Adaptive Graph of Thoughts MCP Server

import httpx import pytest from pytest_httpx import HTTPXMock from adaptive_graph_of_thoughts.config import PubMedConfig, Settings from adaptive_graph_of_thoughts.infrastructure.api_clients.base_client import ( APIRequestError, ) from adaptive_graph_of_thoughts.infrastructure.api_clients.pubmed_client import ( PubMedClient, PubMedClientError, ) # --- Sample Data --- SAMPLE_ESEARCH_RESPONSE_SUCCESS_JSON = { "header": {"type": "esearch", "version": "2.0"}, "esearchresult": { "count": "2", "retmax": "2", "retstart": "0", "idlist": ["123456", "789012"], "translationset": [], "querytranslation": "sample query[all fields]", }, } SAMPLE_ESEARCH_SINGLE_ID_JSON = { "header": {"type": "esearch", "version": "2.0"}, "esearchresult": { "count": "1", "retmax": "1", "retstart": "0", "idlist": ["123456"], "translationset": [], "querytranslation": "sample query[all fields]", }, } SAMPLE_ESUMMARY_RESPONSE_XML_STR = """<eSummaryResult> <DocSum> <Id>123456</Id> <Item Name="Title" Type="String">Sample Article Title 1</Item> <Item Name="AuthorList" Type="List"> <Item Name="Author" Type="String">Author A</Item> <Item Name="Author" Type="String">Author B</Item> </Item> <Item Name="Source" Type="String">Journal of Samples</Item> <Item Name="PubDate" Type="String">2023 Jan</Item> <Item Name="DOI" Type="String">10.1000/sample.doi.1</Item> </DocSum> <DocSum> <Id>789012</Id> <Item Name="Title" Type="String">Sample Article Title 2</Item> <Item Name="AuthorList" Type="List"> <Item Name="Author" Type="String">Author C</Item> </Item> <Item Name="Source" Type="String">Another Journal</Item> <Item Name="PubDate" Type="String">2022 Dec</Item> <Item Name="DOI" Type="String">10.1000/sample.doi.2</Item> </DocSum> </eSummaryResult>""" SAMPLE_ESUMMARY_SINGLE_ARTICLE_XML_STR = """<eSummaryResult> <DocSum> <Id>123456</Id> <Item Name="Title" Type="String">Sample Article Title 1</Item> <Item Name="AuthorList" Type="List"> <Item Name="Author" Type="String">Author A</Item> <Item Name="Author" Type="String">Author B</Item> </Item> <Item Name="Source" Type="String">Journal of Samples</Item> <Item Name="PubDate" Type="String">2023 Jan</Item> <Item Name="DOI" Type="String">10.1000/sample.doi.1</Item> </DocSum> </eSummaryResult>""" SAMPLE_EFETCH_ABSTRACT_XML_STR_123456 = """<PubmedArticleSet> <PubmedArticle> <MedlineCitation Status="MEDLINE" Owner="NLM"> <PMID Version="1">123456</PMID> <Article PubModel="Print"> <Abstract> <AbstractText>This is abstract for PMID 123456.</AbstractText> </Abstract> </Article> </MedlineCitation> </PubmedArticle> </PubmedArticleSet>""" SAMPLE_EFETCH_ABSTRACT_XML_STR_789012 = """<PubmedArticleSet> <PubmedArticle> <MedlineCitation Status="MEDLINE" Owner="NLM"> <PMID Version="1">789012</PMID> <Article PubModel="Print"> <Abstract> <AbstractText>This is abstract for PMID 789012. It is longer.</AbstractText> </Abstract> </Article> </MedlineCitation> </PubmedArticle> </PubmedArticleSet>""" SAMPLE_EFETCH_NO_ABSTRACT_XML_STR = """<PubmedArticleSet> <PubmedArticle> <MedlineCitation Status="MEDLINE" Owner="NLM"> <PMID Version="1">999999</PMID> <Article PubModel="Print"> <!-- No Abstract Node --> </Article> </MedlineCitation> </PubmedArticle> </PubmedArticleSet>""" SAMPLE_EMPTY_ESEARCH_RESPONSE_JSON = { "header": {"type": "esearch", "version": "2.0"}, "esearchresult": {"count": "0", "retmax": "0", "retstart": "0", "idlist": []}, } # --- Fixtures --- @pytest.fixture def mock_settings() -> Settings: """ Creates a Settings instance with a minimal PubMed configuration for testing. Returns: A Settings object containing a PubMedConfig with a base URL and email. """ return Settings( pubmed=PubMedConfig( base_url="https://eutils.ncbi.nlm.nih.gov/entrez/eutils/", email="test@example.com", # Recommended by NCBI ) ) @pytest.fixture async def pubmed_client_fixture(mock_settings: Settings) -> PubMedClient: """ Provides an asynchronous fixture that yields a PubMedClient instance configured with mock settings for testing. """ async with PubMedClient(settings=mock_settings) as client: yield client # --- Test Cases --- def test_pubmed_client_initialization(mock_settings: Settings): """Test successful initialization of PubMedClient.""" client = PubMedClient(settings=mock_settings) assert client is not None assert client.config == mock_settings.pubmed assert client.email == "test@example.com" def test_pubmed_client_initialization_missing_config(): """Test PubMedClientError is raised if PubMed config is missing or incomplete.""" with pytest.raises( PubMedClientError, match="PubMed configuration (base_url) is not set" ): PubMedClient(settings=Settings(pubmed=None)) with pytest.raises( PubMedClientError, match="PubMed configuration (base_url) is not set" ): PubMedClient(settings=Settings(pubmed=PubMedConfig(base_url=None))) # type: ignore async def test_search_articles_success( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a successful article search returns correctly parsed articles with abstracts. Mocks the PubMed API endpoints for `esearch`, `esummary`, and `efetch` to simulate a search query returning two articles. Verifies that the resulting `PubMedArticle` objects contain accurate metadata and abstracts, and asserts that the correct API calls are made with expected parameters. """ client = pubmed_client_fixture # Mock esearch httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", method="GET", json=SAMPLE_ESEARCH_RESPONSE_SUCCESS_JSON, ) # Mock esummary httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esummary.fcgi", method="GET", text=SAMPLE_ESUMMARY_RESPONSE_XML_STR, ) # Mock efetch for PMID 123456 httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/efetch.fcgi?db=pubmed&id=123456&rettype=abstract&retmode=xml&email=test%40example.com", method="GET", text=SAMPLE_EFETCH_ABSTRACT_XML_STR_123456, ) # Mock efetch for PMID 789012 httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/efetch.fcgi?db=pubmed&id=789012&rettype=abstract&retmode=xml&email=test%40example.com", method="GET", text=SAMPLE_EFETCH_ABSTRACT_XML_STR_789012, ) articles = await client.search_articles("sample query", max_results=2) assert len(articles) == 2 article1 = next(a for a in articles if a.pmid == "123456") assert article1.title == "Sample Article Title 1" assert article1.authors == ["Author A", "Author B"] assert article1.journal == "Journal of Samples" assert article1.publication_date == "2023 Jan" assert article1.doi == "10.1000/sample.doi.1" assert article1.url == "https://pubmed.ncbi.nlm.nih.gov/123456/" assert article1.abstract == "This is abstract for PMID 123456." article2 = next(a for a in articles if a.pmid == "789012") assert article2.title == "Sample Article Title 2" assert article2.abstract == "This is abstract for PMID 789012. It is longer." # Verify calls esearch_req = httpx_mock.get_request( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi" ) assert esearch_req is not None assert "term=sample+query" in str(esearch_req.url) assert "retmax=2" in str(esearch_req.url) esummary_req = httpx_mock.get_request( url=f"{client.config.base_url.rstrip('/')}/esummary.fcgi" ) assert esummary_req is not None assert "id=123456%2C789012" in str(esummary_req.url) # comma is %2C efetch_reqs = httpx_mock.get_requests(url__regex=r".*efetch\.fcgi.*") assert len(efetch_reqs) == 2 async def test_search_articles_no_pmids( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that searching for articles returns an empty list when the eSearch API response contains no PMIDs, and verifies that subsequent eSummary and eFetch API calls are not made. """ client = pubmed_client_fixture httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", json=SAMPLE_EMPTY_ESEARCH_RESPONSE_JSON, ) articles = await client.search_articles("query_with_no_results") assert len(articles) == 0 # Ensure esummary and efetch were not called assert len(httpx_mock.get_requests(url__regex=r".*esummary\.fcgi.*")) == 0 assert len(httpx_mock.get_requests(url__regex=r".*efetch\.fcgi.*")) == 0 async def test_search_articles_esearch_http_error( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a PubMedClientError is raised when the eSearch API returns an HTTP error. """ client = pubmed_client_fixture httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", method="GET", status_code=500, text="Internal Server Error", ) with pytest.raises(PubMedClientError, match="PubMed eSearch failed"): await client.search_articles("test query") async def test_search_articles_esummary_http_error( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a PubMedClientError is raised when the esummary API call returns an HTTP error. """ client = pubmed_client_fixture httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", json=SAMPLE_ESEARCH_SINGLE_ID_JSON, # Success for esearch ) httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esummary.fcgi", method="GET", status_code=400, text="Bad Request", ) with pytest.raises(PubMedClientError, match="PubMed eSummary failed"): await client.search_articles("test query") async def test_search_articles_efetch_http_error( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a PubMedClientError is raised when an HTTP error occurs during the efetch step of the search_articles workflow. """ client = pubmed_client_fixture httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", json=SAMPLE_ESEARCH_SINGLE_ID_JSON, ) httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esummary.fcgi", text=SAMPLE_ESUMMARY_SINGLE_ARTICLE_XML_STR, ) httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/efetch.fcgi", # Matches any efetch here method="GET", status_code=503, text="Service Unavailable", ) # The error could be PubMedClientError wrapping APIHTTPError or a more specific message # Depending on how fetch_abstract propagates errors if one abstract fails among many. # The current implementation of search_articles might raise if any efetch fails. with pytest.raises( PubMedClientError, match=r"PubMed eFetch failed \(PMID 123456\)|Could not fetch abstract for PMID 123456", ): await client.search_articles("test query") async def test_fetch_abstract_direct_success( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that fetch_abstract returns the correct abstract string when the efetch API responds successfully. """ client = pubmed_client_fixture pmid = "123456" httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/efetch.fcgi?db=pubmed&id={pmid}&rettype=abstract&retmode=xml&email=test%40example.com", text=SAMPLE_EFETCH_ABSTRACT_XML_STR_123456, ) abstract = await client.fetch_abstract(pmid) assert abstract == "This is abstract for PMID 123456." async def test_fetch_abstract_direct_not_found( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that fetch_abstract returns None when the efetch response contains no abstract for the given PMID. """ client = pubmed_client_fixture pmid = "999999" httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/efetch.fcgi?db=pubmed&id={pmid}&rettype=abstract&retmode=xml&email=test%40example.com", text=SAMPLE_EFETCH_NO_ABSTRACT_XML_STR, ) abstract = await client.fetch_abstract(pmid) assert abstract is None async def test_api_key_and_email_usage(httpx_mock: HTTPXMock): """ Verifies that the PubMedClient includes both api_key and email parameters in all PubMed API requests when configured. This test ensures that the email and api_key are present in the query parameters for esearch, esummary, and efetch requests. """ settings_with_key = Settings( pubmed=PubMedConfig( base_url="https://eutils.ncbi.nlm.nih.gov/entrez/eutils/", email="user_with_key@example.com", api_key="testapikey123", ) ) async with PubMedClient(settings=settings_with_key) as client: # Mock esearch, esummary, efetch to allow search_articles to complete httpx_mock.add_response( url__regex=r".*esearch\.fcgi.*", # Regex to catch any esearch call json=SAMPLE_ESEARCH_SINGLE_ID_JSON, ) httpx_mock.add_response( url__regex=r".*esummary\.fcgi.*", text=SAMPLE_ESUMMARY_SINGLE_ARTICLE_XML_STR, ) httpx_mock.add_response( url__regex=r".*efetch\.fcgi.*", text=SAMPLE_EFETCH_ABSTRACT_XML_STR_123456 ) await client.search_articles("test query for api key", max_results=1) # Check esearch call esearch_request = httpx_mock.get_request(url__regex=r".*esearch\.fcgi.*") assert esearch_request is not None request_params_esearch = esearch_request.url.params assert request_params_esearch.get("email") == "user_with_key@example.com" assert request_params_esearch.get("api_key") == "testapikey123" # Check esummary call esummary_request = httpx_mock.get_request(url__regex=r".*esummary\.fcgi.*") assert esummary_request is not None request_params_esummary = esummary_request.url.params assert request_params_esummary.get("email") == "user_with_key@example.com" assert request_params_esummary.get("api_key") == "testapikey123" # Check efetch call efetch_request = httpx_mock.get_request(url__regex=r".*efetch\.fcgi.*") assert efetch_request is not None request_params_efetch = efetch_request.url.params assert request_params_efetch.get("email") == "user_with_key@example.com" assert request_params_efetch.get("api_key") == "testapikey123" async def test_search_articles_esearch_request_error( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a network exception during the PubMed eSearch API call results in a PubMedClientError being raised. """ client = pubmed_client_fixture httpx_mock.add_exception( pytest.raises( APIRequestError ), # This is not how httpx_mock expects exceptions. It expects an exception instance. url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", method="GET", ) # Correct way to mock an exception with httpx_mock: httpx_mock.add_exception( Exception("Simulated network error"), # General exception for simplicity url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", method="GET", ) with pytest.raises(PubMedClientError, match="PubMed eSearch failed"): await client.search_articles("test query") # Consider adding tests for invalid XML parsing in _parse_esummary_response # and _parse_abstract_from_efetch_xml if robust error handling for malformed XML is critical. # For instance, if ET.fromstring fails. # The current client implementation catches ET.ParseError and re-raises as PubMedClientError. async def test_parse_esummary_malformed_xml( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a PubMedClientError is raised when the eSummary API returns malformed XML. """ client = pubmed_client_fixture httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", json=SAMPLE_ESEARCH_SINGLE_ID_JSON, ) httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esummary.fcgi", text="<malformed_xml>", # Not well-formed XML ) with pytest.raises(PubMedClientError, match="XML parsing error for eSummary"): await client.search_articles("test query") async def test_fetch_abstract_malformed_xml( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a PubMedClientError is raised when eFetch returns malformed XML during abstract retrieval. """ client = pubmed_client_fixture pmid = "123456" httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/efetch.fcgi?db=pubmed&id={pmid}&rettype=abstract&retmode=xml&email=test%40example.com", text="<unclosedTag", ) with pytest.raises( PubMedClientError, match=rf"XML parsing error for eFetch \(PMID {pmid}\)" ): await client.fetch_abstract(pmid) # --- Additional Edge-Case Tests --- def test_pubmed_client_invalid_base_url(): """Test initialization fails with invalid base_url.""" bad_settings = Settings( pubmed=PubMedConfig(base_url="htt://bad url", email="x@y.com") ) with pytest.raises(PubMedClientError, match="base_url"): PubMedClient(settings=bad_settings) @pytest.mark.asyncio async def test_search_articles_timeout_handling( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that a timeout during the eSearch API call raises a PubMedClientError. """ client = pubmed_client_fixture esearch_url = f"{client.config.base_url.rstrip('/')}/esearch.fcgi" httpx_mock.add_exception(httpx.TimeoutException("timeout"), url=esearch_url) with pytest.raises(PubMedClientError, match="eSearch failed"): await client.search_articles("timeout query") @pytest.mark.asyncio async def test_search_articles_large_retmax( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that search_articles correctly passes a large max_results value as the retmax parameter and returns all articles found. Verifies that the client requests the correct number of results from the PubMed API and that all returned articles are processed. """ client = pubmed_client_fixture # Create esearch response with three IDs large_esearch = { "header": {"type": "esearch", "version": "2.0"}, "esearchresult": { "count": "3", "retmax": "3", "retstart": "0", "idlist": ["1", "2", "3"], "translationset": [], "querytranslation": "q", }, } httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esearch.fcgi", method="GET", json=large_esearch, ) # Mock esummary with minimal DocSum entries esummary_xml = ( "<eSummaryResult>" + "".join(f"<DocSum><Id>{pmid}</Id></DocSum>" for pmid in ["1", "2", "3"]) + "</eSummaryResult>" ) httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/esummary.fcgi", method="GET", text=esummary_xml, ) # Mock efetch for each PMID for pmid in ["1", "2", "3"]: efetch_xml = ( "<PubmedArticleSet>" f"<PubmedArticle><MedlineCitation><Article><Abstract><AbstractText>Abstract {pmid}</AbstractText></Abstract></Article></MedlineCitation></PubmedArticleSet>" ) httpx_mock.add_response( url__regex=r".*efetch\.fcgi.*", method="GET", text=efetch_xml ) articles = await client.search_articles("q", max_results=1000) assert len(articles) == 3 esearch_req = httpx_mock.get_request(url__regex=r".*esearch\.fcgi.*") assert esearch_req is not None params = esearch_req.url.params assert params.get("retmax") == "1000" @pytest.mark.parametrize( "xml_str, expected_titles, expected_authors", [ ( "<eSummaryResult><DocSum><Id>1</Id></DocSum></eSummaryResult>", [""], [[]], ), ( '<eSummaryResult><DocSum><Id>1</Id><Item Name="Title" Type="String">Title</Item></DocSum></eSummaryResult>', ["Title"], [[]], ), ( '<eSummaryResult><DocSum><Id>1</Id><Item Name="AuthorList" Type="List"></Item></DocSum></eSummaryResult>', [""], [[]], ), ], ) def test_parse_esummary_response_edge_cases(xml_str, expected_titles, expected_authors): """ Tests that the eSummary XML parser correctly handles edge cases with missing title or author fields. Args: xml_str: The eSummary XML string to parse. expected_titles: List of expected article titles after parsing. expected_authors: List of expected author lists after parsing. """ client = PubMedClient( settings=Settings(pubmed=PubMedConfig(base_url="http://url", email="e@e.com")) ) articles = client._parse_esummary_response(xml_str) titles = [a.title for a in articles] authors = [a.authors for a in articles] assert titles == expected_titles assert authors == expected_authors @pytest.mark.asyncio async def test_fetch_abstract_multiple_articles_in_set( pubmed_client_fixture: PubMedClient, httpx_mock: HTTPXMock ): """ Tests that fetch_abstract returns the abstract of the first article when the efetch response contains multiple articles. Verifies that when multiple articles are present in the efetch XML response, the abstract from the first article is returned. """ client = pubmed_client_fixture combined_xml = SAMPLE_EFETCH_ABSTRACT_XML_STR_123456.replace( "</PubmedArticleSet>", "" ) + SAMPLE_EFETCH_ABSTRACT_XML_STR_789012.replace("<PubmedArticleSet>", "") pmid = "123456" httpx_mock.add_response( url=f"{client.config.base_url.rstrip('/')}/efetch.fcgi?db=pubmed&id={pmid}&rettype=abstract&retmode=xml&email=test%40example.com", method="GET", text=combined_xml, ) abstract = await client.fetch_abstract(pmid) # Based on implementation, should return the first AbstractText encountered assert abstract == "This is abstract for PMID 123456."