naas-abi 1.0.0__py3-none-any.whl

naas_abi/workflows/SearchIndividualWorkflow.py

@@ -0,0 +1,190 @@
+from dataclasses import dataclass
+from enum import Enum
+from typing import Annotated, Optional
+
+from fastapi import APIRouter
+from langchain_core.tools import BaseTool, StructuredTool
+from naas_abi_core.services.triple_store.TripleStorePorts import ITripleStoreService
+from naas_abi_core.utils.String import normalize
+from naas_abi_core.workflow import Workflow, WorkflowConfiguration
+from naas_abi_core.workflow.workflow import WorkflowParameters
+from pydantic import Field
+from rdflib import query
+from thefuzz import fuzz  # type: ignore
+
+
+@dataclass
+class SearchIndividualWorkflowConfiguration(WorkflowConfiguration):
+    """Configuration for SearchIndividual workflow."""
+
+    triple_store: ITripleStoreService
+
+
+class SearchIndividualWorkflowParameters(WorkflowParameters):
+    """Parameters for SearchIndividual workflow."""
+
+    search_label: Annotated[
+        str,
+        Field(
+            ...,
+            description="Individual label to search for in the ontology schema.",
+            example="Naas.ai",
+        ),
+    ]
+    class_uri: Optional[
+        Annotated[
+            str,
+            Field(
+                ...,
+                description="Class URI to use to search for individuals.",
+                example="https://www.commoncoreontologies.org/ont00000443",
+            ),
+        ]
+    ] = None
+    limit: Optional[
+        Annotated[
+            int,
+            Field(
+                description="Maximum number of results to return.",
+                ge=1,
+                le=100,
+            ),
+        ]
+    ] = 10
+    query: Optional[
+        Annotated[
+            str,
+            Field(
+                description="Custom SPARQL query to use to search for individuals.",
+            ),
+        ]
+    ] = None
+
+
+class SearchIndividualWorkflow(Workflow):
+    """Workflow for searching ontology individuals."""
+
+    __configuration: SearchIndividualWorkflowConfiguration
+
+    def __init__(self, configuration: SearchIndividualWorkflowConfiguration):
+        super().__init__(configuration)
+        self.__configuration = configuration
+
+    def search_individual(
+        self, parameters: SearchIndividualWorkflowParameters
+    ) -> dict | list[dict]:
+        class_uri_filter = (
+            f"?individual_uri a <{parameters.class_uri}> ;"
+            if parameters.class_uri
+            else ""
+        )
+        search_label_norm = normalize(parameters.search_label)
+        sparql_query = f"""
+        SELECT DISTINCT ?individual_uri ?label
+        WHERE {{
+            # Ensure individual is a NamedIndividual
+            ?individual_uri a owl:NamedIndividual .
+            
+            {class_uri_filter}
+
+            {{
+                # Search by rdfs:label
+                ?individual_uri rdfs:label ?label .
+                
+                # Convert labels to lowercase for comparison
+                BIND(LCASE(STR(?label)) AS ?lower_label)
+                BIND(LCASE("{parameters.search_label}") AS ?lower_search)
+                
+                # Split search label into words and check if any word matches
+                {{
+                    # Full label match
+                    FILTER(CONTAINS(?lower_label, ?lower_search) || CONTAINS(?lower_search, ?lower_label))
+                }}
+                UNION
+                {{
+                    # Word-by-word match
+                    VALUES ?search_word {{ {" ".join([f'"{word.lower()}"' for word in parameters.search_label.split()])} }}
+                    FILTER(CONTAINS(?lower_label, ?search_word))
+                }}
+            }}
+            UNION
+            {{
+                # Search by skos:altLabel
+                ?individual_uri skos:altLabel ?label .
+                
+                # Convert labels to lowercase for comparison
+                BIND(LCASE(STR(?label)) AS ?lower_label)
+                BIND(LCASE("{search_label_norm}") AS ?lower_search)
+                
+                # Split search label into words and check if any word matches
+                {{
+                    # Full label match
+                    FILTER(CONTAINS(?lower_label, ?lower_search) || CONTAINS(?lower_search, ?lower_label))
+                }}
+                UNION
+                {{
+                    # Word-by-word match
+                    VALUES ?search_word {{ {" ".join([f'"{word.lower()}"' for word in search_label_norm.split()])} }}
+                    FILTER(CONTAINS(?lower_label, ?search_word))
+                }}
+            }}
+        }}
+        ORDER BY ?label
+        LIMIT {parameters.limit}
+        """
+        if parameters.query is not None:
+            sparql_query = parameters.query
+        results = self.__configuration.triple_store.query(sparql_query)
+
+        # Process results with fuzzy matching
+        data = []
+        for row in results:
+            assert isinstance(row, query.ResultRow)
+            data_dict = {}
+            for key in row.labels:
+                data_dict[key] = str(row[key]) if row[key] else None
+                if key == "label" and row[key]:
+                    score = fuzz.token_set_ratio(
+                        search_label_norm, normalize(str(row[key]))
+                    )
+                    data_dict["score"] = score
+            data.append(data_dict)
+
+        # Sort by best match and remove duplicates
+        if len(data) > 0:
+            data.sort(
+                key=lambda x: x["score"] if x["score"] is not None else 0, reverse=True
+            )
+            seen = set()
+            unique_data = []
+            for x in data:
+                if x["individual_uri"] not in seen:
+                    seen.add(x["individual_uri"])
+                    unique_data.append(x)
+            data = unique_data
+        return data
+
+    def as_tools(self) -> list[BaseTool]:
+        return [
+            StructuredTool(
+                name="search_individual",
+                description="Search an ontology individual based on its label. It will return the most relevant individual using matching with rdfs:label",
+                func=lambda **kwargs: self.search_individual(
+                    SearchIndividualWorkflowParameters(**kwargs)
+                ),
+                args_schema=SearchIndividualWorkflowParameters,
+            )
+        ]
+
+    def as_api(
+        self,
+        router: APIRouter,
+        route_name: str = "",
+        name: str = "",
+        description: str = "",
+        description_stream: str = "",
+        tags: list[str | Enum] | None = None,
+    ) -> None:
+        if tags is None:
+            tags = []
+        return None

naas_abi/workflows/SearchIndividualWorkflow_test.py

@@ -0,0 +1,98 @@
+import pytest
+from naas_abi.workflows.SearchIndividualWorkflow import (
+    SearchIndividualWorkflow,
+    SearchIndividualWorkflowConfiguration,
+    SearchIndividualWorkflowParameters,
+)
+
+
+@pytest.fixture
+def search_individual_workflow() -> SearchIndividualWorkflow:
+    from naas_abi import services
+
+    # Configurations
+    search_individual_workflow_configuration = SearchIndividualWorkflowConfiguration(
+        triple_store=services.triple_store_service
+    )
+
+    # Workflow
+    search_individual_workflow = SearchIndividualWorkflow(
+        configuration=search_individual_workflow_configuration
+    )
+
+    return search_individual_workflow
+
+
+def test_search_individual_workflow(
+    search_individual_workflow: SearchIndividualWorkflow,
+):
+    from uuid import uuid4
+
+    from naas_abi import services
+    from naas_abi_core.utils.Graph import TEST
+    from rdflib import OWL, RDFS, Graph, Literal, URIRef
+
+    graph = Graph()
+    node_id = str(uuid4())
+    graph.add(
+        (
+            TEST[node_id],
+            URIRef("http://www.w3.org/1999/02/22-rdf-syntax-ns#type"),
+            URIRef("https://www.commoncoreontologies.org/ont00000443"),
+        )
+    )
+    graph.add(
+        (
+            TEST[node_id],
+            URIRef("http://www.w3.org/1999/02/22-rdf-syntax-ns#type"),
+            OWL.NamedIndividual,
+        )
+    )
+    graph.add((TEST[node_id], RDFS.label, Literal(node_id)))
+
+    services.triple_store_service.insert(graph)
+
+    # Test
+    result = search_individual_workflow.search_individual(
+        SearchIndividualWorkflowParameters(
+            search_label=node_id,
+            class_uri="https://www.commoncoreontologies.org/ont00000443",
+            limit=10,
+            query=None,
+        )
+    )
+
+    assert isinstance(result, list), result
+    assert len(result) == 1, result
+    assert result[0]["label"] == node_id, result[0]
+
+    services.triple_store_service.remove(graph)
+
+    # Test
+    result = search_individual_workflow.search_individual(
+        SearchIndividualWorkflowParameters(
+            search_label=node_id,
+            class_uri="https://www.commoncoreontologies.org/ont00000443",
+            limit=10,
+            query=None,
+        )
+    )
+
+    assert len(result) == 0, result
+
+
+def test_search_unexisting_individual(
+    search_individual_workflow: SearchIndividualWorkflow,
+):
+    # Test
+    result = search_individual_workflow.search_individual(
+        SearchIndividualWorkflowParameters(
+            search_label="kndfggkljkdfjgldfklgdkfgdfg_does_not_exist",
+            class_uri="https://www.commoncoreontologies.org/ont00000443",
+            limit=10,
+            query=None,
+        )
+    )
+
+    assert isinstance(result, list), result
+    assert len(result) == 0, result

naas_abi-1.0.0.dist-info/METADATA

@@ -0,0 +1,9 @@
+Metadata-Version: 2.4
+Name: naas-abi
+Version: 1.0.0
+Summary: Add your description here
+Author-email: Maxime Jublou <maxime@naas.ai>, Florent Ravenel <florent@naas.ai>, Jeremy Ravenel <jeremy@naas.ai>
+Requires-Python: <4,>=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: naas-abi-core>=1.0.0-dev.1
+Requires-Dist: naas-abi-marketplace>=1.0.0-dev.1

naas_abi-1.0.0.dist-info/RECORD

@@ -0,0 +1,62 @@
+naas_abi/__init__.py,sha256=xOa5ED-cLAdFrosaPUJGPT3ThluEfx-0hWSmKzfzcmM,1193
+naas_abi/cli.py,sha256=bfCcxGoQ_iSMhQDDX0-dRCyhc53Nawsy3WAruCNCL7M,27549
+naas_abi/mappings.py,sha256=qu17k8cPwPNKRIqchDMnxViDx0Kn0MTy2XT8XYUIW34,4872
+naas_abi/triggers.py,sha256=ljkgZP2zsTlmkK02ivTPW-yV02pn2TdyRk5CP4kUtvo,4658
+naas_abi/agents/AbiAgent.py,sha256=xbpjtsGIqIOMrg5jL8acpKRCAE2Ik_us5sr8ttbnQRk,18071
+naas_abi/agents/AbiAgent_test.py,sha256=AKggG5RS3zDWnql1DFE7cF7riAHeOdKTZp44WN65Arg,5906
+naas_abi/agents/EntitytoSPARQLAgent.py,sha256=Muf4L2A4B8vV4JKYkj_P0p-5u1XyWIHfWY9S-dmxHmg,57095
+naas_abi/agents/EntitytoSPARQLAgent_test.py,sha256=kx0gkbJoQZqt2_f_RINajwQU5EnHR1PsQx53rh7wJvA,2267
+naas_abi/agents/KnowledgeGraphBuilderAgent.py,sha256=dx0_2VzVsLRi3cRiA0eqmVr-O7qbBrrC5ykW9nbW4_A,13022
+naas_abi/agents/KnowledgeGraphBuilderAgent_test.py,sha256=pKgakywjYCkjEngbaG-D9q9gwNMQEgXeD1C2xb9irTo,2718
+naas_abi/agents/OntologyEngineerAgent.py,sha256=QZRQPRoXfgoT3qqH0U6wM8_o3Zn1KLv0CtYiQkHtC1M,4571
+naas_abi/agents/OntologyEngineerAgent_test.py,sha256=Z5A9u-GKPMlaaUQmI5Yvm9N9mwZTOuyyrqIyrSwrCEI,1319
+naas_abi/apps/oxigraph_admin/main.py,sha256=PHrzl4KOKWMZRLg32SEbEHIhPUj3Uf-Ccqj3072OFYc,13970
+naas_abi/apps/oxigraph_admin/terminal_style.py,sha256=ByiWVYM5Ju8iGMKbs1YwfEwLfLIpaQbB5nPHzb9zdUY,5152
+naas_abi/apps/sparql_terminal/main.py,sha256=v2TD5qFBGlPGajceGdBTTT8QM0f4v-knF-uIKOGLr_c,1824
+naas_abi/apps/sparql_terminal/terminal_style.py,sha256=TYBM2ChnOLSgVw3TA33obBSc_5gAC5kJCuD2dqMOmF4,6560
+naas_abi/apps/terminal_agent/main.py,sha256=7MN88mivIV7gP1lELnqNxP_huwTbavhnTjkNmC9NMw8,19425
+naas_abi/apps/terminal_agent/terminal_style.py,sha256=YOpfvBlKU52wNyGEKbZPiRQVKxug-hI92H8ScV5L8Ew,5254
+naas_abi/models/airgap_gemma.py,sha256=I9pWN_DSgy4h6rT7OiYWDVijyTMoEgONm4xfZYWvp6I,8469
+naas_abi/models/airgap_qwen.py,sha256=EQZ-0tAKZCTCqnovLiGCx4ss0yh2Jjk0xh7wCwDomG8,682
+naas_abi/models/default.py,sha256=Ylj0GzyUDXjmuU0a2V759128bIVmvs-YXGc7o7oJSEQ,648
+naas_abi/models/gpt_4_1.py,sha256=OhwRr49VULUYxatj0cvM5g4mUkaOnl9qb9XK6AuLIiQ,818
+naas_abi/pipelines/AIAgentOntologyGenerationPipeline.py,sha256=ou-0FWBPFc5zkG3q_H3ri9_6AEm9JFDH3r4t5ThpxeU,24621
+naas_abi/pipelines/AIAgentOntologyGenerationPipeline_test.py,sha256=lDsuZKfQlCtLNKQ09nkSHemsvjSNdeZAhaDw-cPYbjo,4130
+naas_abi/pipelines/AddIndividualPipeline.py,sha256=KdrE4eY6Toh9pcJorhZUQ122zt2G5R0ewl9f97A4SVI,8426
+naas_abi/pipelines/AddIndividualPipeline_test.py,sha256=TXCdMdCZCy6Ap1Ar3GCQvQqtIHdEjcECegFCXxbJZeU,2275
+naas_abi/pipelines/InsertDataSPARQLPipeline.py,sha256=EcYRAr9ZIA-bWnKj9vclbgsN9uzXkACXzcJdH5HNl4k,6779
+naas_abi/pipelines/InsertDataSPARQLPipeline_test.py,sha256=XW_CEJfbg-Azi0Y9JjFfEN6ar30a8b2yHSmn0N1oe2A,3245
+naas_abi/pipelines/MergeIndividualsPipeline.py,sha256=6Z644v2OU47jFpfKj6uH6DK0hlbdMTjoR477Iy5dF04,9187
+naas_abi/pipelines/MergeIndividualsPipeline_test.py,sha256=c3QVeXRkEOxZfIPPDQdEpI_9v95R1rXPH8EUWlYtPqk,2990
+naas_abi/pipelines/RemoveIndividualPipeline.py,sha256=aU6ZRWisVQNPVkIwCaJ5BbthyOR5ZW5NGxnUr-Q_QYk,5514
+naas_abi/pipelines/RemoveIndividualPipeline_test.py,sha256=e5kgfT44bFl5wvgr3hSs1mu2tEq1xEyTIiF3KosKiNk,1697
+naas_abi/pipelines/UpdateCommercialOrganizationPipeline.py,sha256=JLuRbVa7Ic6FafEnDL8IoDf6rZqeVTJjRKHPipv_uBU,6749
+naas_abi/pipelines/UpdateDataPropertyPipeline.py,sha256=wtF5akFg6OpOakK4kCjfXhrc3fQkUYNDJuQVei8-qvM,6704
+naas_abi/pipelines/UpdateLegalNamePipeline.py,sha256=5gWMJ8_Uf39g-rhhHb6xBF4g7CTkZH9lrs6HEz5mdEA,3432
+naas_abi/pipelines/UpdateLinkedInPagePipeline.py,sha256=Amb-PH1JIB83axd2Gv_auQHWJmhE1JUBgPv0HOEcOdQ,6161
+naas_abi/pipelines/UpdatePersonPipeline.py,sha256=rycR5IhI7ZP0-YkCflaZLhWX6yzrM2Pd8aYWuhrUxYQ,6244
+naas_abi/pipelines/UpdateSkillPipeline.py,sha256=mv-rnxD_tcZEvJrskRbzIjoIe_Guo1-illcIUeO6YM0,3888
+naas_abi/pipelines/UpdateTickerPipeline.py,sha256=xRnZHKibyJCgsDXGzEsWCHCIWllnecZANSKkjJI87jA,3378
+naas_abi/pipelines/UpdateWebsitePipeline.py,sha256=LboGip58H8dE04jCNAF8jygt21CtmKua-xDwDa1B_Fw,3537
+naas_abi/workflows/AgentRecommendationWorkflow.py,sha256=QSAKib22LiWQQIbuwblQk8AE0b0WAX_huJLJeSC6KdE,12974
+naas_abi/workflows/AgentRecommendationWorkflow_test.py,sha256=9LaYrn3qNbPtKFOBxxL9o6pFKZ3aSxlU3IkKmvh2ok8,5205
+naas_abi/workflows/ArtificialAnalysisWorkflow.py,sha256=tieCFk-JIKbyVLlbdPDFIZLouO58B9ECVGIklgx8KV0,12114
+naas_abi/workflows/ArtificialAnalysisWorkflow_test.py,sha256=glb0WJa19EpyuTqg7SCDkSI5oo_7UCTiDSeHdpaDDUY,1928
+naas_abi/workflows/ConvertOntologyGraphToYamlWorkflow.py,sha256=T5AlTAVDgtoi8joLr9XDVuedEibzKV2E3j7Rss3hreM,7800
+naas_abi/workflows/ConvertOntologyGraphToYamlWorkflow_test.py,sha256=mwKtMzXPe5Iv19n5-XbiIk45IdOuLvHLeTUqSiTyEKY,2467
+naas_abi/workflows/CreateClassOntologyYamlWorkflow.py,sha256=H8znWADOAzlU_0RZCfsmnGH2sFJn8gkIM3p__PIRWqo,7747
+naas_abi/workflows/CreateClassOntologyYamlWorkflow_test.py,sha256=ql_cYfFBFod3sCKEXU0AR_oHtvd5Ykb0OG-tWhYorfk,2091
+naas_abi/workflows/CreateIndividualOntologyYamlWorkflow.py,sha256=kgJ_G4C4VUYfxW1oeDN1rzEYCvujsd_ENil1A3zLjgM,6814
+naas_abi/workflows/CreateIndividualOntologyYamlWorkflow_test.py,sha256=IM5nuTuEG2iYQMCkXBxxNQ0N2I7ua_aWpIVUmIfsELI,2791
+naas_abi/workflows/ExportGraphInstancesToExcelWorkflow.py,sha256=tgSZDqX70QeM7kam4JAK-C2MAQX-qQ8JP-97XafAaPY,19439
+naas_abi/workflows/ExportGraphInstancesToExcelWorkflow_test.py,sha256=N9QKd7lX7N_Q16DAJLQkzKcT07eGJ9zpmiYUXU21OLc,1170
+naas_abi/workflows/GetObjectPropertiesFromClassWorkflow.py,sha256=8Hcqm3bqj-KLjxTFIfGf9Ku9c6U0d2A9VdQpY0RS9FE,14505
+naas_abi/workflows/GetObjectPropertiesFromClassWorkflow_test.py,sha256=GUcIU9m_qwUqRXq8HwQrVW520PGNwq2N5nsNnUbbAkU,1965
+naas_abi/workflows/GetSubjectGraphWorkflow.py,sha256=sKNddfFrk1VNvogcoVxv_Bk-LKlw6XFe5Tgwk3HvH3U,2825
+naas_abi/workflows/GetSubjectGraphWorkflow_test.py,sha256=fzzFYUwojM39x5Ld-59jHMDbGy4Lcq6uO8_BAvZV0ug,1886
+naas_abi/workflows/SearchIndividualWorkflow.py,sha256=d0JasMTP7AEOyvEnZr1KHgagkBo9Tugi2gsk_m6g0E4,6553
+naas_abi/workflows/SearchIndividualWorkflow_test.py,sha256=KBhd58TATMSWWdThYPg7veAySR4hts5LuoygQqQIiaU,2749
+naas_abi-1.0.0.dist-info/METADATA,sha256=yKXsj05nPs1AK97s_fMSmm4pXPURBS2eh8bT4eMtCc8,358
+naas_abi-1.0.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+naas_abi-1.0.0.dist-info/top_level.txt,sha256=Dhf3mhATB9-JuvBR8Y55OabvI-PCeQUqb5ae3EMmbeA,9
+naas_abi-1.0.0.dist-info/RECORD,,

naas_abi-1.0.0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (80.9.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

naas_abi-1.0.0.dist-info/top_level.txt

@@ -0,0 +1 @@
+naas_abi