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

fead: Add pubmed in toolkits (#1806)

Rishabh3d7d6924ed04
diff --git a/camel/toolkits/__init__.py b/camel/toolkits/__init__.py
index b7f5ee859..c892954d5 100644
--- a/camel/toolkits/__init__.py
+++ b/camel/toolkits/__init__.py
@@ -58,6 +58,7 @@ from .mcp_toolkit import MCPToolkit
from .web_toolkit import WebToolkit
from .file_write_toolkit import FileWriteToolkit
from .terminal_toolkit import TerminalToolkit
+from .pubmed_toolkit import PubMedToolkit
__all__ = [
@@ -104,4 +105,5 @@ __all__ = [
'WebToolkit',
'FileWriteToolkit',
'TerminalToolkit',
+ 'PubMedToolkit',
]
diff --git a/camel/toolkits/pubmed_toolkit.py b/camel/toolkits/pubmed_toolkit.py
new file mode 100644
index 000000000..e3bd7d3be
--- /dev/null
+++ b/camel/toolkits/pubmed_toolkit.py
@@ -0,0 +1,346 @@
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+
+from typing import Any, Dict, List, Optional, Union, cast
+
+import requests
+
+from camel.logger import get_logger
+from camel.toolkits import BaseToolkit, FunctionTool
+
+logger = get_logger(__name__)
+
+
+class PubMedToolkit(BaseToolkit):
+ r"""A toolkit for interacting with PubMed's E-utilities API to access
+ MEDLINE data.
+
+ This toolkit provides functionality to search and retrieve papers from the
+ PubMed database, including abstracts, citations, and other metadata.
+
+ Args:
+ timeout (Optional[float]): The timeout for API requests in seconds.
+ (default: :obj:`None`)
+ """
+
+ BASE_URL = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"
+
+ def __init__(self, timeout: Optional[float] = None) -> None:
+ r"""Initializes the PubMedToolkit."""
+ super().__init__(timeout=timeout)
+
+ def _make_request(
+ self,
+ endpoint: str,
+ params: Dict[str, Union[str, int]],
+ retries: int = 3,
+ ) -> Optional[Dict[str, Any]]:
+ r"""Makes a request to the PubMed/MEDLINE API with error handling and
+ retries.
+
+ Args:
+ endpoint (str): The API endpoint to call.
+ params (Dict[str, Union[str, int]]): Query parameters.
+ retries (int, optional): Number of retry attempts.
+ (default: :obj:`3`)
+
+ Returns:
+ Optional[Dict[str, Any]]: JSON response if successful, else None.
+ """
+ url = f"{self.BASE_URL}/{endpoint}"
+ request_params = cast(Dict[str, Union[str, int]], params)
+
+ for attempt in range(retries):
+ try:
+ response = requests.get(
+ url, params=request_params, timeout=self.timeout
+ )
+ response.raise_for_status()
+
+ if not response.text:
+ logger.warning(
+ f"Empty response from PubMed API: {endpoint}"
+ )
+ return None
+
+ return response.json()
+ except requests.RequestException as e:
+ if attempt == retries - 1:
+ logger.error(f"Failed to fetch data from PubMed: {e!s}")
+ return None
+ logger.warning(f"Request attempt {attempt + 1} failed: {e!s}")
+ except ValueError as e:
+ logger.error(f"Failed to parse JSON response: {e!s}")
+ return None
+ return None
+
+ def search_papers(
+ self,
+ query: str,
+ max_results: int = 10,
+ sort: str = "relevance",
+ date_range: Optional[Dict[str, str]] = None,
+ publication_type: Optional[List[str]] = None,
+ ) -> List[Dict[str, str]]:
+ r"""Search for biomedical papers in MEDLINE via PubMed with advanced
+ filtering options.
+
+ Args:
+ query (str): The search query string.
+ max_results (int, optional): Maximum number of results to return.
+ (default: :obj:`10`)
+ sort (str, optional): Sort order - 'relevance' or 'date'.
+ (default: :obj:`"relevance"`)
+ date_range (Optional[Dict[str, str]], optional): Date range filter
+ with 'from' and 'to' dates in YYYY/MM/DD format.
+ (default: :obj:`None`)
+ publication_type (Optional[List[str]], optional): Filter by
+ publication types (e.g., ["Journal Article", "Review"]).
+ (default: :obj:`None`)
+
+ Returns:
+ List[Dict[str, str]]: List of papers with their metadata.
+ """
+ # Build query with filters
+ filtered_query = query
+ if publication_type:
+ type_filter = " OR ".join(
+ [f'"{pt}"[Publication Type]' for pt in publication_type]
+ )
+ filtered_query = f"({query}) AND ({type_filter})"
+ if date_range:
+ date_filter = (
+ f"{date_range.get('from', '')}:"
+ f"{date_range.get('to', '')}[Date - Publication]"
+ )
+ filtered_query = f"({filtered_query}) AND ({date_filter})"
+
+ # Search for paper IDs
+ search_params: Dict[str, Union[str, int]] = {
+ "db": "pubmed",
+ "term": filtered_query,
+ "retmax": max_results,
+ "sort": "relevance" if sort == "relevance" else "pub+date",
+ "retmode": "json",
+ }
+
+ search_data = self._make_request("esearch.fcgi", search_params)
+ if not search_data or "esearchresult" not in search_data:
+ logger.error("Failed to retrieve search results")
+ return []
+
+ paper_ids = search_data["esearchresult"].get("idlist", [])
+ if not paper_ids:
+ return []
+
+ # Fetch details for papers
+ results = []
+ for paper_id in paper_ids:
+ paper_details = self.get_paper_details(paper_id)
+ if paper_details:
+ results.append(paper_details)
+
+ return results
+
+ def get_paper_details(
+ self,
+ paper_id: Union[str, int],
+ include_references: bool = False,
+ ) -> Optional[Dict[str, Any]]:
+ r"""Get detailed information about a specific biomedical paper from
+ MEDLINE/PubMed.
+
+ Args:
+ paper_id (Union[str, int]): PubMed ID of the paper.
+ include_references (bool, optional): Whether to include referenced
+ papers. (default: :obj:`False`)
+
+ Returns:
+ Optional[Dict[str, Any]]: Paper details including title, authors,
+ abstract, etc., or None if retrieval fails.
+ """
+ # Fetch summary
+ summary_params: Dict[str, Union[str, int]] = {
+ "db": "pubmed",
+ "id": str(paper_id),
+ "retmode": "json",
+ }
+ summary_data = self._make_request("esummary.fcgi", summary_params)
+
+ if not summary_data or "result" not in summary_data:
+ logger.error(
+ f"Failed to retrieve paper details for ID: {paper_id}"
+ )
+ return None
+
+ paper_data = summary_data["result"][str(paper_id)]
+
+ # Handle authors - they come as a list of dicts with 'name' key
+ authors = paper_data.get("authors", [])
+ author_names = []
+ for author in authors:
+ if isinstance(author, dict) and "name" in author:
+ author_names.append(author["name"])
+ elif isinstance(author, str):
+ author_names.append(author)
+
+ # Get abstract
+ abstract = self.get_abstract(paper_id)
+
+ # Get references if requested
+ references = []
+ if include_references:
+ ref_params: Dict[str, Union[str, int]] = {
+ "db": "pubmed",
+ "id": str(paper_id),
+ "linkname": "pubmed_pubmed_refs",
+ "retmode": "json",
+ }
+ ref_data = self._make_request("elink.fcgi", ref_params)
+ if ref_data and "linksets" in ref_data:
+ try:
+ references = ref_data["linksets"][0]["linksetdbs"][0][
+ "links"
+ ]
+ except (KeyError, IndexError):
+ logger.warning(
+ f"No references found for paper ID: {paper_id}"
+ )
+
+ return cast(
+ Dict[str, Any],
+ {
+ "id": str(paper_id),
+ "title": paper_data.get("title", ""),
+ "authors": ", ".join(author_names),
+ "journal": paper_data.get("source", ""),
+ "pub_date": paper_data.get("pubdate", ""),
+ "abstract": abstract,
+ "doi": paper_data.get("elocationid", ""),
+ "keywords": paper_data.get("keywords", []),
+ "mesh_terms": paper_data.get("mesh", []),
+ "publication_types": paper_data.get("pubtype", []),
+ "references": references if include_references else None,
+ },
+ )
+
+ def get_abstract(self, paper_id: Union[str, int]) -> str:
+ r"""Get the abstract of a specific biomedical paper from MEDLINE/
+ PubMed.
+
+ Args:
+ paper_id (Union[str, int]): PubMed ID of the paper.
+
+ Returns:
+ str: The abstract text.
+ """
+ params: Dict[str, Union[str, int]] = {
+ "db": "pubmed",
+ "id": str(paper_id),
+ "rettype": "abstract",
+ "retmode": "text",
+ }
+
+ try:
+ response = requests.get(
+ f"{self.BASE_URL}/efetch.fcgi", params=params
+ )
+ response.raise_for_status()
+ return response.text.strip()
+ except requests.exceptions.RequestException as e:
+ logger.error(
+ f"Failed to retrieve abstract for ID {paper_id}: {e!s}"
+ )
+ return ""
+
+ def get_citation_count(self, paper_id: Union[str, int]) -> int:
+ r"""Get the number of citations for a biomedical paper in MEDLINE/
+ PubMed.
+
+ Args:
+ paper_id (Union[str, int]): PubMed ID of the paper.
+
+ Returns:
+ int: Number of citations, or 0 if retrieval fails.
+ """
+ params: Dict[str, Union[str, int]] = {
+ "db": "pubmed",
+ "id": str(paper_id),
+ "linkname": "pubmed_pubmed_citedin",
+ "retmode": "json",
+ }
+
+ data = self._make_request("elink.fcgi", params)
+ if not data or "linksets" not in data:
+ return 0
+
+ try:
+ return len(data["linksets"][0]["linksetdbs"][0]["links"])
+ except (KeyError, IndexError):
+ return 0
+
+ def get_related_papers(
+ self,
+ paper_id: Union[str, int],
+ max_results: int = 10,
+ ) -> List[Dict[str, Any]]:
+ r"""Get biomedical papers related to a specific paper in MEDLINE/
+ PubMed.
+
+ Args:
+ paper_id (Union[str, int]): PubMed ID of the paper.
+ max_results (int, optional): Maximum number of results to return.
+ (default: :obj:`10`)
+
+ Returns:
+ List[Dict[str, Any]]: List of related papers with their metadata.
+ """
+ params: Dict[str, Union[str, int]] = {
+ "db": "pubmed",
+ "id": str(paper_id),
+ "linkname": "pubmed_pubmed",
+ "retmode": "json",
+ }
+
+ data = self._make_request("elink.fcgi", params)
+ if not data or "linksets" not in data:
+ return []
+
+ try:
+ related_ids = data["linksets"][0]["linksetdbs"][0]["links"][
+ :max_results
+ ]
+ related_papers: List[Dict[str, Any]] = []
+
+ for pid in related_ids:
+ if paper := self.get_paper_details(pid):
+ related_papers.append(paper)
+
+ return related_papers
+ except (KeyError, IndexError):
+ return []
+
+ def get_tools(self) -> List[FunctionTool]:
+ r"""Returns a list of tools provided by the PubMed toolkit.
+
+ Returns:
+ List[FunctionTool]: List of available tools.
+ """
+ return [
+ FunctionTool(self.search_papers),
+ FunctionTool(self.get_paper_details),
+ FunctionTool(self.get_abstract),
+ FunctionTool(self.get_citation_count),
+ FunctionTool(self.get_related_papers),
+ ]
diff --git a/examples/toolkits/pubmed_toolkit.py b/examples/toolkits/pubmed_toolkit.py
new file mode 100644
index 000000000..ff731fd3c
--- /dev/null
+++ b/examples/toolkits/pubmed_toolkit.py
@@ -0,0 +1,283 @@
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+
+from camel.agents import ChatAgent
+from camel.configs.openai_config import ChatGPTConfig
+from camel.models import ModelFactory
+from camel.toolkits import PubMedToolkit
+from camel.types import ModelPlatformType, ModelType
+
+# Initialize PubMed toolkit and get tools
+tools = PubMedToolkit().get_tools()
+
+# Set up model configuration
+model_config_dict = ChatGPTConfig(
+ temperature=0.0,
+).as_dict()
+
+model = ModelFactory.create(
+ model_platform=ModelPlatformType.DEFAULT,
+ model_type=ModelType.DEFAULT,
+ model_config_dict=model_config_dict,
+)
+
+# Create chat agent
+system_msg = (
+ "You are a research assistant specialized in medical literature. "
+ "Help researchers find and analyze scientific papers from PubMed."
+)
+camel_agent = ChatAgent(
+ system_message=system_msg,
+ model=model,
+ tools=tools,
+)
+camel_agent.reset()
+
+# Example 1: Search for recent papers about mRNA vaccine technology
+print("\nExample 1: Search for recent papers about mRNA vaccine technology")
+print("=" * 80)
+
+usr_msg = (
+ "Find recent review papers about mRNA vaccine technology published "
+ "in 2024, with a focus on therapeutic applications and clinical trials. "
+ "Limit to 3 papers."
+)
+
+response = camel_agent.step(usr_msg)
+print(str(response.info['tool_calls'])[:2000])
+
+"""
+===============================================================================
+ToolCallingRecord(
+ tool_name='search_papers',
+ args={
+ 'query': 'mRNA vaccine tech therapeutic applications trials',
+ 'max_results': 10,
+ 'sort': 'date',
+ 'date_range': {'from': '2024/01/01', 'to': '2024/12/31'},
+ 'publication_type': ['Review'],
+ },
+ result=[
+ {
+ 'id': '39601789',
+ 'title': 'Example Title',
+ 'authors': 'First Author, Second Author',
+ 'journal': 'Example Journal',
+ 'pub_date': '2025 Jan 6',
+ 'abstract': 'Abstract of the paper',
+===============================================================================
+"""
+
+
+# Example 2: Get detailed information about a specific paper
+print("\nExample 2: Get detailed paper information")
+print("=" * 80)
+
+usr_msg = (
+ "Get detailed information about PubMed ID 39601789 "
+ "(a key paper about mRNA vaccine technology)."
+)
+camel_agent.reset()
+response = camel_agent.step(usr_msg)
+print(str(response.info['tool_calls'])[:2000])
+
+"""
+===============================================================================
+[ToolCallingRecord(
+ tool_name='get_paper_details',
+ args={'paper_id': 37840631, 'include_references': True},
+ result={
+ 'id': '37840631',
+ 'title': 'Chinese guideline for lipid management (2023):
+ a new guideline rich in domestic elements for
+ controlling dyslipidemia.',
+ 'authors': 'Li JJ',
+ 'journal': 'J Geriatr Cardiol',
+ 'pub_date': '2023 Sep 28',
+ 'abstract': '1. J Geriatr Cardiol.
+ 2023 Sep 28;20(9):618-620.
+ doi: 10.26599/1671-5411.2023.09.007.
+ Chinese guideline for lipid management (2023):
+ a new guideline rich in domestic elements for
+ controlling dyslipidemia.Li JJ(1).\Author information:
+ (1)Division of Cardio-Metabolic Center,
+ State Key Laboratory of Cardiovascular
+ Disease, Fu Wai Hospital, National Center
+ for Cardiovascular Disease, Chinese Academy
+ of Medical Sciences, Peking Union Medical College,
+ Beijing, China.DOI: 10.26599/1671-5411.2023.09.007
+ PMCID: PMC10568543\nPMID: 37840631',
+ 'doi': 'doi: 10.26599/1671-5411.2023.09.007',
+ 'keywords': [],
+ 'mesh_terms': [],
+ 'publication_types': ['Journal Article'],
+ 'references': ['35729555', '34734202', '34404993',
+ '31172370', '30586774', '30526649',
+ '29434622', '20350253']
+ },
+ tool_call_id='call_k8s7oFcRvDBKuEKvk48uoWXZ'
+)]
+===============================================================================
+"""
+
+# Example 3: Find related papers and citation metrics
+print("\nExample 3: Find related papers and citation metrics")
+print("=" * 80)
+
+usr_msg = (
+ "Find papers related to PubMed ID 39601789 (limit to 3 papers) and "
+ "show its citation count."
+)
+camel_agent.reset()
+response = camel_agent.step(usr_msg)
+print(str(response.info['tool_calls'])[:2000])
+
+"""
+===============================================================================
+[ToolCallingRecord(
+ tool_name='get_related_papers',
+ args={'paper_id': 37840631, 'max_results': 5},
+ result=[
+ {'id': '37840631',
+ 'title': 'Chinese guideline for lipid management (2023):
+ a new guideline rich in domestic elements for
+ controlling dyslipidemia.',
+ 'authors': 'Li JJ',
+ 'journal': 'J Geriatr Cardiol',
+ 'pub_date': '2023 Sep 28',
+ 'abstract': (
+ '1. J Geriatr Cardiol. 2023 Sep 28;20(9):618-620. doi: '
+ '10.26599/1671-5411.2023.09.007.'
+ 'Chinese guideline for lipid management (2023): a new guideline'
+ 'rich in domestic elements for controlling dyslipidemia.'
+ 'Li JJ(1).Author information:(1)Division of Cardio-Metabolic '
+ 'Center, State Key Laboratory of Cardiovascular Disease, Fu Wai '
+ 'Hospital, National Center for Cardiovascular Disease, Chinese '
+ 'Academy of Medical Sciences, Peking Union Medical College, '
+ 'Beijing, China.DOI: 10.26599/1671-5411.2023.09.007'
+ 'PMCID: PMC10568543 PMID: 37840631'
+ ),
+ 'doi': 'doi: 10.26599/1671-5411.2023.09.007',
+ 'keywords': [],
+ 'mesh_terms': [],
+ 'publication_types': ['Journal Article'],
+ 'references': None},
+ {'id': '22801311',
+ 'title': (
+ '[Short-term impact of modified blood-lipid reports on physicians'
+ 'lipid lowering drug prescribing behavior and knowledge '
+ 'improvement on dyslipidemia].'
+ ),
+ 'authors': 'Li JH, Jiang H, Sun XH, Li CC, Ke YN, Yan SK, Wu YF',
+ 'journal': 'Zhonghua Xin Xue Guan Bing Za Zhi',
+ 'pub_date': '2012 Apr',
+ 'abstract': (
+ '1. Zhonghua Xin Xue Guan Bing Za Zhi. 2012 Apr;40(4):318-22.'
+ '[Short-term impact modified blood-lipid reports on physicians'
+ 'lipid lowering drug prescribing behavior and knowledge '
+ 'improvement on dyslipidemia].Article in Chinese]'
+ 'Li JH(1), Jiang H, Sun XH, Li CC, Ke YN, Yan SK, Wu YF.'
+ 'Author information:(1)Department of Cardiology, China-Japan'
+===============================================================================
+"""
+
+# Example 4: Advanced search with multiple filters
+print("\nExample 4: Advanced search with multiple filters")
+print("=" * 80)
+
+usr_msg = (
+ "Find clinical trial papers about mRNA-based cancer vaccines published "
+ "between 2023/01/01 and 2024/03/01, focusing on phase III trials. "
+ "Limit to 3 papers."
+)
+camel_agent.reset()
+response = camel_agent.step(usr_msg)
+print(str(response.info['tool_calls'])[:2000])
+
+"""
+===============================================================================
+[ToolCallingRecord(
+ tool_name='search_papers',
+ args={
+ 'query': 'mRNA cancer vaccine phase III clinical trial',
+ 'max_results': 10,
+ 'sort': 'date',
+ 'date_range': {'from': '2023/01/01', 'to': '2024/03/01'},
+ 'publication_type': ['Clinical Trial']
+ },
+ result=[
+ {
+ 'id': '37820782',
+ 'title': 'Stochastic interventional approach to assessing immune '
+ 'correlates of protection: Application to the COVE '
+ 'RNA-1273 vaccine trial.',
+ 'authors': (
+ 'Hejazi NS, Shen X, Carpp LN, Benkeser D, Follmann D,
+ Janes HE, Baden LR, El Sahly HM, Deng W, Zhou H,
+ Leav B, Montefiori DC, 'Gilbert PB'
+ ),
+ 'journal': 'Int J Infect Dis',
+ 'pub_date': '2023 Dec',
+ 'abstract': Abstract of the paper
+===============================================================================
+"""
+
+# Example 5: Get abstract and analyze citations
+print("\nExample 5: Get abstract and analyze citations")
+print("=" * 80)
+
+usr_msg = (
+ "Get the abstract of PubMed ID 39601789 and find out how many times "
+ "it has been cited."
+)
+camel_agent.reset()
+response = camel_agent.step(usr_msg)
+print(str(response.info['tool_calls'])[:2000])
+
+"""
+===============================================================================
+[
+ ToolCallingRecord(
+ tool_name='get_abstract',
+ args={'paper_id': 37840631},
+ result='''
+ 1. J Geriatr Cardiol. 2023 Sep 28;20(9):618-620. doi:
+ 10.26599/1671-5411.2023.09.007.
+
+ Chinese guideline for lipid management (2023):a new guideline
+ rich in domestic elements for controlling dyslipidemia.
+
+ Li JJ(1).
+
+ Author information:
+ (1)Division of Cardio-Metabolic Center, State Key Laboratory
+ of Cardiovascular Disease, Fu Wai Hospital, National Center
+ for Cardiovascular Disease, Chinese Academy of Medical Sciences,
+ Peking Union Medical College, Beijing, China.
+
+ DOI: 10.26599/1671-5411.2023.09.007
+ PMCID: PMC10568543
+ PMID: 37840631
+ ''',
+ tool_call_id='call_AFG6jLkdvWidaVGrj9UblTci'
+ ),
+ ToolCallingRecord(
+ tool_name='get_citation_count',
+ args={'paper_id': 37840631},
+ result=0,
+ tool_call_id='call_ZM3p59gtYmeR9DPdONNHV4Qw'
+ )
+]
+===============================================================================
+"""
diff --git a/test/toolkits/test_pubmed_toolkit.py b/test/toolkits/test_pubmed_toolkit.py
new file mode 100644
index 000000000..39e93e26c
--- /dev/null
+++ b/test/toolkits/test_pubmed_toolkit.py
@@ -0,0 +1,283 @@
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+
+from unittest.mock import MagicMock, patch
+
+import pytest
+import requests
+
+from camel.toolkits import PubMedToolkit
+
+
+def test_init():
+ toolkit = PubMedToolkit()
+ assert toolkit.BASE_URL == "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"
+ assert toolkit.timeout is None
+
+ toolkit_with_timeout = PubMedToolkit(timeout=30)
+ assert toolkit_with_timeout.timeout == 30
+
+
+def test_get_tools():
+ toolkit = PubMedToolkit()
+ tools = toolkit.get_tools()
+
+ assert isinstance(tools, list)
+ assert len(tools) == 5
+
+ # Extract function names from FunctionTool objects
+ tool_functions = {tool.func.__name__ for tool in tools}
+ expected_functions = {
+ "search_papers",
+ "get_paper_details",
+ "get_abstract",
+ "get_citation_count",
+ "get_related_papers",
+ }
+ assert tool_functions == expected_functions
+
+
+@patch('requests.get')
+def test_search_papers_success(mock_get):
+ # Mock the search response
+ mock_search_response = MagicMock()
+ mock_search_response.json.return_value = {
+ "esearchresult": {"idlist": ["12345", "67890"]}
+ }
+ mock_search_response.text = (
+ '{"esearchresult": {"idlist": ["12345", "67890"]}}'
+ )
+
+ # Mock the summary responses for each paper ID
+ mock_summary_response1 = MagicMock()
+ mock_summary_response1.json.return_value = {
+ "result": {
+ "12345": {
+ "title": "Test Paper 1",
+ "authors": [{"name": "Author 1"}, {"name": "Author 2"}],
+ "source": "Test Journal",
+ "pubdate": "2024",
+ }
+ }
+ }
+ mock_summary_response1.text = (
+ '{"result": {"12345": {"title": "Test Paper 1"}}}'
+ )
+
+ mock_summary_response2 = MagicMock()
+ mock_summary_response2.json.return_value = {
+ "result": {
+ "67890": {
+ "title": "Test Paper 2",
+ "authors": [{"name": "Author 3"}],
+ "source": "Another Journal",
+ "pubdate": "2023",
+ }
+ }
+ }
+ mock_summary_response2.text = (
+ '{"result": {"67890": {"title": "Test Paper 2"}}}'
+ )
+
+ # Mock the abstract responses
+ mock_abstract_response1 = MagicMock()
+ mock_abstract_response1.text = "Test abstract 1"
+
+ mock_abstract_response2 = MagicMock()
+ mock_abstract_response2.text = "Test abstract 2"
+
+ # Set up the side effect sequence
+ mock_get.side_effect = [
+ mock_search_response, # Search for papers
+ mock_summary_response1, # Get details for paper 12345
+ mock_abstract_response1, # Get abstract for paper 12345
+ mock_summary_response2, # Get details for paper 67890
+ mock_abstract_response2, # Get abstract for paper 67890
+ ]
+
+ toolkit = PubMedToolkit()
+ results = toolkit.search_papers(
+ query="test query",
+ max_results=2,
+ date_range={"from": "2023/01/01", "to": "2024/12/31"},
+ publication_type=["Journal Article"],
+ )
+
+ assert len(results) == 2
+ assert results[0]["title"] == "Test Paper 1"
+ assert results[1]["title"] == "Test Paper 2"
+
+ # Verify the search query construction
+ first_call_args = mock_get.call_args_list[0][1]['params']
+ assert "test query" in first_call_args['term']
+ assert '"Journal Article"[Publication Type]' in first_call_args['term']
+ assert (
+ "2023/01/01:2024/12/31[Date - Publication]" in first_call_args['term']
+ )
+
+
+@patch('requests.get')
+def test_search_papers_error(mock_get):
+ mock_get.side_effect = requests.RequestException("API Error")
+
+ toolkit = PubMedToolkit()
+ results = toolkit.search_papers("test query")
+
+ assert results == []
+
+
+@patch('requests.get')
+def test_get_paper_details_success(mock_get):
+ # Mock the summary response
+ mock_summary_response = MagicMock()
+ mock_summary_response.json.return_value = {
+ "result": {
+ "12345": {
+ "title": "Test Paper",
+ "authors": [{"name": "Author 1"}, {"name": "Author 2"}],
+ "source": "Test Journal",
+ "pubdate": "2024",
+ "mesh": ["Term 1", "Term 2"],
+ "keywords": ["keyword1", "keyword2"],
+ "pubtype": ["Journal Article"],
+ }
+ }
+ }
+ mock_summary_response.text = (
+ '{"result": {"12345": {"title": "Test Paper"}}}'
+ )
+
+ # Mock the abstract response
+ mock_abstract_response = MagicMock()
+ mock_abstract_response.text = "Test abstract"
+
+ mock_get.side_effect = [mock_summary_response, mock_abstract_response]
+
+ toolkit = PubMedToolkit()
+ result = toolkit.get_paper_details("12345")
+
+ assert result["title"] == "Test Paper"
+ assert result["authors"] == "Author 1, Author 2"
+ assert result["abstract"] == "Test abstract"
+ assert result["mesh_terms"] == ["Term 1", "Term 2"]
+ assert result["keywords"] == ["keyword1", "keyword2"]
+ assert result["publication_types"] == ["Journal Article"]
+
+
+@patch('requests.get')
+def test_get_paper_details_with_references(mock_get):
+ # Mock responses
+ mock_summary_response = MagicMock()
+ mock_summary_response.json.return_value = {
+ "result": {
+ "12345": {
+ "title": "Test Paper",
+ "authors": [{"name": "Author 1"}],
+ "source": "Test Journal",
+ "pubdate": "2024",
+ }
+ }
+ }
+ mock_summary_response.text = (
+ '{"result": {"12345": {"title": "Test Paper"}}}'
+ )
+
+ mock_abstract_response = MagicMock()
+ mock_abstract_response.text = "Test abstract"
+
+ mock_ref_response = MagicMock()
+ mock_ref_response.json.return_value = {
+ "linksets": [{"linksetdbs": [{"links": ["67890", "11111"]}]}]
+ }
+ mock_ref_response.text = (
+ '{"linksets": [{"linksetdbs": [{"links": ["67890", "11111"]}]}]}'
+ )
+
+ mock_get.side_effect = [
+ mock_summary_response,
+ mock_abstract_response,
+ mock_ref_response,
+ ]
+
+ toolkit = PubMedToolkit()
+ result = toolkit.get_paper_details("12345", include_references=True)
+
+ assert result["references"] == ["67890", "11111"]
+
+
+@patch('requests.get')
+def test_get_citation_count(mock_get):
+ mock_response = MagicMock()
+ mock_response.json.return_value = {
+ "linksets": [{"linksetdbs": [{"links": ["1", "2", "3"]}]}]
+ }
+ mock_response.text = (
+ '{"linksets": [{"linksetdbs": [{"links": ["1", "2", "3"]}]}]}'
+ )
+
+ mock_get.return_value = mock_response
+
+ toolkit = PubMedToolkit()
+ count = toolkit.get_citation_count("12345")
+
+ assert count == 3
+
+
+@patch('requests.get')
+def test_get_related_papers(mock_get):
+ # Mock the related papers response
+ mock_related_response = MagicMock()
+ mock_related_response.json.return_value = {
+ "linksets": [{"linksetdbs": [{"links": ["12345"]}]}]
+ }
+ mock_related_response.text = (
+ '{"linksets": [{"linksetdbs": [{"links": ["12345"]}]}]}'
+ )
+
+ # Mock the paper details response
+ mock_details_response = MagicMock()
+ mock_details_response.json.return_value = {
+ "result": {
+ "12345": {
+ "title": "Related Paper",
+ "authors": [{"name": "Author 1"}],
+ "source": "Test Journal",
+ "pubdate": "2024",
+ }
+ }
+ }
+ mock_details_response.text = (
+ '{"result": {"12345": {"title": "Related Paper"}}}'
+ )
+
+ # Mock the abstract response
+ mock_abstract_response = MagicMock()
+ mock_abstract_response.text = "Test abstract"
+
+ mock_get.side_effect = [
+ mock_related_response,
+ mock_details_response,
+ mock_abstract_response,
+ ]
+
+ toolkit = PubMedToolkit()
+ results = toolkit.get_related_papers("67890", max_results=1)
+
+ assert len(results) == 1
+ assert results[0]["title"] == "Related Paper"
+
+
+def test_invalid_timeout():
+ with pytest.raises(ValueError):
+ PubMedToolkit(timeout=-1)