Intangiriro
Ubwihindurize bwihuse bwa Artificial Intelligence (AI) hamwe no Kwiga Imashini (ML) byatumye habaho iterambere rya agentic AI, ubwoko bwa AI bushobora gufata ibyemezo no gufata ibyemezo wenyine. Iri koranabuhanga rifite ubushobozi bwo guhindura uburyo twegera ubwubatsi bwa software hamwe na automatike ya DevOps. Muri iki kiganiro, tuzasesengura guhuza ibikorwa bya AI, Ururimi runini (LLMs), hamwe nibikoresho bya orchestre nka LangChain na DSPy kugirango uhindure software yububiko, ibikorwa remezo, hamwe na DevOps.
Agentic AI ni iki?
Agentic AI bivuga ubwoko bwa AI bushobora gufata ibyemezo no gufata ibyemezo wenyine. Ibi bitandukanye na sisitemu gakondo ya AI, isanzwe igenewe gukora imirimo yihariye no gukurikiza amategeko yateganijwe mbere. Sisitemu ya Agentic AI, yashyizweho kugirango irusheho kwigenga no guhuza n'imiterere, ibemerera kwigira kubidukikije no gufata ibyemezo bishingiye ku ntego zabo n'intego zabo.
Uruhare rwa LLM muri Agentic AI
Ururimi runini rwindimi (LLMs) ni ubwoko bwa AI butozwa kumibare myinshi yamakuru kandi bushobora kubyara imvugo imeze nkabantu. LLMs nigice cyingenzi cya sisitemu ya AI ikora, kuko itanga ubushobozi bwo gusobanukirwa no kubyara ururimi rwabantu. Ibi bituma sisitemu ya AI ikora ivugana nabantu hamwe nubundi buryo muburyo busanzwe kandi bwimbitse.
Ibikoresho bya Orchestre: LangChain na DSPy
Ibikoresho bya Orchestre nka LangChain na DSPy byashizweho kugirango bicunge kandi bihuze imikoranire hagati ya sisitemu zitandukanye nibigize. Ibi bikoresho bitanga uburyo bwo gusobanura no gushyira mubikorwa ibikorwa, bishobora kubamo imirimo nko gutunganya amakuru, amahugurwa yicyitegererezo, no kohereza.
Ururimi
LangChain nigikoresho gifungura-orchestre igikoresho gitanga uburyo bworoshye kandi bwagutse bwo kubaka no gucunga ibikorwa. LangChain yagenewe gukorana nurwego runini rwa sisitemu n'ibigize, harimo LLMs, data base, na APIs.
DSPy
DSPy nigikoresho cya Python gishingiye kumurongo utanga uburyo bworoshye kandi bwimbitse bwo gusobanura no gukora akazi. DSPy yagenewe gukorana nurwego runini rwa sisitemu nibigize, harimo LLMs, data base, na APIs.
Gutangiza software yububiko hamwe na DevOps inzira
Agentic AI, LLMs, hamwe nibikoresho bya orchestre nka LangChain na DSPy birashobora gukoreshwa mugutangiza ibintu byinshi byububiko bwa software hamwe nibikorwa bya DevOps. Ingero zimwe zirimo:
** **
- ** Gukoresha byikora **: Sisitemu ya Agentic AI irashobora gukoreshwa mugutangiza gahunda, harimo kohereza code, iboneza, nibindi bikoresho.
- ** Gukurikirana byikora **: Sisitemu ya Agentic AI irashobora gukoreshwa mugukoresha mu buryo bwikora, harimo kugenzura imikorere ya sisitemu, umutekano, nibindi bipimo.
Gushyira mubikorwa
Muri iki gice, tuzatanga urugero rworoshye rwuburyo bwo gukoresha LangChain na DSPy kugirango uhindure software yububiko bwa software hamwe na DevOps.
Urugero: Gusubiramo Kode Yisubiramo
Muriyi ngero, tuzakoresha LangChain na DSPy kugirango dusubiremo kode isubiramo. Tuzasobanura ibikorwa byakazi birimo intambwe zikurikira:
- ** Isesengura rya kode **: Urupapuro rwakazi ruzakoresha LLM mu gusesengura kode no kumenya ibibazo bishobora kuvuka.
- ** Gusubiramo kode **: Urupapuro rwakazi ruzakoresha sisitemu ya AI yo gusuzuma kode no kumenya ibibazo bishobora kuvuka.
- ** Kohereza kode **: Urupapuro rwakazi ruzakoresha igikoresho cyo kohereza kode kubidukikije.
Dore urugero rwuburyo bwo gushyira mubikorwa iki gikorwa ukoresheje LangChain na DSPy: `` python gutumiza mu mahanga kuva muri langchain itumiza LLM
Sobanura LLM
llm = LLM ("inyandiko-davinci-003")
Sobanura sisitemu ya AI
ai_system = langchain.AgenticAI ()
Sobanura akazi
akazi = langchain.Umurimo ( intambwe = [ Ururimi. Intambwe ( izina = "Isesengura rya Kode", imikorere = llm.analyze_code, inyongeramusaruro = {"code": "def muraho_isi (): icapa ('Mwaramutse Isi')"}, ), Ururimi. Intambwe ( izina = "Gusubiramo Kode", imikorere = ai_system.review_code, inyongeramusaruro = {"code": "def muraho_isi (): icapa ('Mwaramutse Isi')"}, ), Ururimi. Intambwe ( izina = "Kode yoherejwe", imikorere = langchain.deploy_code, inyongeramusaruro = {"code": "def muraho_isi (): icapa ('Mwaramutse Isi')"}, ), ] )
Kora akazi
akazi.kora () `` ` Iyi code isobanura urujya n'uruza rurimo intambwe eshatu: gusesengura kode, gusubiramo kode, no kohereza kode. Urupapuro rwakazi rukoresha LLM kugirango isesengure kode, sisitemu ya AI ikora kugirango isuzume kode, nigikoresho cyo kohereza kugirango kode ikorwe.
Hejuru 5-10 Amabwiriza Yakoreshejwe Byinshi / Uduce
Dore top 5-10 zikoreshwa cyane / ibisobanuro kuri LangChain na DSPy:
langchain.Umurimo wo gukora (intambwe = [...]): Sobanura urujya n'uruza rw'intambwe nyinshi.langchain. Intambwe (izina =" ... ", imikorere = ..., inyongeramusaruro = {...}): Sobanura intambwe imwe mubikorwa.Llm.analyze_code (inyongeramusaruro = {" code ": ...}): Koresha LLM gusesengura kode. 4 ..langchain.deploy_code (inyongeramusaruro = {" code ": ...}): Kohereza code mubidukikije.langchain.AgenticAI (): Kora urugero rwa sisitemu ya AI ikora.langchain.LLM (): Kora urugero rwa LLM.langchain.deploy_code (): Kohereza code mubidukikije.langchain.execute_workflow (): Kora akazi.langchain.get_workflow_status (): Kubona imiterere yumurimo.
Kugaragara
Dore urugero rwuburyo bwo kwiyumvisha ibikorwa byakazi ukoresheje Chart.js:
json { "ubwoko": "akabari", "amakuru": { "ibirango": ["Isesengura rya Kode", "Isubiramo rya Kode", "Kohereza Kode"], "imibare": [ { "label": "Intambwe Yakazi", "amakuru": [10, 20, 30], "backgroundColor": [ "rgba (255, 99, 132, 0.2)", "rgba (54, 162, 235, 0.2)", "rgba (255, 206, 86, 0.2)" ], "umupaka w'amabara": [ "rgba (255, 99, 132, 1)", "rgba (54, 162, 235, 1)", "rgba (255, 206, 86, 1)" ], "Ubugari bwumupaka": 1 } ] }, "amahitamo": {} } `
Iyi kode isobanura imbonerahamwe yumurongo ifite ibirango bitatu: Isesengura rya Kode, Isubiramo rya Kode, hamwe no kohereza kode. Imbonerahamwe yerekana umubare wintambwe muri buri cyiciro cyakazi.
Umwanzuro
Agentic AI, LLMs, hamwe nibikoresho bya orchestre nka LangChain na DSPy bifite ubushobozi bwo guhindura uburyo twegera ubwubatsi bwa software hamwe na automatike ya DevOps. Muguhindura imirimo nko gusubiramo kode, kugerageza, no kohereza, turashobora kunoza imikorere nibikorwa byiterambere ryiterambere. Muri iki kiganiro, twasuzumye uburyo bwo guhuza ikoranabuhanga kandi dutanga urugero rworoshye rwuburyo bwo kuzikoresha mugutangiza software yububiko bwa software hamwe na DevOps.