codebase.py

from owlready2 import *
import uuid
import json
import pandas as pd

import os
import ollama

#### Alex' Funktionen
def dict2kg(onto,data):            
    user_id = uuid.uuid4().hex
    # get references to classes and data properties
    clPerson = onto.search(iri = "*Person")[0]  
    clOrganization = onto.search(iri = "*Organization")[0]
    #dpName = onto.search(iri = "*name")[0] # Data Property
    #dpAge = onto.search(iri = "*age")[0] # Data Property
    #rMember = onto.search(iri = "*member")[0] # Object Property
    # gibt klasse reaction name aus
    clReactionName = onto.search(iri = "*0007009")[0]
    # gibt klasse reaction name aus
    clCatalysisResearchField = onto.search(iri = "*0000196")[0]

    if type(data["voc4cat:CatalysisResearchField"]) != list:
        data["voc4cat:CatalysisResearchField"] = [data["voc4cat:CatalysisResearchField"]]

    if type(data["foaf:Organization"]) != list:
        data["foaf:Organization"] = [data["foaf:Organization"]]


    #person member organization
    if type(data["foaf:Organization"]) == list:
        for org in data["foaf:Organization"]:
            for resField in data["voc4cat:CatalysisResearchField"]:
                for resname in data["voc4cat:ReactionName"]:
                    codestring = """with onto:
                                                Indv = onto.search(label = "person_{}").first() # gibt liste zurueck
                                                if Indv == None: 
                                                    Indv = clPerson('{}')
                                                    Indv.label = "person_{}"
                                                    Indv.name = "{}"
                                                    Indv.givenName.append("{}")
                                                    Indv.age = "{}"       

                                                org_indv = onto.search(label = "{}").first() # gibt liste zurueck
                                                if org_indv == None: 
                                                    org_indv = clOrganization('{}')
                                                    org_indv.label = "{}"
                                                Indv.member.append(org_indv)

                                                reac_indv = onto.search(label = "{}").first() # gibt liste zurueck
                                                if reac_indv == None: 
                                                    reac_indv = clReactionName('{}')
                                                    reac_indv.label = "{}"
                                                Indv.favReaction.append(reac_indv)

                                                research_indv = onto.search(label = "{}").first() # gibt liste zurueck
                                                if research_indv == None: 
                                                    research_indv = clCatalysisResearchField('{}')
                                                    research_indv.label = "{}"
                                                Indv.favResearch.append(research_indv)
                                        """.format(user_id,
                                                user_id,
                                                user_id,
                                                user_id,
                                                data["foaf:name"],
                                                data["foaf:age"],
                                                org,
                                                org,
                                                org,
                                                resname,
                                                resname,
                                                resname,
                                                resField,
                                                resField,
                                                resField,
                                                )                    
                    code = compile(codestring, "<string>","exec")
                    exec(code)
    else:
        codestring = """with onto:
                                    Indv = onto.search(label = "person_{}").first() # gibt liste zurueck
                                    if Indv == None: 
                                        Indv = clPerson('{}')
                                        Indv.label = "person_{}"
                                        Indv.name = "{}"
                                        Indv.givenName.append("{}")
                                        Indv.age = "{}"       

                                    org_indv = onto.search(label = "{}").first() # gibt liste zurueck
                                    if org_indv == None: 
                                        org_indv = clOrganization('{}')
                                        org_indv.label = "{}"
                                    Indv.member.append(org_indv)

                                    reac_indv = onto.search(label = "{}").first() # gibt liste zurueck
                                    if reac_indv == None: 
                                        reac_indv = clReactionName('{}')
                                        reac_indv.label = "{}"
                                    Indv.favReaction.append(reac_indv)

                                    research_indv = onto.search(label = "{}").first() # gibt liste zurueck
                                    if research_indv == None: 
                                        research_indv = clCatalysisResearchField('{}')
                                        research_indv.label = "{}"
                                    Indv.favResearch.append(research_indv)
                            """.format(user_id,
                                    user_id,
                                    user_id,
                                    user_id,
                                    data["foaf:name"],
                                    data["foaf:age"],
                                    data["foaf:Organization"],
                                    data["foaf:Organization"],
                                    data["foaf:Organization"],
                                    data["voc4cat:ReactionName"],
                                    data["voc4cat:ReactionName"],
                                    data["voc4cat:ReactionName"],
                                    data["voc4cat:CatalysisResearchField"],
                                    data["voc4cat:CatalysisResearchField"],
                                    data["voc4cat:CatalysisResearchField"],
                                    )                    
        code = compile(codestring, "<string>","exec")
        exec(code)

        
    return onto

#####


def kg_query(onto_world, search_str, query_number):
    if query_number == 0:    
        # Give me all people whose favorite reaction contains search_str (case insensitive)
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?person ?name ?org ?reaction ?research 
            WHERE {{  ?person a foaf:Person.
                    ?person foaf:givenName ?name.
                    ?person foaf:member ?org.
                    ?person voc4cat:favReaction ?reaction.
                    ?person voc4cat:favResearch ?research.
                    ?reaction rdfs:label ?reacLab.
                    FILTER(CONTAINS(LCASE(?reacLab),LCASE("{}"))) 
            }} 
        """.format(search_str)
    elif query_number == 1:
        # Give me all people whose organization contains search_str (case insensitive)
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?person ?name ?org ?reaction ?research 
            WHERE {{  ?person a foaf:Person.
                    ?person foaf:givenName ?name.
                    ?person foaf:member ?org.
                    ?person voc4cat:favReaction ?reaction.
                    ?person voc4cat:favResearch ?research.
                    ?org rdfs:label ?orgLab.
                    FILTER(CONTAINS(LCASE(?orgLab),LCASE("{}"))) 
            }} 
        """.format(search_str)
    # search for people with search_string in research field
    elif query_number == 2:
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?person ?name ?org ?reaction ?research
            WHERE {{  ?person a foaf:Person.
                    ?person foaf:givenName ?name.
                    ?person foaf:member ?org.
                    ?person voc4cat:favReaction ?reaction.
                    ?person voc4cat:favResearch ?research.
                    ?research rdfs:label ?researchLab.
                    FILTER(CONTAINS(LCASE(?researchLab),LCASE("{}"))) 
            }} 
        """.format(search_str)
    # Search in Names
    elif query_number == 3:
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?person ?name ?org ?reaction ?research
            WHERE {{  ?person a foaf:Person.
                    ?person foaf:givenName ?name.
                    ?person foaf:member ?org.
                    ?person voc4cat:favReaction ?reaction.
                    ?person voc4cat:favResearch ?research.
                    FILTER(CONTAINS(LCASE(?name),LCASE("{}"))) 
            }} 
        """.format(search_str)
    
    # list all organizations
    elif query_number == 4:
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?orgL 
            WHERE {{?org a foaf:Organization.
                    ?org rdfs:label ?orgL.
            }} 
        """.format(search_str)
        
    # list all researchFields
    elif query_number == 5:
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?resFieldL 
            WHERE {{?resField a voc4cat:voc4cat_0000196.
                    ?resField rdfs:label ?resFieldL.
            }} 
        """.format(search_str)

    # list all voc4cat:ReactionName
    elif query_number == 6:
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?resFieldL 
            WHERE {{?resField a voc4cat:voc4cat_0007009.
                    ?resField rdfs:label ?resFieldL.
            }} 
        """.format(search_str)
    # people with same favourite reaction
    elif query_number == 7:
        query = """
        PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
        PREFIX owl: <http://www.w3.org/2002/07/owl#>
        PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
        PREFIX foaf: <http://xmlns.com/foaf/0.1/>
        PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

        SELECT DISTINCT ?person ?name ?org ?reaction ?research 
          WHERE {{  ?reaction voc4cat:favReactionOf ?person.
                    ?person foaf:givenName ?name.
                    ?person foaf:member ?org.
                    ?person voc4cat:favResearch ?research.
                    FILTER(CONTAINS(LCASE(?reaction),LCASE("{}"))) 
            }} 
        """.format(search_str)



    else:
        raise ValueError("query_number must be 0, 1, 2, 3, 4, 5, 6 or 7")

    results = list(onto_world.sparql(query))
    rows = []
    for r in results:
        row = [i if isinstance(i, str) else i.label.first() for i in r]
        rows.append(row)

    if query_number == 4:
        df = pd.DataFrame(rows, columns = ['org'])

    elif query_number == 5:
        df = pd.DataFrame(rows, columns = ['resField'])
    
    elif query_number == 6:
        df = pd.DataFrame(rows, columns = ['reacName'])

    else: 
        columns = ['person', 'name', 'org', 'reaction', 'research']
        df = pd.DataFrame(rows, columns=columns)

    return df
#####




##### Marcs Funktionen
def call_ollama(user_input):
    prompt = user_input

    # Extract JSON string from response
    try:
#        response = ollama.generate(model="summit_demo_new", prompt=prompt)
        response = ollama.generate(model="demo_software_metadata_extraction", prompt=prompt)
    except Exception as e:
        print(f"Error calling Ollama API: {e}")
        return None

    json_str = response["response"].split("json")[1][:-3].strip()

    # Parse JSON string to dict
    try:
        json_data = json.loads(json_str)
    except json.JSONDecodeError as e:
        print(f"Error parsing JSON: {e}")
        return None
    
    # Load existing data if file exists
    if os.path.exists("testData_copy.json"):
        with open("testData_copy.json", "r", encoding="utf-8") as f:
            try:
                result = json.load(f)
            except json.JSONDecodeError:
                result = {"person": {}}
    else:
        result = {"person": {}}

    # Save updated data
    with open("testData_copy.json", "w", encoding="utf-8") as f:
        json.dump(json_data, f, indent=4)
    
    return json_data
#####


# SPARQL QUERY TO BE TESTED

#    PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
#    PREFIX owl: <http://www.w3.org/2002/07/owl#>
#    PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
#    PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
#    PREFIX foaf: <http://xmlns.com/foaf/0.1/>
#    PREFIX voc4cat: <https://w3id.org/nfdi4cat/>

#         SELECT DISTINCT ?favReaction ?name 
#          WHERE {  ?favReaction a voc4cat:voc4cat_0007009.
#                   ?favReaction rdfs:label ?name.
#                   ?favReaction voc4cat:favReactionOf ?person.
#                  # FILTER(CONTAINS(LCASE(?name),LCASE("Ma"))) 
#             }