I-Spring AI isebenza: Ukwakha Izinhlelo Zokusebenza Ze-AI Zangempela nge-Spring Boot

Ikhaya » Python » I-Spring AI isebenza: Ukwakha Izinhlelo Zokusebenza Ze-AI Zangempela nge-Spring Boot

I-Spring AI in Action isiba yindawo okufanele isetshenziswe kakhulu kubathuthukisi be-Java ne-Spring Boot abafuna ukuletha i-AI yokukhiqiza yesimanje kumaphrojekthi abo ansuku zonke ngaphandle kokushiya i-JVM stack. Esikhundleni sokukuphoqa ukuthi ungene ezindabeni ze-Python noma ekusebenziseni amathuluzi angacacile, incwadi kanye nohlaka kusebenza ngokubambisana ukuze uqhubeke nokubhala ikhodi ku-Java noma ku-Kotlin ngenkathi usahlanganisa ama-Large Language Models (LLMs) anamandla, i-Retrieval Augmented Generation (RAG), ama-ejenti, amathuluzi kanye nezici ze-multimodal.

Okwenza le ndlela yokusebenza ibe mnandi kangaka ukuhlanganiswa kohlaka olulungele ukukhiqizwa (i-Spring AI) kanye nomhlahlandlela osebenza kahle kakhulu, oqhutshwa yizibonelo (i-Spring AI in Action nguCraig Walls). Ndawonye zibonisa indlela yokuxhumanisa amamodeli e-AI, izizindalwazi ze-vector, inkumbulo yengxoxo namathuluzi okuhlola kuzinhlelo zokusebenza ezijwayelekile ze-Spring Boot kusetshenziswa ama-POJO alula, ukucushwa okuzenzakalelayo kanye ne-API ehlanzekile, ephathekayo efihla ubunzima obuningi obuqondene nomhlinzeki.

Kuyini i-Spring AI nokuthi kungani kubalulekile kubathuthukisi beJava

I-Spring AI iwuhlaka lohlelo lokusebenza olwenzelwe ukuletha izimiso zakudala zaseNtwasahlobo—ukuthwaleka, ukwakheka kwe-modular kanye nomklamo ogxile ku-POJO—ezweni lobunjiniyela be-AI. Enhliziyweni yayo, i-Spring AI igxile ekuxazululeni inkinga enzima kakhulu esebenzayo ku-AI yebhizinisi: ukuxhumanisa i-AI yenhlangano yakho idatha futhi Ama-API ngesimanje Amamodeli e-AI ngendlela ehlala njalo, eqaphelekayo futhi elula ukuyishintsha ngokuhamba kwesikhathi.

Esikhundleni sokukuvalela kumthengisi oyedwa we-LLM, i-Spring AI ifinyeza ngaphezu kwabaningi babahlinzeki abakhulu. Ngaphandle kwebhokisi, ungakhuluma namamodeli avela ku-OpenAI, I-Azure OpenAI, Anthropic, Amazon Bedrock, Google, MistralAI ngisho namamodeli endawo asetshenziswa nge-Ollama. Imodeli efanayo yokuhlela isekela kokubili izimpendulo ezihambisanayo kanye nezokusakaza, futhi usakwazi ukufinyelela amakhono athile omhlinzeki uma uwadinga ngempela.

Enye insika ebalulekile ye-Spring AI ukusekela kwayo okuqinile komkhiqizo ohleliwe. Esikhundleni sokuhlaziya umbhalo ongakahlelwa ngesandla, ungahlela izimpendulo zamamodeli ngqo kumakilasi namarekhodi e-Java, uguqule ulimi lwemvelo olungcolile lube ama-POJO ahlanzekile. Lokhu kubalulekile uma wakha ama-ejenti, amathuluzi noma imisebenzi yokusebenza okumele icabange ngedatha ebikezelwayo esikhundleni sombhalo ongahlelekile.

I-Spring AI iphinde ihlangane ngokujulile nezizindalwazi ze-vector ukuze ukwazi ukusebenzisa i-Retrieval Augmented Generation ngaphandle kokuvuselela isondo. Isekela abahlinzeki abanjengo-Apache Cassandra, i-Azure Vector Search, i-Chroma, i-Milvus, i-MongoDB Atlas, i-Neo4j, Ukuhlanganiswa kwe-Oracle, I-PostgreSQL ene-PGVector, i-Pinecone, i-Qdrant, i-Redis kanye ne-Weaviate. I-API ye-Vector Store ephathekayo kanye nolimi lokuhlunga lwe-metadata olufana ne-SQL likuvumela ukuthi ushintshe ama-vector backend ngokushintsha okuncane kwekhodi.

Ngaphezu kwakho konke lokho, i-Spring AI ifika namathuluzi okubona, amapayipi okungenisa amadokhumenti, ukuhlolwa kwamamodeli kanye namaphethini e-AI akhiqizayo. Uthola ikhono lokukhuluma kahle ChatClient fana no WebClient/RestClient, Abeluleki bamaphethini avamile e-AI njenge-RAG kanye nenkumbulo yengxoxo, ukucushwa okuzenzakalelayo ngeziqalisi ze-Spring Boot, kanye nezinsiza zokuqapha ukusetshenziswa kwamathokheni kanye nokuthola ukubona izinto ezingekho.

Ngaphakathi "kwe-Spring AI in Action": kusukela ku-Hello AI World kuya kuma-ejenti aphelele

I-"Spring AI in Action" kaCraig Walls iyisiqondiso esisebenzayo nesisebenzayo esikukhombisa ukuthi ungawasebenzisa kanjani wonke lawa makhono e-Spring AI ezinhlelweni zangempela. Le ncwadi ihloselwe ngqo abathuthukisi be-Spring futhi icabanga ukuthi usuvele uyazi i-Spring Boot, kodwa ayidingi ulwazi lwangaphambilini lwe-AI yokukhiqiza; akudingeki ube usosayensi wedatha noma "uchwepheshe we-AI" ukuze ulandele.

Uhambo olusencwadini luqala ngesibonelo esilula esithi “Hello AI World” futhi kancane kancane luveza amasu athuthukile njengoba uzizwa ukhululekile. Uqala ngokuxhuma ucingo oluyisisekelo lwe-LLM ngaphakathi kohlelo lokusebenza lwe-Spring Boot, bese udlulela ekudaleni izifinyezo zombhalo, wakhe abasizi abahlala ngaphakathi kwezinsizakalo zakho zewebhu noma ze-backend ezikhona, futhi wakhe izimpendulo ukuze izimpendulo zibe usizo kakhulu futhi zibikezelwe.

Njengoba uqhubeka, okuqukethwe kugxila ku-RAG, izitolo ze-vector kanye nezimo eziningi lapho amamodeli asebenza khona nombhalo kanye nezithombe. Ufunda indlela yokubuza imibuzo mayelana namadokhumenti ayimfihlo imodeli engakaze iqeqeshwe kuwo, indlela yokuguqula izithombe zibe umbhalo kanye nokuphikisana nalokho, kanye nendlela yokusekela izimpendulo ze-LLM kudatha yakho ukuze ziyeke ukucabanga ngezinto ezingekho lapho zibhekene nemibuzo ethile yesizinda.

Ingxenye yesibili yencwadi iphakamisa izinga ngokuhlola izinto ezisetshenziswayo, ukusetshenziswa kwamathuluzi, inkulumo, kanye nokubonakala. Lapha ubona ukuthi ungakha kanjani ama-ejenti e-AI anganquma ukuthi amathuluzi noma ama-API azobizwa nini, ukuthi angahambisa kanjani imisebenzi iye ezicelweni ezikhethekile, ukuthi angalandelela kanjani okwenzekayo ngama-metrics kanye nokulandelelwa, nokuthi ungagcina kanjani uhlelo lwakho luphephile ngokuhlola kanye nokuvikela okuqukethwe okukhiqizwe.

Kuyo yonke incwadi, uCraig Walls ugcina uphawu lwakhe lokuhweba, isitayela esiqhutshwa yisibonelo, ehlala egxile “ekwenzeni izinto” esikhundleni sokukucwilisa emcabangweni. Izahluko zigcwele izingcezu ezisebenzayo kanye nezimo ezingokoqobo: ama-chatbot azi idatha yakho ngempela, abasizi abafakwe emisebenzini yebhizinisi, kanye nama-ejenti ahlukanisa imisebenzi eyinkimbinkimbi ibe yizicucu ezincane nezilawulekayo.

Izihloko ezibalulekile kanye nesakhiwo sencwadi

Ithebula lokuqukethwe kwe-“Spring AI in Action” linikeza isithombe esicacile sobubanzi balokho ozokwakha. Kusukela kumabhlogo okwakha ayisisekelo kuya kumaphethini athuthukile, isahluko ngasinye sigxila endaweni ethile yokuhlanganiswa kwe-AI neNtwasahlobo:

• Ukuqala nge-Spring AI: ukuqala kabusha iphrojekthi, ukulungiselela abahlinzeki, ukuthumela izixwayiso zakho zokuqala.
• Ukuhlola izimpendulo ezikhiqizwe: ukulinganisa ikhwalithi, ukuthola izinkinga, nokuvikela okuqukethwe kwekhwalithi ephansi noma okungaqondakali.
• Ukuthumela izimemezelo zokukhiqiza: ukuklama izikhuthazo, ukusebenzisa amathempulethi nokulawula ukuziphatha kwemodeli.
• Ukukhuluma namadokhumenti akho: ukusebenzisa i-RAG ukuze ama-LLM akwazi ukuphendula imibuzo mayelana nedatha yangasese engaqeqeshiwe.
• Ukunika amandla inkumbulo yengxoxo: ukugcina umongo wengxoxo ephendula ama-multi-turn usebenzisa abeluleki bememori be-Spring AI.
• Ukukhiqiza okuqhutshwa ngamathuluzi okusebenza: ukuvumela amamodeli ukuthi abize imisebenzi namathuluzi aseceleni kweklayenti lapho edinga idatha entsha noma yangaphandle.
• Ukusebenzisa i-Model Context Protocol (MCP): ukuphatha umongo ocebile kanye nokusebenzisana namathuluzi nemithombo yedatha.
• Ukukhiqiza ngezwi nezithombe: ukwamukela amakhono amaningi okukhuluma nezithombe.
• Ukubuka imisebenzi ye-AI: ukwengeza ukubonwa nokuqapha emipayipini yakho ye-AI.
• Ukuvikela i-AI yokukhiqiza: ukusebenzisa izithiyo zokuvikela, izihlungi zokuqukethwe kanye nezinye izindlela zokuvikela.
• Ukusebenzisa amaphethini e-AI akhiqizayo: ukuthwebula amaphethini angasetshenziswa kabusha emisebenzini ye-AI.
• Ama-ejenti aqashayo: ukwakha izinhlelo ze-ejensi ezingahlela, ziqondise futhi zithuthukise umsebenzi.

Ukubuyekezwa okuvela kubantu abahlonishwayo emiphakathini yaseNtwasahlobo naseJava kugqamisa ukuthi ulwazi lufinyeleleka kalula futhi luwusizo kangakanani. Ababhali nababuyekezi besandulelo badumisa le ncwadi ngezincazelo ezicacile, ama-demos abanzi kanye nokujula "kwengcebo" ngobuchwepheshe obusha, okugcizelela ukuthi ihlala isekelwe ekuthuthukisweni komhlaba wangempela kunokuba ibe yindaba ecashile yezemfundo.

Uma uthenga uhlelo oluphrintiwe ku-Manning uthola ne-eBook yamahhala (i-PDF noma i-ePub) kanye nokufinyelela inguqulo yabo ye-liveBook eku-inthanethi. Ipulatifomu ye-liveBook ngokwayo ifaka umsizi we-AI okwazi ukuphendula imibuzo yakho ngezilimi eziningi, ukuze ukwazi ukuhlola izibonelo, useshe umbhalo futhi ucacise izihloko ngenkathi ufunda.

Izici eziyinhloko ze-AI yasentwasahlobo zezinhlelo zokusebenza ze-AI zebanga lebhizinisi

Ngaphandle kwencwadi, uhlaka lwe-Spring AI luveza isethi yezici eziphelele ezenzelwe izinhlelo zokusebenza ze-AI zezinga lokukhiqiza. Akukhona nje ukubiza i-LLM; kumayelana nokwakha izinhlelo eziphelele eziphephile, ezibonakala kalula, ezivivinywayo futhi eziphathekayo kubahlinzeki nasezindaweni ezizungezile.

Izinga elifanayo lokuguquguquka lifinyelela ezitolo ze-vector. Ngokusekelwa kwe-Apache Cassandra, i-Azure Vector Search, i-Chroma, i-Milvus, i-MongoDB Atlas, i-Neo4j, i-Oracle, i-PostgreSQL/PGVector, i-Pinecone, i-Qdrant, i-Redis, i-Weaviate nabanye, ungasebenzisa i-RAG kanye nokusesha kwe-semantic ngaphandle kokuxhuma uhlelo lwakho lokusebenza kusixazululo esisodwa sesitoreji. I-API ephathekayo kanye nezihlungi ze-metadata ezivezayo kwenza kube lula ukusebenzisa imibuzo eyinkimbinkimbi yokufana.

Amathuluzi kanye nobizo lomsebenzi kuyizakhamuzi zesigaba sokuqala e-Spring AI. Amamodeli angacela ukwenziwa kwamathuluzi nemisebenzi eseceleni kweklayenti ukuze kutholakale idatha yesikhathi sangempela noma kuqalwe izenzo. Lokhu kuguqula i-LLM yakho isuke ekubeni yi-generator yombhalo engasebenzi ibe yingxenye esebenzayo engakwazi ukubuza ama-API, ukubiza izizindalwazi noma ukuhlela izinsizakalo ngamakholi omsebenzi athayishiwe.

Ukuqaphela kufakwe ohlakeni ukuze ubone ukuthi i-AI yakho yenzani ngaphansi kwe-hood. Ungaqoqa amamethrikhi ekusetshenzisweni kwamathokheni, amazinga okubambezeleka kanye namaphutha, ulandelele izingcingo ngohlelo lwakho futhi uhlobanise umsebenzi we-LLM nezinye izinsizakalo zakho ezincane. Lokhu kubalulekile lapho i-AI ishintsha kusuka ekuhlolweni iye emisebenzini ebalulekile yebhizinisi.

I-Spring AI ihlanganisa nohlaka lokungenisa amadokhumenti esitayela se-ETL lwemisebenzi yobunjiniyela bedatha. Kukusiza ukuthi ulayishe, uqoqe futhi ubhale amadokhumenti ezitolo ze-vector ukuze amapayipi akho e-RAG aqine futhi aphindeke, kunokuba abe yiqoqo lezikripthi ezingezona ezejwayelekile.

I-ChatClient, Abeluleki kanye namakhono okuxoxa

Ezingeni lokubhala ikhodi, iningi lokuxhumana kwe-AI yaseNtwasahlobo lijikeleza ku- ChatClient I-API, isikhombikubona esicacile esiphefumulelwe amaphethini ajwayelekile e-Spring WebClient kanye ne-RestClient. Uyakha futhi uthumela izixwayiso, uthola izimpendulo, usakaze amathokheni njengoba efika futhi uphatha amaphutha ngendlela ezwakala ingokwemvelo kubathuthukisi be-Spring.

Abeluleki bangenye into ebalulekile ehlanganisa amaphethini e-AI avamile okukhiqiza. Ziguqula idatha eya futhi iphuma kuma-LLM, zibeke ungqimba ekuziphatheni okufana ne-RAG noma inkumbulo, futhi zinikeza ukuphatheka kalula kuwo wonke amamodeli kanye nezimo zokusetshenziswa. Esikhundleni sokuxhuma ngesandla yonke into noma umongo, uxhuma i-Advisors ukuze uthole ukuziphatha okuqinile nge-boilerplate encane.

Inkumbulo yengxoxo iphathwa ngabeluleki bememori yengxoxo abakhethekile abaphatha ingxoxo yokujika okuningi. Njengoba ama-LLM ngokwawo engenabo umbuso futhi “akhohlwa” izikhathi ezedlule, laba beluleki balandelela umlando wengxoxo futhi banikeze izingcezu ezifanele zomongo emuva kulokho okushiwoyo. Ungakhetha phakathi kwamasu ahlukene futhi usebenzise ngisho nenkumbulo eqhubekayo, yesikhathi eside ngezindlela ezisekelwe ku-vector.

Ukuhlanganiswa kwememori yengxoxo kanye ne-RAG Advisors kukuvumela ukuthi wakhe abasizi abangakhuluma namadokhumenti akho ngokushintshana okuningi. Umsebenzisi angabuza ukulandelwa, alungise imibuzo yakhe futhi abhekisele ezingxenyeni zangaphambilini zengxoxo, kuyilapho i-Spring AI ithola ngokuzenzakalelayo futhi ifake izingcezu zedokhumenti ezifanele kakhulu kuzo zonke izicelo.

Amathempulethi e-Prompt enza kube lula ukuwakhipha futhi uwasebenzise kabusha. Uchaza amathempulethi ajwayelekile amukela amapharamitha, afaka imiyalelo eyengeziwe futhi ucacise ifomethi yokuphuma oyifunayo (isibonelo i-JSON ehambisana ngqo nezinto ze-Java). Ngaphambi kokuba kuthunyelwe i-prompt, i-Spring AI igcwalisa izikhala, isebenzise umongo futhi iqinisekise ukuthi imiyalelo icacile kumodeli.

I-RAG, ukunciphisa imibono engekho kanye nabasizi abaqaphela amadokhumenti

I-Retrieval Augmented Generation (RAG) ingenye yamaphethini abaluleke kakhulu ahlanganiswe yilolu hlaka kanye nencwadi. Kuxazulula umkhawulo obalulekile wama-LLM angaguquki: azi kuphela lokho aqeqeshwe kukho, okusho ukuthi awakwazi ukubona amadokhumenti akho angaphakathi, idatha yamakhasimende noma ulwazi oluyimfihlo ngokuzenzakalelayo.

Nge-RAG, uhlelo lwakho lokusebenza luqala ngokuthola isethi encane yamadokhumenti afana ngokomqondo nombuzo womsebenzisi bese luwaphakela kumodeli njengomongo. I-Spring AI ifingqa okuningi kwalo msebenzi, ihlanganisa nezitolo eziningi ze-vector futhi inikeze i-API yokubuza ngokufana, ukuhlunga nge-metadata bese ulungisa indlela ohlukanisa futhi ufaka ngayo okuqukethwe kwakho.

I-RAG esetshenziswe kahle inciphisa kakhulu ukubona izinto ezingekho. Esikhundleni sokuqagela uma ingenalo ulwazi noma iqeqeshwe ngokweqile kudatha ejwayelekile ye-inthanethi, imodeli iqondiswa ezingcezu ezisezingeni eliphezulu, eziqondene nesizinda. Incwadi idlula ezimweni zokusebenzisa "ingxoxo namadokhumenti akho" kanye "nemibuzo nezimpendulo ngamadokhumenti akho" ezibonisa leli phethini kusukela ekuqaleni kuze kube sekupheleni.

Ngokusebenzisa QuestionAnswerAdvisor futhi ChatClient, ungaqhuba ukugeleza kwe-RAG yonke ngokucacile noma uvumele uMluleki ahlele ukushumeka, ukubuyisa kanye nokufakwa komongo kuwe. Lokho kukunika ukuguquguquka: qala ngendlela elula yokuhamba ngokushesha, bese wehlisa izinga uma udinga ukuziphatha ngokwezifiso noma ukulungiswa okujulile.

Ngenxa yokuthi i-Spring AI isekela izimpendulo zokusakaza, lezo zimpendulo eziqaphela amadokhumenti zingasakazwa ku-UI njengoba zikhiqizwa. Lokhu kulingisa ukuthayipha komuntu ngesikhathi sangempela futhi kunikeza ulwazi olungcono lomsebenzisi, ikakhulukazi uma izimpendulo zinde noma ukubambezeleka kwemodeli kuphezulu.

Amaphethini e-Ejenti aphefumulelwe ucwaningo lwe-Anthropic

I-Spring AI iphinde isebenzise isethi yamaphethini e-ejenti aphefumulelwe ucwaningo luka-Anthropic mayelana nokwakha ama-ejenti e-LLM asebenzayo. Kugcizelelwa ubulula kanye nokuhlanganiswa kunokuba kube yizinhlaka ze-ejenti ezisindayo nezingabonakali, ezihambisana kahle nezidingo zebhizinisi zezinhlelo ezigcinwayo nezihlolwayo.

Iphethini yokuqala, i-Chain Workflow, ihlukanisa imisebenzi emikhulu ibe uchungechunge lwezinyathelo ezincane, ezihlelekile. Isinyathelo ngasinye sisebenzisa i-prompt yaso, sisebenzisa umphumela wesinyathelo sangaphambilini futhi sikhiqize imiphumela ephakathi ethuthukisiwe. Ku-AI yaseNtwasahlobo, lokhu kubukeka njengokuphindaphinda phezu kwezicelo zesistimu kanye nokubiza ChatClient ngokuphindaphindiwe, ukudlulisa impendulo yangaphambilini njengengxenye yokufaka okulandelayo, ukudala ipayipi elicacile nelinwebekayo.

I-Parallelization Workflow imayelana nokusebenzisa izingcingo eziningi ze-LLM ngesikhathi esisodwa nokuhlanganisa imiphumela yazo. Ungayisebenzisela “ukuhlukanisa” (ukuhlukanisa umsebenzi ube yizicucu ezizimele) noma “ukuvota” (ukuba nezinhlobo eziningana zemisebenzi ezibhekana nomfutho ofanayo bese uhlanganisa imiphumela). Isibonelo, ungase ucele imodeli ukuthi ihlaziye umthelela wezinguquko zemakethe kumakhasimende, abasebenzi, abatshalizimali kanye nabaphakeli ngesikhathi esisodwa, bese uhlanganisa leyo mibono.

I-Routing Workflow yethula ukuthumela okuhlakaniphile kule ngxube. I-LLM iqala ngokuhlukanisa okufakiwe bese inquma ukuthi yimuphi umlayezo noma umphathi okhethekile okufanele awucubungule: imibuzo yokukhokha iya kumbuzo owodwa wochwepheshe, izinkinga zobuchwepheshe ziye komunye, imibuzo ejwayelekile kumsizi ojwayelekile. Ukuhamba komsebenzi kwe-AI yaseNtwasahlobo kuhlanganisa lo mqondo ndawonye ngokusebenzisa ChatClient kanye nemephu yemizila.

Izisebenzi ze-Orchestrator ziyisibonelo esithuthuke kakhulu esisagwema ukuzimela okungalawulwa. Imodeli "ye-orchestrator" ephakathi ihlukanisa umsebenzi oyinkimbinkimbi ube yimisebenzi engaphansi, bese izisebenzi ezikhethekile zibhekana naleyo misebenzi engaphansi, ngokuvamile ngesikhathi esifanayo. Uma izisebenzi seziqedile, imiphumela yazo ihlanganiswa ibe umphumela wokugcina. I-Spring AI inikeza izakhi zokwakha ukuze kuqaliswe le ndlela ngenkathi kugcinwa imithwalo yemfanelo icacile futhi ingabikezelwa.

Okokugcina, iphethini ye-Evaluator‑Optimizer isebenzisa amamodeli amabili abambisanayo. Imodeli eyodwa isebenza njengomkhiqizi ophakamisa izixazululo, kanti imodeli yesibili iziphatha njengomgxeki noma umhlaziyi, ihlola ikhambi ngokumelene nezindinganiso ezicacile futhi inikeze impendulo ngokuthuthukiswa. Lokhu kuqhubeka kuze kube yilapho umhloli enelisekile, okukhiqiza impendulo ecwengekile kanye nomzila wokuvela kwekhambi.

Imikhuba emihle kakhulu, ukwethembeka kanye nokuvela kwekusasa

Amaphethini nezici ku-Spring AI zihambisana nemikhuba emihle ecacile evela ocwaningweni luka-Anthropic kanye nokuhlangenwe nakho kokukhiqiza kwe-Spring ecosystem. Iseluleko esivamile ukuthi uqale ngomsebenzi olula kakhulu ongasebenza, bese ubeka ubunzima kuphela uma bufaka inani ngokusobala.

Ukuthembeka kufanele kube yinto ebaluleke kakhulu kunoma yiluphi uhlelo olusebenzisa i-LLM. Lokho kusho ukusebenzisa umkhiqizo ohleliwe ngendlela ephephile noma kuphi lapho kungenzeka khona, ukuqinisekisa izimpendulo, ukwengeza ukuphathwa kwamaphutha okunamandla kanye nokuzama kabusha, kanye nokufaka amapayipi akho ngezilinganiso namalogi. Uma kukhona okungahambi kahle, kufanele ukwazi ukuqonda ukuthi kungani futhi ukulungise ngokushesha.

Abathuthukisi bakhuthazwa ukuthi bahlole ngokucophelela ukuhwebelana kokubambezeleka nokunemba. Ukuhlanganisa izinyathelo eziningi noma ukwengeza ama-loop okuhlola kungathuthukisa kakhulu ikhwalithi kodwa kuzokwandisa nezikhathi zokuphendula kanye nokusetshenziswa kwamathokheni. Ukulinganisa kungasiza ekubuyiseni isivinini, kodwa kuphela uma imisebenzi izimele ngempela.

Umsebenzi wesikhathi esizayo ohlelweni lwe-AI lwaseNtwasahlobo uzojulisa amakhono azungeze ukwakheka kwamaphethini, amasu enkumbulo athuthukisiwe kanye nokuhlanganiswa kwamathuluzi. Ukubhala amaphethini amaningi—njengokufaka izintambo, ukuhambisa kanye nokuhlola—kukuvumela ukuthi wakhe ama-ejenti ahlakaniphile asaqondakala. Ukuphathwa kwememori okuthuthukisiwe kuhlola umongo oqhubekayo, amafasitela omongo asebenzayo kanye nokugcinwa kolwazi lwesikhathi eside.

Ukuhlanganiswa kwamathuluzi kanye ne-Model Context Protocol (MCP) kungenye indawo esebenzayo. Izixhumi ezijwayelekile zamathuluzi angaphandle kanye nephrothokholi ecebile yomongo wemodeli kusho ukuthi ama-ejenti angafinyelela ngokuphephile nangokuguquguqukayo kumasevisi akho, ama-API kanye nemithombo yedatha, konke ngaphansi kwe-tack yakho yokuphatha kanye nokubona.

I-AI yasentwasahlobo epulatifomu ebanzi: Izixazululo ze-Tanzu Gen AI

Ezinhlanganweni ezakha phezu kwe-VMware's Tanzu stack, i-Spring AI nayo isekela i-Tanzu Gen AI Solutions. I-Tanzu AI Server, enikwe amandla yi-Spring AI, inikeza indawo elungele ukukhiqizwa yokufaka izinhlelo zokusebenza ze-AI ku-Tanzu Platform enokuphepha kwezinga lebhizinisi, ukubusa kanye nokukhula.

Lokhu kuhlanganiswa kwenza kube lula ukufinyelela kumamodeli afana ne-Amazon Bedrock Nova ngokusebenzisa isikhombikubona esihlanganisiwe. Esikhundleni sokuthi iqembu ngalinye lixhumanise ukuxhumana kwalo kwemodeli, ipulatifomu ibeka izinga elifanele lokufinyelela, izinqubomgomo zokuphepha kanye namathuluzi okusebenza. I-Spring AI iphatha ukuphatheka kwemodeli, kuyilapho i-Tanzu inikeza ingqalasizinda eqinile, ukukala okuzenzakalelayo kanye nokubuka okulindelekile kusuka kupulatifomu yesimanje ye-Kubernetes.

Ngenxa yokuthi i-Spring AI inesibopho sokuqashelwa kwezinga lesicelo, amaqembu angahamba phakathi kwabahlinzeki noma amukele amamodeli amasha ngaphandle kokubhala kabusha i-logic yebhizinisi lawo. Lokho kuzivumelanisa nezimo kubalulekile endaweni ye-AI esheshayo lapho kuvela khona amamodeli amasha njalo futhi amanani noma amakhono angashintsha ngokushesha.

Izici zokuphepha kanye nokuphatha ku-Tanzu Gen AI Solutions zisonga lezi zinhlelo zokusebenza ze-AI ezilawulweni ezifanayo zebhizinisi ezisetshenziselwa ezinye izinsizakalo ezincane. Izinqubomgomo, ukulawulwa kokufinyelela, izindlela zokuhlola kanye namathuluzi okuthobela imithetho kudlulela ngokwemvelo emisebenzini ye-LLM, okwenza kube lula ukuqhuba amacala okusetshenziswa abucayi noma alawulwayo.

Zonke lezi zendlalelo—uhlaka, incwadi, amaphethini kanye neplatifomu—zihlangana zibheke emgomweni ofanayo: ukuvumela abathuthukisi baseNtwasahlobo ukwengeza izici ze-AI ezibiza kakhulu njengabasizi ababonakalayo, ukusesha okuhlakaniphile, ukufingqa umbhalo kanye nezincomo ngqo kuzinhlelo zokusebenza zeJava ngaphandle kokulahla ukuthembeka noma ukulawula. Njengoba i-Spring AI in Action ingumhlahlandlela wakho osebenzayo kanye ne-Spring AI ingumgogodla wakho wobunjiniyela, ungasuka ekuhlolweni uye ezinsizakalweni eziqinile ezisebenzisa i-AI ngenkathi uhlala ngaphakathi kwendawo yaseNtwasahlobo osuvele uyazi kahle.