Complete Guide to Implement Knowledge Graph Using Python

It consists of sub fields which cannot be easily solved. Therefore, an approach to store data in a structured manner is Knowledge Graph which is ... Publishedon October31,2020 In DevelopersCorner CompleteGuidetoImplementKnowledgeGraphUsingPython InformationExtractionisaprocessofextractinginformationinamorestructuredwayi.e.,theinformationwhichismachine-understandable.Itconsistsofsubfieldswhichcannotbeeasilysolved.Therefore,anapproachtostoredatainastructuredmannerisKnowledgeGraphwhichisasetofthree-itemsetscalledTriplewherethesetcombinesasubject,apredicateandanobject. By AnkitDas InformationExtractionisaprocessofextractinginformationinamorestructuredwayi.e.,theinformationwhichismachine-understandable.Itconsistsofsubfieldswhichcannotbeeasilysolved.Therefore,anapproachtostoredatainastructuredmannerisKnowledgeGraphwhichisasetofthree-itemsetscalledTriplewherethesetcombinesasubject,apredicateandanobject. Inthisarticle,wewilldiscusshowtobuildaknowledgegraphusingPythonandSpacy. THEBELAMY Signupforyourweeklydoseofwhat'supinemergingtechnology. Email Signup Let’sgetstarted. CodeImplementation Importallthelibrariesrequiredforthisproject. importspacy fromspacy.lang.enimportEnglish importnetworkxasnx importmatplotlib.pyplotasplt ThesehubswillbetheelementsthatareavailableinWikipedia.Edgesaretheconnectionsinterfacingtheseelementstoeachother.Wewillextricatethesecomponentsinanunaidedway,i.e.,wewillutilizethepunctuationofthesentences.  Theprimarythoughtistoexperienceasentenceandconcentratethesubjectandtheitemasandwhentheyareexperienced.First,weneedtopassthetexttothefunction.Thetextwillbebrokendownandplaceeachtokenorwordinacategory.Afterwehavearrivedatthefinishofasentence,weclearupthewhitespaceswhichmayhaveremainedandafterwardswe’reallset,wehavegottenatriple.Forexampleinthestatement“BhubaneswariscategorisedasaTier-2city”itwillgiveatriplefocusingonthemainsubject(Bhubaneswar,categorised,Tier-2city). Belowwehavedefinedthecodetogettriplesthatcanbeusedtobuildknowledgegraphs. defgetSentences(text):     nlp=English()     nlp.add_pipe(nlp.create_pipe('sentencizer'))     document=nlp(text)     return[sent.string.strip()forsentindocument.sents] defprintToken(token):     print(token.text,"->",token.dep_) defappendChunk(original,chunk):     returnoriginal+''+chunk defisRelationCandidate(token):     deps=["ROOT","adj","attr","agent","amod"]     returnany(subsintoken.dep_forsubsindeps) defisConstructionCandidate(token):     deps=["compound","prep","conj","mod"]     returnany(subsintoken.dep_forsubsindeps) defprocessSubjectObjectPairs(tokens):     subject=''     object=''     relation=''     subjectConstruction=''     objectConstruction=''     fortokenintokens:         printToken(token)         if"punct"intoken.dep_:             continue         ifisRelationCandidate(token):             relation=appendChunk(relation,token.lemma_)         ifisConstructionCandidate(token):             ifsubjectConstruction:                 subjectConstruction=appendChunk(subjectConstruction,token.text)             ifobjectConstruction:                 objectConstruction=appendChunk(objectConstruction,token.text)         if"subj"intoken.dep_:             subject=appendChunk(subject,token.text)             subject=appendChunk(subjectConstruction,subject)             subjectConstruction=''         if"obj"intoken.dep_:             object=appendChunk(object,token.text)             object=appendChunk(objectConstruction,object)             objectConstruction=''     print(subject.strip(),",",relation.strip(),",",object.strip())     return(subject.strip(),relation.strip(),object.strip()) defprocessSentence(sentence):     tokens=nlp_model(sentence)     returnprocessSubjectObjectPairs(tokens) defprintGraph(triples):     G=nx.Graph()     fortripleintriples:         G.add_node(triple[0])         G.add_node(triple[1])         G.add_node(triple[2])         G.add_edge(triple[0],triple[1])         G.add_edge(triple[1],triple[2])     pos=nx.spring_layout(G)     plt.figure(figsize=(12,8))     nx.draw(G,pos,edge_color='black',width=1,linewidths=1,             node_size=500,node_color='skyblue',alpha=0.9,             labels={node:nodefornodeinG.nodes()})     plt.axis('off')     plt.show() WecanusepyplotlibrariestobuildtheKnowledgeGraph.Theabovecodeisfordisplayingthegraph. if__name__=="__main__":        text="BhubaneswaristhecapitalandlargestcityoftheIndianstateofOdisha.ThecityisboundedbytheDayaRiver"\             "tothesouthandtheKuakhaiRivertotheeast;theChandakaWildlifeSanctuary"\             "andNandankananZoolieinthewesternandnorthernpartsofBhubaneswar."\             "BhubaneswariscategorisedasaTier-2city."\             "BhubaneswarandCuttackareoftenreferredtoasthe'twincitiesofOdisha'."\             "Thecityhasapopulationof1163000."         sentences=getSentences(text) nlp_model=spacy.load('en_core_web_sm') triples=[] print(text) forsentenceinsentences: triples.append(processSentence(sentence)) printGraph(triples) Whendataprocessingisbeingdone,theSpacylibraryattachesatagtoeverywordsothatweknowawordiseitherasubjectoranobject.Givenbelowisanexample. KnowledgeGraph FinalThoughts Inthisarticle,wefiguredouthowtoextricatedatafromagivenbookastriplesandfabricateaninformationdiagramfromit.Further,wecanexplorethisfieldofdataextractioninmoredetailstolearnextractionofmoreperplexingconnections.Hopethisarticleisusefultoyou. MoreGreatAIMStories BehindClaraParabricks:NVIDIA’sFrameworkForGenomicResearch IntroductionGuideToFP-TreeAlgorithm What’sMoreSecure:BlockchainOrDatabase? Salesforce’sDeepakPargaonkarOnCloudAccelerationInIndia GuideToImageSuper-ResolutionByESRGAN BeginnersGuideToTruncatedSVDForDimensionalityReduction Adataanalystwithexpertiseinstatisticalanalysis,datavisualizationreadytoservetheindustryusingvariousanalyticalplatforms.Ilookforwardtohavingin-depthknowledgeofmachinelearninganddatascience.Outsidework,youcanfindmeasafun-lovingpersonwithhobbiessuchassportsandmusic. OurUpcomingEvents Masterclass,VirtualHowtopowerapplicationsforthedata-driveneconomy20thJul Register Masterclass,VirtualBuildingacareerinArtificialIntelligence23rdJul Register Conference,in-person(Bangalore)Cypher202221-23rdSep Register Conference,VirtualDeepLearningDevCon202229thOct Register Conference,in-person(Bangalore)MachineLearningDevelopersSummit(MLDS)202319-20thJan Register 3WaystoJoinourCommunity DiscordServerStayConnectedwithalargerecosystemofdatascienceandMLProfessionals JoinDiscordCommunity TelegramChannelDiscoverspecialoffers,topstories,upcomingevents,andmore. JoinTelegram Subscribetoournewsletter GetthelatestupdatesfromAIM Email Subscribe MOSTPOPULAR BuildyourfirstmachinelearningmodelwithBigQueryML SourabhMehta ThisarticlehascoveredthestepstocreateamachinelearningmodelusingBigQueryML. It’shightimeSpaceXenteredIndia AmitRajaNaik InJune2020,theIndiangovernmentlaunchedIN-SPACetoincreaseprivateparticipationinthespacesector. AIcouldendthestockimageindustryasweknowit PritamBordoloi TheglobalstockimagesandvideosmarketwasvaluedatUSD4.68billionin2021. ThefreeWeb3coursesonYoutube TasmiaAnsari Here’sacuratedlistofaffordableandfreeonlineWeb3courses. AredatascientistsmovingawayfromJupyterNotebooks? PoulomiChatterjee InMarchthisyear,AquaSecurity’sTeamNautilusdiscoveredaPython-basedransomwarethatwasusingJupyterNotebookstoaccessandtargetotherenvironments. AIMLaunchesTheCampusAmbassadorProgram Dr.VaibhavKumar CampusAmbassadorProgram2022hasbeendesignedtonurturethedatasciencetalentattheuniversitylevel&boostthedatascienceecosystemoncampuses.#AIMCAP EnhancingDataGovernanceCapabilitieswithDatabricks’DeltaLiveTables JasonYip Theseconceptssoundstraightforward,butwhyexactlyareorganizationsnotdoingdatagovernanceornotdoingitcorrectly?  Whyisthefairnessinrecommendersystemsrequired? SourabhMehta Thesystemisconsideredfairwhentherecommendationisunbiasedtowardsanygrouporindividualsconsumerorevenproviders. IsHermitthenewPegasus? SriKrishna AccordingtoWikileaks,RCSLabisalsoaknownbusinessassociateoftheItalianspywarevendorMementoLabs. DatabricksgivesawayitsdataLakeproductforfree,butwhy? ZinniaBanerjee DeltaLakeisanopenformatstoragelayerthatbringsreliabilitytodatalakes. Ourmissionistobringaboutbetter-informedandmoreconsciousdecisionsabouttechnologythroughauthoritative,influential,andtrustworthyjournalism. ShapeTheFutureofTech CONTACTUS⟶ ©AnalyticsIndiaMagazinePvtLtd2022 Termsofuse PrivacyPolicy Copyright