registered for LLM\n",
-    "# We need to register the tools for execution with the UserProxyAgent\n",
-    "for tool in websurfer_tools:\n",
-    "    tool.register_for_execution(user_proxy)"
+    "websurfer.run(\n",
+    "    message=\"Get info from https://docs.ag2.ai/docs/Home\",\n",
+    "    tools=websurfer.tools,\n",
+    "    max_turns=2,\n",
+    "    user_input=False,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "#### **Manual Setup:** Using `initiate_chat` Method\n",
+    "This method requires manually creating a `UserProxyAgent` and registering tools for execution.\n",
+    "\n",
+    "- ⚠️ **More setup required**\n",
+    "- ⚠️ **Must manually register tools**"
    ]
   },
   {
@@ -186,6 +257,14 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "user_proxy = UserProxyAgent(name=\"user_proxy\", human_input_mode=\"NEVER\")\n",
+    "websurfer = WebSurferAgent(name=\"WebSurfer\", llm_config=llm_config, web_tool=\"crawl4ai\")\n",
+    "\n",
+    "# WebSurferAgent has a list of tools which are registered for LLM\n",
+    "# We need to register the tools for execution with the UserProxyAgent\n",
+    "for tool in websurfer.tools:\n",
+    "    tool.register_for_execution(user_proxy)\n",
+    "\n",
     "user_proxy.initiate_chat(\n",
     "    recipient=websurfer,\n",
     "    message=\"Get info from https://docs.ag2.ai/docs/Home\",\n",

diff --git a/test/agentchat/test_function_call.py b/test/agentchat/test_function_call.py
@@ -252,7 +252,6 @@ def test_update_function(credentials_gpt_4o_mini: Credentials):
         summary_method="reflection_with_llm",
     )
     messages1 = assistant.chat_messages[user_proxy][-1]["content"]
-    print(messages1)
     print("Chat summary and cost", res1.summary, res1.cost)
 
     assistant.update_function_signature("greet_user", is_remove=True)
@@ -262,7 +261,6 @@ def test_update_function(credentials_gpt_4o_mini: Credentials):
         summary_method="reflection_with_llm",
     )
     messages2 = assistant.chat_messages[user_proxy][-1]["content"]
-    print(messages2)
     # The model should know about the function in the context of the conversation
     assert "greet_user" in messages1
     assert "greet_user" not in messages2

diff --git a/test/agentchat/test_tool_calls.py b/test/agentchat/test_tool_calls.py
@@ -167,7 +167,6 @@ def test_update_tool(credentials_gpt_4o: Credentials):
         message="What functions do you know about in the context of this conversation? End your response with 'TERMINATE'.",
     )
     messages1 = assistant.chat_messages[user_proxy][-1]["content"]
-    print("Message:", messages1)
     print("Summary:", res.summary)
     assert messages1.replace("TERMINATE", "") == res.summary, (
         "Message (removing TERMINATE) and summary should be the same"
@@ -180,7 +179,6 @@ def test_update_tool(credentials_gpt_4o: Credentials):
         summary_method="reflection_with_llm",
     )
     messages2 = assistant.chat_messages[user_proxy][-1]["content"]
-    print("Message2:", messages2)
     # The model should know about the function in the context of the conversation
     assert "greet_user" in messages1
     assert "greet_user" not in messages2

diff --git a/website/_blogs/2025-01-31-WebSurferAgent/index.mdx b/website/_blogs/2025-01-31-WebSurferAgent/index.mdx
@@ -51,6 +51,7 @@ To get started with the [`Browser Use`](https://github.com/browser-use/browser-u
 
    ```bash
    playwright install
+   playwright install-deps
    ```
 
 3. For running the code in Jupyter, use `nest_asyncio` to allow nested event loops.
@@ -74,26 +75,48 @@ nest_asyncio.apply()
 ```
 
 ### Configure WebSurferAgent with Browser Use
-[`WebSurferAgent`](/docs/api-reference/autogen/agents/WebSurferAgent) is the one responsible for browsing the web and retrieving information. The `web_tool="browser_use"` tells the agent to use the [`BrowserUseTool`](/docs/api-reference/autogen/tools/experimental/BrowserUseTool) to surf the web.
-
-After creating the [`WebSurferAgent`](/docs/api-reference/autogen/agents/WebSurferAgent), its tools (such as the [`BrowserUseTool`](/docs/api-reference/autogen/tools/experimental/BrowserUseTool)) are registered with the [`UserProxyAgent`](/docs/api-reference/autogen/UserProxyAgent) so they can be used during the interaction.
-
-The user can ask the [`WebSurferAgent`](/docs/api-reference/autogen/agents/WebSurferAgent) to fetch information from a specific webpage, in this case, the AG2 documentation home page.
 
 ```python
 config_list = [{"model": "gpt-4o-mini", "api_key": os.environ["OPENAI_API_KEY"]}]
 
 llm_config = {
     "config_list": config_list,
 }
+````
 
-user_proxy = UserProxyAgent(name="user_proxy", human_input_mode="NEVER")
+[`WebSurferAgent`](/docs/api-reference/autogen/agents/WebSurferAgent) is the one responsible for browsing the web and retrieving information. The `web_tool="browser_use"` tells the agent to use the [`BrowserUseTool`](/docs/api-reference/autogen/tools/experimental/BrowserUseTool) to surf the web.
+
+After creating the [`WebSurferAgent`](/docs/api-reference/autogen/agents/WebSurferAgent) there are two ways to start the chat session:
+
+#### **Recommended:** Using the [`run`](/docs/api-reference/autogen/ConversableAgent#run) Method
+
+The new [`run`](/docs/api-reference/autogen/ConversableAgent#run) method simplifies the process by eliminating the need for manual [`UserProxyAgent`](/docs/api-reference/autogen/UserProxyAgent) creation.
+
+- ✅ **Easier setup** – No need to manually register tools
+
+```python
 websurfer = WebSurferAgent(name="WebSurfer", llm_config=llm_config, web_tool="browser_use")
 
-websurfer_tools = websurfer.tools
+websurfer.run(
+    message="Get info from https://docs.ag2.ai/docs/Home",
+    tools=websurfer.tools,
+    max_turns=2,
+    user_input=False,
+)
+```
+
+#### **Manual Setup:** Using [`initiate_chat`](/docs/api-reference/autogen/ConversableAgent#initiate-chat) Method
+This method requires manually creating a [`UserProxyAgent`](/docs/api-reference/autogen/UserProxyAgent) and registering tools for execution.
+
+- ⚠️ **More setup required**
+- ⚠️ **Must manually register tools**
+
+```python
+websurfer = WebSurferAgent(name="WebSurfer", llm_config=llm_config, web_tool="browser_use")
+user_proxy = UserProxyAgent(name="user_proxy", human_input_mode="NEVER")
 # WebSurferAgent has a list of tools which are registered for LLM
 # We need to register the tools for execution with the UserProxyAgent
-for tool in websurfer_tools:
+for tool in websurfer.tools:
     tool.register_for_execution(user_proxy)
 
 user_proxy.initiate_chat(
@@ -103,6 +126,7 @@ user_proxy.initiate_chat(
 )
 ```
 
+### Output
 ```console
 user_proxy (to WebSurfer):
 
@@ -331,6 +355,7 @@ To integrate [`Crawl4AI`](https://github.com/unclecode/crawl4ai) with AG2, follo
 
    ```bash
    playwright install
+   playwright install-deps
    ```
 3. For running the code in Jupyter, use `nest_asyncio` to allow nested event loops.
     ```bash
@@ -353,15 +378,45 @@ nest_asyncio.apply()
 ```
 
 ### Configure WebSurferAgent with Crawl4AI
-The only difference from the previous example is that the `web_tool` parameter must be set to `crawl4ai` in order for the [`Crawl4AITool`](/docs/api-reference/autogen/tools/experimental/Crawl4AITool) to be used.
 
 ```python
 config_list = [{"model": "gpt-4o-mini", "api_key": os.environ["OPENAI_API_KEY"]}]
 
 llm_config = {
     "config_list": config_list,
 }
+```
+
+The only difference from the previous example is that the `web_tool` parameter must be set to `crawl4ai` in order for the [`Crawl4AITool`](/docs/api-reference/autogen/tools/experimental/Crawl4AITool) to be used.
+
 
+As in previous example, there are two ways to start the chat session:
+
+#### **Recommended:** Using the [`run`](/docs/api-reference/autogen/ConversableAgent#run) Method
+
+The new [`run`](/docs/api-reference/autogen/ConversableAgent#run) method simplifies the process by eliminating the need for manual [`UserProxyAgent`](/docs/api-reference/autogen/UserProxyAgent) creation.
+
+- ✅ **Easier setup** – No need to manually register tools
+
+```python
+# `web_tool` parameter must be set to `crawl4ai` in order for the `Crawl4AITool` to be used.
+websurfer = WebSurferAgent(name="WebSurfer", llm_config=llm_config, web_tool="crawl4ai")
+
+websurfer.run(
+    message="Get info from https://docs.ag2.ai/docs/Home",
+    tools=websurfer.tools,
+    max_turns=2,
+    user_input=False,
+)
+```
+
+#### **Manual Setup:** Using [`initiate_chat`](/docs/api-reference/autogen/ConversableAgent#initiate-chat) Method
+This method requires manually creating a [`UserProxyAgent`](/docs/api-reference/autogen/UserProxyAgent) and registering tools for execution.
+
+- ⚠️ **More setup required**
+- ⚠️ **Must manually register tools**
+
+```python
 user_proxy = UserProxyAgent(name="user_proxy", human_input_mode="NEVER")
 websurfer = WebSurferAgent(name="WebSurfer", llm_config=llm_config, web_tool="crawl4ai")
 
@@ -378,6 +433,7 @@ user_proxy.initiate_chat(
 )
 ```
 
+### Output
 ```console
 user_proxy (to WebSurfer):