Skip to content

armgit / camel

Public Imported from GitHub
0

🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org

feat: add metaso.cn API search (#2986)

@hu-mengliangce56cf0e4e6f
diff --git a/.env b/.env
index 3238f48d4..1e2b5de6d 100644
--- a/.env
+++ b/.env
@@ -124,6 +124,9 @@
#LANGFUSE_SECRET_KEY="Fill your API key here"
#LANGFUSE_PUBLIC_KEY="Fill your API key here"
+#METASO API(https://metaso.cn/search-api/api-keys)
+#METASO_API_KEY="Fill your API key here"
+
# E2B
# E2B_API_KEY="Fill your e2b or e2b-compatible sandbox provider API Key here"
# E2B_DOMAIN="Fill your custom e2b domain here"
diff --git a/camel/toolkits/search_toolkit.py b/camel/toolkits/search_toolkit.py
index 4e3c1585d..587ffd003 100644
--- a/camel/toolkits/search_toolkit.py
+++ b/camel/toolkits/search_toolkit.py
@@ -1317,6 +1317,67 @@ class SearchToolkit(BaseToolkit):
f"search: {e!s}"
}
+ @api_keys_required([(None, 'METASO_API_KEY')])
+ def search_metaso(
+ self,
+ query: str,
+ page: int = 1,
+ include_summary: bool = False,
+ include_raw_content: bool = False,
+ concise_snippet: bool = False,
+ scope: Literal[
+ "webpage", "document", "scholar", "image", "video", "podcast"
+ ] = "webpage",
+ ) -> dict[str, Any]:
+ r"""Perform a web search using the metaso.cn API.
+
+ Args:
+ query (str): The search query string.
+ page (int): Page number, default is 1.
+ include_summary (bool): Whether to include summary in the result.
+ include_raw_content (bool): Whether to include raw content in the result.
+ concise_snippet (bool): Whether to return concise snippet.
+ scope (Literal[
+ "webpage", "document", "scholar", "image", "video", "podcast"
+ ]): Search scope, default is 'webpage'.
+
+ Returns:
+ dict: Search results or error information.
+ """
+ import http.client
+ import json
+
+ # It is recommended to put the token in environment variable for security
+ import os
+ METASO_API_KEY = os.getenv("METASO_API_KEY")
+
+ conn = http.client.HTTPSConnection("metaso.cn")
+ payload = json.dumps({
+ "q": query,
+ "scope": scope,
+ "includeSummary": include_summary,
+ "page": str(page),
+ "includeRawContent": include_raw_content,
+ "conciseSnippet": concise_snippet
+ })
+ headers = {
+ 'Authorization': f'Bearer {METASO_API_KEY}',
+ 'Accept': 'application/json',
+ 'Content-Type': 'application/json'
+ }
+ try:
+ conn.request("POST", "/api/v1/search", payload, headers)
+ res = conn.getresponse()
+ data = res.read()
+ result = data.decode("utf-8")
+ try:
+ return json.loads(result)
+ except Exception:
+ return {"error": f"Metaso returned content could not be parsed: {result}"}
+ except Exception as e:
+ return {"error": f"Metaso search failed: {e}"}
+
+
def get_tools(self) -> List[FunctionTool]:
r"""Returns a list of FunctionTool objects representing the
functions in the toolkit.
@@ -1337,4 +1398,5 @@ class SearchToolkit(BaseToolkit):
FunctionTool(self.search_bing),
FunctionTool(self.search_exa),
FunctionTool(self.search_alibaba_tongxiao),
+ FunctionTool(self.search_metaso),
]
diff --git a/examples/toolkits/search_toolkit.py b/examples/toolkits/search_toolkit.py
index 692d00ceb..b59a370e7 100644
--- a/examples/toolkits/search_toolkit.py
+++ b/examples/toolkits/search_toolkit.py
@@ -378,3 +378,42 @@ print(search_alibaba_tongxiao_response)
]}
===============================================================================
"""
+# Example using metaso response
+metaso_response = SearchToolkit().search_metaso(
+ query="What is CAMEL-AI?",
+ scope="webpage",
+ page= 1,
+)
+
+print(metaso_response)
+"""
+===============================================================================
+{'credits': 10, 'searchParameters': {'q': 'What is CAMEL-AI?', 'scope': 'webpage',
+ 'page': 1, 'includeSummary': False, 'includeRawContent': False, 'conciseSnippet':
+ False, 'format': 'chat_completions'}, 'webpages': [{'title': '使用CAMEL框架和Qwen
+ 模型自动进行数据获取及报告与知识 ...', 'link': 'https://developer.aliyun.com/articl
+ e/1647631', 'score': 'high', 'snippet': '**关于CAMEL AI**\n* * *\n**"Finding the
+ Scaling Law of Agents"**\nCAMEL AI (https://www.camel-ai.org) 是一个致力于构建基
+ 于大语言模型(LLMs)的多智能体系统的开源平台,旨在推动人工智能的发展。就像AI领域的“沙
+ 漠之舟”,CAMEL不仅在信息的广袤领域里稳步前行,更为实现AGI探索出一条前沿之路。CAMEL框架
+ 是最早基于ChatGPT的autonomous agents的知名项目,已被NeurlPS 2023录用。', 'position'
+ : 1, 'authors': ['阿里云开发者社区'], 'date': '2024年12月31日'}, {'title':
+ 'CAMEL-AI团队参与发表Nature子刊啦!聚焦LLM如何重塑 ...', 'link': 'https://cloud.
+ tencent.com/developer/article/2475624', 'score': 'high', 'snippet': 'Finding
+ the Scaling Law of Agents"\nCAMEL AI (https://www.camel-ai.org) 是一个致力于构建
+ 基于大语言模型)LLMs)的多智能体系统的开源平台,旨在推动人工智能的发展。\n就像AI领域的
+ “沙漠之舟”,CAMEL不仅在信息的广袤领域里稳步前行,更为实现AGI探索出一条前沿之路。\n
+ CAMEL框架是最早基于ChatGPT的autonomous agents的知名项目,已被NeurlPS 2023录用。\n
+ CAMEL AI框架设计上极具灵活性与模块化,包含多种智能体实现、专业领域提示示例以及AI数据探
+ 索功能等模块,适应各种高效决策与复杂数据处理需求。\n通过协作运作的智能体队列,CAMEL AI
+ 不仅在金融、零售、医疗等领域展示出强大的适用性和创新性,还大大提升了自适应能力......
+ 'position': 9, 'authors': ['KubeNine'], 'date': '2025年04月15日'}, {'title': '
+ RAG应用的评估-CSDN博客', 'link': 'https://blog.csdn.net/weixin_44245188/article/
+ details/148157069', 'score': 'high', 'snippet': "Query: 什么是CAMEL AI?\n
+ Expected Answers: ['CAMEL AI 是一个开源的、社区驱动的AI框架。']\nRetrieved
+ Results: ['CAMEL AI 介绍\\n\\nCAMEL AI 是一个开源的、社区驱动的AI框架,旨在简化AI应
+ 用的开发和部署。该框架提供了多种功能模块,包括数据加载、特征工程、模型训练和部署等。']
+ \nPrecision: 1.0000\nRecall: 1.0000\nF1 Score: 1.0000\nAverage
+ Similarity: 0.9584", 'position': 10, 'date': '2025年05月23日'}], 'total': 38}
+===============================================================================
+"""
\ No newline at end of file
diff --git a/test/toolkits/test_search_functions.py b/test/toolkits/test_search_functions.py
index 861c13d63..5ce215511 100644
--- a/test/toolkits/test_search_functions.py
+++ b/test/toolkits/test_search_functions.py
@@ -833,3 +833,118 @@ def test_search_alibaba_tongxiao(mock_get, search_toolkit):
}
],
}
+
+@patch('http.client.HTTPSConnection')
+def test_search_metaso_success(mock_https_connection, search_toolkit):
+ # Mock the connection and response
+ mock_conn = MagicMock()
+ mock_https_connection.return_value = mock_conn
+
+ mock_response = MagicMock()
+ mock_response.read.return_value = b'{"results": [{"title": "Test Title", "url": "https://example.com", "snippet": "Test snippet"}], "total": 1}'
+ mock_conn.getresponse.return_value = mock_response
+
+ # Mock environment variable
+ with patch.dict(os.environ, {'METASO_API_KEY': 'test_api_key'}):
+ result = search_toolkit.search_metaso(query="test query")
+
+ # Verify the result
+ expected_result = {
+ "results": [
+ {
+ "title": "Test Title",
+ "url": "https://example.com",
+ "snippet": "Test snippet"
+ }
+ ],
+ "total": 1
+ }
+ assert result == expected_result
+
+ # Verify the API call
+ mock_https_connection.assert_called_once_with("metaso.cn")
+ mock_conn.request.assert_called_once()
+
+ # Verify request parameters
+ call_args = mock_conn.request.call_args
+ assert call_args[0][0] == "POST"
+ assert call_args[0][1] == "/api/v1/search"
+
+ # Verify headers
+ headers = call_args[0][3]
+ assert headers['Authorization'] == 'Bearer test_api_key'
+ assert headers['Accept'] == 'application/json'
+ assert headers['Content-Type'] == 'application/json'
+
+
+@patch('http.client.HTTPSConnection')
+def test_search_metaso_with_parameters(mock_https_connection, search_toolkit):
+ # Mock the connection and response
+ mock_conn = MagicMock()
+ mock_https_connection.return_value = mock_conn
+
+ mock_response = MagicMock()
+ mock_response.read.return_value = b'{"results": [], "total": 0}'
+ mock_conn.getresponse.return_value = mock_response
+
+ with patch.dict(os.environ, {'METASO_API_KEY': 'test_api_key'}):
+ search_toolkit.search_metaso(
+ query="test query",
+ page=2,
+ include_summary=True,
+ include_raw_content=True,
+ concise_snippet=True,
+ scope="document"
+ )
+
+ # Verify the payload contains correct parameters
+ call_args = mock_conn.request.call_args
+ payload = call_args[0][2]
+ import json
+ payload_data = json.loads(payload)
+
+ assert payload_data["q"] == "test query"
+ assert payload_data["page"] == "2"
+ assert payload_data["includeSummary"] is True
+ assert payload_data["includeRawContent"] is True
+ assert payload_data["conciseSnippet"] is True
+ assert payload_data["scope"] == "document"
+
+
+@patch('http.client.HTTPSConnection')
+def test_search_metaso_connection_error(mock_https_connection, search_toolkit):
+ # Mock connection error
+ mock_conn = MagicMock()
+ mock_https_connection.return_value = mock_conn
+ mock_conn.request.side_effect = Exception("Connection failed")
+
+ with patch.dict(os.environ, {'METASO_API_KEY': 'test_api_key'}):
+ result = search_toolkit.search_metaso(query="test query")
+
+ assert "error" in result
+ assert "Metaso search failed: Connection failed" in result["error"]
+
+ # Verify connection was attempted
+ mock_https_connection.assert_called_once_with("metaso.cn")
+ mock_conn.request.assert_called_once()
+
+
+@patch('http.client.HTTPSConnection')
+def test_search_metaso_invalid_json(mock_https_connection, search_toolkit):
+ # Mock invalid JSON response
+ mock_conn = MagicMock()
+ mock_https_connection.return_value = mock_conn
+
+ mock_response = MagicMock()
+ mock_response.read.return_value = b'invalid json response'
+ mock_conn.getresponse.return_value = mock_response
+
+ with patch.dict(os.environ, {'METASO_API_KEY': 'test_api_key'}):
+ result = search_toolkit.search_metaso(query="test query")
+
+ assert "error" in result
+ assert "Metaso returned content could not be parsed" in result["error"]
+
+ # Verify connection was attempted
+ mock_https_connection.assert_called_once_with("metaso.cn")
+ mock_conn.request.assert_called_once()