Ukusebenzisa i-Agentic AI yeSoftware Architecture kunye neDevOps Automation

Intshayelelo

Ukuguquka ngokukhawuleza kwe-Artificial Intelligence (AI) kunye ne-Machine Learning (ML) kuye kwakhokelela ekuphuhlisweni kwe-AI ye-agent, uhlobo lwe-AI olunokuthi luthathe amanyathelo kwaye luzenzele izigqibo ngokwalo. Le teknoloji inamandla okuguqula indlela esisondela ngayo kuyilo lwesoftware kunye ne-DevOps automation. Kule nqaku, siza kuphonononga ukuhlanganiswa kwe-AI ye-agent, iiModeli zoLwimi olukhulu (LLMs), kunye nezixhobo zomculo ezifana neLangChain kunye ne-DSPy ukwenza ngokuzenzekelayo i-architecture ye-software, iziseko zophuhliso, kunye neenkqubo ze-DevOps.

Yintoni iAgent AI?

I-Agent AI ibhekisela kuhlobo lwe-AI olunokuthi luthathe amanyathelo kwaye luzenzele izigqibo ngokwalo. Oku kuchasene neenkqubo ze-AI zemveli, ezilungiselelwe ukwenza imisebenzi ethile kwaye zilandele uluhlu lwemithetho echazwe kwangaphambili. Iinkqubo ze-Agent ze-AI, kwelinye icala, zenzelwe ukuzimela ngakumbi kwaye zilungelelanise, zibavumela ukuba bafunde kwindawo yabo kwaye benze izigqibo ezisekelwe kwiinjongo zabo kunye neenjongo zabo.

Indima yeeLLMs kwiAgentic AI

IiModeli zoLwimi olukhulu (LLMs) luhlobo lwe-AI oluqeqeshwe kumthamo omkhulu wedatha yokubhaliweyo kwaye ekwaziyo ukuvelisa ulwimi olufana nomntu. Ii-LLMs ziyinxalenye ephambili yeenkqubo ze-AI, njengoko zibonelela ngesakhono sokuqonda nokuvelisa ulwimi lwabantu. Oku kuvumela iinkqubo ze-AI ukuba zinxibelelane nabantu kunye nezinye iinkqubo ngendlela yendalo kunye nengqiqo.

Izixhobo zeOkhestra: LangChain kunye neDSPy

Izixhobo ze-Orchestration ezifana neLangChain kunye ne-DSPy zenzelwe ukulawula nokulungelelanisa ukusebenzisana phakathi kweenkqubo ezahlukeneyo kunye namacandelo. Ezi zixhobo zibonelela ngendlela yokuchaza kunye nokuqhuba ukuhamba komsebenzi, okungabandakanya imisebenzi efana nokulungiswa kwedatha, uqeqesho lomzekelo, kunye nokuthunyelwa.

LangChain

I-LangChain sisixhobo se-ochestration esivulelekileyo esibonelela ngesakhelo esiguquguqukayo kunye nesandisiweyo sokwakha nokulawula ukuhamba komsebenzi. I-LangChain yenzelwe ukusebenza kunye noluhlu olubanzi lweenkqubo kunye namacandelo, kuquka i-LLM, i-database, kunye ne-APIs.

DSPy

I-DSPy sisixhobo se-orchestration esekwe kwiPython esinika indlela elula necacileyo yokuchaza kunye nokuqhuba ukuhamba komsebenzi. I-DSPy yenzelwe ukusebenza ngoluhlu olubanzi lweenkqubo kunye namacandelo, kuquka ii-LLMs, ii-database, kunye nee-APIs.

Ukuzenzekela koYilo lweSoftware kunye neeNkqubo zeDevOps

I-Agentic AI, ii-LLMs, kunye nezixhobo zomculo ezifana neLangChain kunye ne-DSPy zingasetyenziselwa ukuzenzekelayo uluhlu olubanzi lwe-architecture ye-software kunye neenkqubo ze-DevOps. Eminye imizekelo ibandakanya:

Uphononongo lwekhowudi oluzenzekelayo: Iinkqubo ze-Agent ze-AI zingasetyenziselwa ukuphonononga ikhowudi kunye nokuchonga imiba enokubakho, efana neebhugi kunye nobuthathaka bokhuseleko.
Uvavanyo oluzenzekelayo: Iinkqubo ze-Agentic AI zingasetyenziselwa ukuvavanywa ngokuzenzekelayo, kubandakanywa uvavanyo lweyunithi, uvavanyo lokudibanisa, kunye novavanyo lokuphela.
** Ukuthunyelwa ngokuzenzekelayo **: Iinkqubo ze-Agent ze-AI zingasetyenziselwa ukuthunyelwa ngokuzenzekelayo, kubandakanywa ukuthunyelwa kwekhowudi, ukucwangciswa, kunye nezinye izixhobo.
Ukubeka iliso okuzenzekelayo: Iisistim ze-Agenttic AI zingasetyenziselwa ukubeka iliso ngokuzenzekelayo, kubandakanywa ukubeka esweni ukusebenza kwenkqubo, ukhuseleko, kunye nezinye iimethrikhi.

Uphumezo olusebenzayo

Kweli candelo, siya kunika umzekelo olula wendlela yokusebenzisa iLangChain kunye ne-DSPy ukwenza ngokuzenzekelayo i-architecture yesofthiwe kunye nenkqubo ye-DevOps.

Umzekelo: Uphononongo lweKhowudi oluzenzekelayo

Kulo mzekelo, siya kusebenzisa i-LangChain kunye ne-DSPy ukuhlaziya ikhowudi ngokuzenzekelayo. Siza kuchaza ukuhamba komsebenzi okubandakanya la manyathelo alandelayo:

Uhlalutyo lwekhowudi: Ukuhamba komsebenzi kuya kusebenzisa i-LLM ukuhlalutya ikhowudi kunye nokuchonga imiba enokubakho.
** Uphononongo lwekhowudi **: Ukuhamba komsebenzi kuya kusebenzisa inkqubo ye-AI ye-agenti ukuphonononga ikhowudi kunye nokuchonga imiba enokubakho.
** Ukuhanjiswa kwekhowudi **: Ukuhamba komsebenzi kuya kusebenzisa isixhobo sokuhambisa ukuhambisa ikhowudi kwindawo yokuvelisa.

Nanku umzekelo wendlela yokuphumeza oku kuhamba komsebenzi usebenzisa iLangChain kunye neDSPy:

ngenisa i-langchain
ukusuka kwi-LLM yokungenisa i-langchain

# Chaza iLLM
llm = LLM("text-davinci-003")

# Chaza inkqubo ye-AI yearhente
ai_system = langchain.AgenticAI ()

# Chaza indlela yokusebenza
ukuhamba komsebenzi = langchain.Ukuhamba komsebenzi(
    amanyathelo=[
        langchain.Step(
            igama = "Uhlalutyo lwekhowudi",
            func=llm.analyze_code,
            inputs={"code": "def hello_world(): print('Molo Hlabathi')"},
        ),
        langchain.Step(
            igama="Uphononongo lwekhowudi",
            func=ai_system.review_code,
            inputs={"code": "def hello_world(): print('Molo Hlabathi')"},
        ),
        langchain.Step(
            igama = "Ukuthunyelwa kwekhowudi",
            func=langchain.deploy_code,
            inputs={"code": "def hello_world(): print('Molo Hlabathi')"},
        ),
    ]
)

# Yenza inkqubo yokusebenza
ukuhamba komsebenzi.phumeza()

Le khowudi ichaza ukuhamba komsebenzi okubandakanya amanyathelo amathathu: uhlalutyo lwekhowudi, ukuhlaziywa kwekhowudi, kunye nokuthunyelwa kwekhowudi. Ukuhamba komsebenzi kusebenzisa i-LLM ukuhlalutya ikhowudi, inkqubo ye-AI ye-agent yokuphonononga ikhowudi, kunye nesixhobo sokuhambisa ukuhambisa ikhowudi kwindawo yokuvelisa.

Top 5-10 Uninzi Imiyalelo / iziqwengana ezisetyenziswayo

Nantsi eyona miyalelo/iziqwengana eziphezulu ezi-5-10 zeLangChain kunye neDSPy:

langchain.Workflow(amanyathelo=[...]): Chaza ukuqhutywa komsebenzi onamanyathelo amaninzi.
langchain.Inyathelo(igama="...", func=..., inputs={...}): Chaza inyathelo elinye kuhambo lomsebenzi.
llm.analyze_code(inputs={"code": ...}): Sebenzisa iLLM ukuhlalutya ikhowudi.
ai_system.review_code(inputs={"code": ...}): Sebenzisa inkqubo ye-AI yokujonga kwakhona ikhowudi.
langchain.deploy_code(inputs={"code": ...}): Faka ikhowudi kwindawo yemveliso.
langchain.AgenticAI(): Yenza umzekelo we-agetic system ye-AI.
langchain.LLM(): Yenza umzekelo weLLM.
langchain.deploy_code(): Faka ikhowudi kwindawo yemveliso.
langchain.execute_workflow(): Phumeza ukuhamba komsebenzi.
langchain.get_workflow_status(): Fumana iwonga lokuhamba komsebenzi.

Umboniso

Nanku umzekelo wendlela yokubona ukuhamba komsebenzi usebenzisa iChart.js:

{
  "type": "bar",
  "idatha": {
    "labels": ["Uhlahlelo lweKhowudi", "Uphononongo lweKhowudi", "Ukusasazwa kweKhowudi"],
    "iiseti zedatha": [
      {
        "label": "Amanyathelo okuhamba komsebenzi",
        "idatha": [10, 20, 30],
        "umbala ongasemva": [
          "rgba(255, 99, 132, 0.2)",
          "rgba(54, 162, 235, 0.2)",
          "rgba(255, 206, 86, 0.2)"
        ],
        "Umbala womda": [
          "rgba(255, 99, 132, 1)",
          "rgba(54, 162, 235, 1)",
          "rgba(255, 206, 86, 1)"
        ],
        "BorderWidth": 1
      }
    ]
  },
  "ukhetho": {}
}

Le khowudi ichaza itshathi yebha eneelebhile ezintathu: Uhlalutyo lweKhowudi, ukuPhononongwa kweKhowudi, kunye nokuhanjiswa kweKhowudi. Itshathi ibonisa inani lamanyathelo kwinqanaba ngalinye lokuhamba komsebenzi.

Isiphelo

I-Agentic AI, ii-LLMs, kunye nezixhobo zomculo ezifana neLangChain kunye ne-DSPy zinamandla okuguqula indlela esisondela ngayo kwi-software architecture kunye ne-DevOps automation. Ngokuzenzekelayo imisebenzi efana nokuhlaziywa kwekhowudi, ukuvavanywa, kunye nokuthunyelwa, sinokuphucula ukusebenza kunye nokusebenza kweenkqubo zethu zophuhliso. Kweli nqaku, siye sahlola ukudityaniswa kwezi teknoloji kwaye sinika umzekelo olula wendlela yokuzisebenzisa ukuzenzekelayo i-software architecture kunye nenkqubo ye-DevOps.