Gabatarwa

Saurin juyin halitta na Intelligence Artificial Intelligence (AI) da Injin Learning (ML) ya haifar da haɓaka AI mai aiki, nau'in AI wanda zai iya ɗaukar ayyuka da yanke shawara da kansa. Wannan fasaha tana da yuwuwar yin juyin juya hali yadda muke kusanci tsarin gine-ginen software da sarrafa kansa na DevOps. A cikin wannan labarin, za mu bincika haɗin kai na AI mai aiki, Manyan Harshe Model (LLMs), da kayan aikin ƙungiyar kamar LangChain da DSPy don sarrafa tsarin gine-ginen software, abubuwan more rayuwa, da ayyukan DevOps.

Menene Agent AI?

Agent AI yana nufin nau'in AI wanda zai iya ɗaukar ayyuka kuma ya yanke shawara da kansa. Wannan ya bambanta da tsarin AI na gargajiya, waɗanda galibi an tsara su don yin takamaiman ayyuka da bin ƙa'idodin ƙayyadaddun ƙayyadaddun ƙa'idodi. Agentic AI tsarin, a gefe guda, an tsara su don zama masu zaman kansu da daidaitawa, ba su damar koyo daga yanayin su da kuma yanke shawara bisa ga burinsu da manufofinsu.

Matsayin LLMs a cikin AI

Manyan Harshe Model (LLMs) nau'in AI ne waɗanda aka horar da su akan ɗimbin bayanan rubutu kuma suna da ikon samar da harshe irin na ɗan adam. LLMs sune mahimmin ɓangaren tsarin AI na aiki, saboda suna ba da ikon fahimta da samar da harshen ɗan adam. Wannan yana ba da damar tsarin AI na aiki don sadarwa tare da mutane da sauran tsarin ta hanyar da ta fi dacewa da hankali.

Kayan aikin Orchestration: LangChain da DSPy

Kayan aikin Orchestration kamar LangChain da DSPy an tsara su don sarrafawa da daidaita hulɗar tsakanin tsarin daban-daban da sassa. Wadannan kayan aikin suna ba da hanyar da za a ƙayyade da aiwatar da ayyukan aiki, wanda zai iya haɗawa da ayyuka kamar sarrafa bayanai, horar da samfurin, da ƙaddamarwa.

LangChain

LangChain kayan aiki ne na buɗaɗɗen kaɗe-kaɗe wanda ke ba da tsari mai sassauƙa da haɓakawa don ginawa da sarrafa ayyukan aiki. An tsara LangChain don yin aiki tare da nau'ikan tsarin da aka gyara, ciki har da LLMs, bayanai, da APIs.

DSPy

DSPy kayan aiki ne na tushen Python wanda ke ba da hanya mai sauƙi da fahimta don ayyana da aiwatar da ayyukan aiki. An ƙera DSPy don yin aiki tare da kewayon tsari da abubuwan haɗin gwiwa, gami da LLMs, bayanan bayanai, da APIs.

Gine-ginen Software na atomatik da Tsarin DevOps

AI, LLMs, da kayan aikin kade-kade kamar LangChain da DSPy ana iya amfani da su don sarrafa kewayon kewayon gine-ginen software da ayyukan DevOps. Wasu misalan sun haɗa da:

** Bita na lamba ta atomatik ***: Ana iya amfani da tsarin AI na asali don duba lamba da gano abubuwan da za su iya yiwuwa, kamar kwari da raunin tsaro. ** Gwajin sarrafa kansa ***: Ana iya amfani da tsarin AI na asali don sarrafa gwaji, gami da gwajin naúrar, gwajin haɗin kai, da gwajin ƙarshe zuwa ƙarshe. ** Aiki ta atomatik ***: Za a iya amfani da tsarin AI na asali don yin aiki da kai, gami da ƙaddamar da lamba, daidaitawa, da sauran albarkatu.

** Sa ido ta atomatik ***: Ana iya amfani da tsarin AI na asali don sarrafa sarrafa kansa, gami da sa ido kan aikin tsarin, tsaro, da sauran ma'auni.

Aiki Mai Aiki

A cikin wannan sashe, za mu ba da misali mai sauƙi na yadda ake amfani da LangChain da DSPy don sarrafa tsarin gine-ginen software da DevOps.

Misali: Nau'in Code Mai sarrafa kansa

A cikin wannan misali, za mu yi amfani da LangChain da DSPy don sarrafa bitar lambar. Za mu ayyana tsarin aiki wanda ya haɗa da matakai masu zuwa:

** Binciken lambar ***: Aikin aiki zai yi amfani da LLM don nazarin lambar kuma gano abubuwan da za su iya yiwuwa.
** Bita na lamba ***: Gudun aikin zai yi amfani da tsarin AI mai aiki don duba lambar kuma gano abubuwan da za su iya yiwuwa.
** Ƙaddamar da lambar ***: Aikin aiki zai yi amfani da kayan aiki na kayan aiki don ƙaddamar da lambar zuwa yanayin samarwa.

Anan ga misalin yadda ake aiwatar da wannan aikin ta amfani da LangChain da DSPy: `` Python shigo da langchain daga langchain shigo da LLM

ayyana LLM

llm = LLM ("rubutu-davinci-003")

Ƙayyade tsarin AI mai aiki

ai_system = langchain.AgenticAI()

Bayyana tsarin aiki

aikin aiki = langchain. Gudun aiki ( matakai=[ langchain. Mataki( suna = "Nazarin lamba", func=llm.analyze_code, shigarwar = {"code": "def hello_world(): buga ('Hello Duniya')"}, ), langchain. Mataki( name="Bita na lamba", func=ai_system.review_code, shigarwar = {"code": "def hello_world(): buga ('Hello Duniya')"}, ), langchain. Mataki( name="Tsarin lambar", func=langchain.deploy_code, shigarwar = {"code": "def hello_world(): buga ('Hello Duniya')"}, ), ] )

Gudanar da aikin

aikin aiki.execute()

Wannan lambar tana bayyana tsarin aiki wanda ya haɗa da matakai uku: ƙididdigar lamba, sake duba lambar, da tura lambar. Gudun aikin yana amfani da LLM don nazarin lambar, tsarin AI mai aiki don duba lambar, da kayan aiki na turawa don ƙaddamar da lambar zuwa yanayin samarwa.

## Manyan Dokoki 5-10 da Akafi Amfani da su/Snippets

Anan akwai manyan umarni 5-10 da aka fi amfani da su/snippets don LangChain da DSPy:

1. `langchain. Gudun aiki (matakai = [...])`: Ƙayyade aikin aiki tare da matakai masu yawa.
2. `langchain.Mataki(name="...", func=..., inputs={...})": Ƙayyade mataki ɗaya a cikin tafiyar aiki.
3. `llm.analyze_code(inputs={"code": ...})`: Yi amfani da LLM don tantance lamba.
4. `ai_system.review_code(inputs={"code": ...})': Yi amfani da tsarin AI mai aiki don duba lamba.
5. `langchain.deploy_code(input={"code": ...})': Sanya lamba zuwa yanayin samarwa.
6. `langchain.AgenticAI()': Ƙirƙiri misali na tsarin AI mai aiki.
7. `langchain.LLM()`: Ƙirƙiri misali na LLM.
8. `langchain.deploy_code()`: Sanya lamba zuwa yanayin samarwa.
9. `langchain.execute_workflow()`: Yi aikin aiki.
10. `langchain.get_workflow_status()`: Sami matsayin tafiyar aiki.

## Kallon gani

Anan ga misalin yadda ake ganin tafiyar aiki ta amfani da Chart.js:
Json
{
  "type": "bar",
  "data": {
    "labels": ["Nazarin lambar", "Bita na lamba", "Tsarin lambar"],
    "datasets": [
      {
        "label": "Matakan Gudun Aiki",
        "bayanai": [10, 20, 30],
        "BackgroundLauni": [
          "rgba (255, 99, 132, 0.2)",
          "rgba (54, 162, 235, 0.2)",
          "rgba (255, 206, 86, 0.2)"
        ],
        "Launi Border": [
          "rgba (255, 99, 132, 1)",
          "rgba (54, 162, 235, 1)",
          "rgba (255, 206, 86, 1)"
        ],
        "BorderWidth": 1
      }
    ]
  },
  "zaɓi": {}
}

Wannan lambar tana bayyana taswirar mashaya tare da alamomi guda uku: Lambobin Lambobi, Bita na lamba, da Ƙaddamar da Code. Jadawalin yana nuna adadin matakai a kowane mataki na tafiyar aiki.

Kammalawa

AI, LLMs, da kayan aikin kade-kade kamar LangChain da DSPy suna da yuwuwar sauya yadda muke kusanci tsarin gine-ginen software da sarrafa kansa na DevOps. Ta hanyar sarrafa ayyuka ta atomatik kamar bita na lamba, gwaji, da turawa, za mu iya inganta inganci da tasiri na ayyukan ci gaban mu. A cikin wannan labarin, mun bincika haɗin waɗannan fasahohin kuma mun ba da misali mai sauƙi na yadda ake amfani da su don sarrafa tsarin gine-ginen software da DevOps.

Bayar da Agent AI don Gine-ginen Software da DevOps Automation