Ho hlophisa liphaephe tsa AI-Powered DevOps tse nang le DSPy le LangChain bakeng sa Patlisiso e Tsoang Pele ea Ketsahalo le Qeto. | B4N1

Ho hlophisa liphaephe tsa AI-Powered DevOps tse nang le DSPy le LangChain bakeng sa Patlisiso e Tsoang Pele ea Ketsahalo le Qeto.

3 months ago · b4n1 Read time: 4 min

Kenyelletso

Ho rarahana ho ntseng ho hola ha lits'ebetso tsa sejoale-joale ho lebisitse keketsehong ea makhetlo le ho teba ha liketsahalo, ho etsa hore lipatlisiso le tharollo ea liketsahalo e be karolo ea bohlokoa ea DevOps. Mahlale a maiketsetso (AI) le ho ithuta ka mochini (ML) li hlahile e le lisebelisoa tse matla tsa ho ntlafatsa liphaephe tsa DevOps, tse nolofalletsang boits'oaro bo tsamaisoang ke AI le orchestration ho bona le ho rarolla liketsahalo pele li ama khoebo. Sengoliloeng sena, re tla hlahloba mokhoa oa ho kenya tšebetsong lipeipi tsa AI-powered DevOps re sebelisa DSPy le LangChain, re tsepamisitse maikutlo ho lipatlisiso le tharollo ea liketsahalo tsa AIOps le RAG, le ho hokahana le AWS DevOps Agent bakeng sa lipatlisiso tsa liketsahalo tse ikemetseng.

Kakaretso ea DSPy le LangChain

DSPy ke laebrari ea Python e fanang ka mokhoa o bonolo le o sebetsang oa ho aha le ho tsamaisa lipeipi tsa AI-powered DevOps. E lumella bahlahisi ho kopanya mefuta e fapaneng ea AI le ML, ho kenyeletsoa ts'ebetso ea puo ea tlhaho (NLP), pono ea komporo, le li-analytics tse lebelloang, liphaepheng tsa bona tsa DevOps. LangChain, ka lehlakoreng le leng, ke laebrari ea Python e nolofalletsang batho ba ntlafatsang ho haha ​​​​le ho tsamaisa mefuta e meholo ea AI ka mokhoa oa ho haha ​​​​modular le scalable.

Ho kenya tšebetsong Lipeipi tsa AI-Powered DevOps ka DSPy le LangChain

Ho kenya tšebetsong liphaephe tsa AI-powered DevOps ka DSPy le LangChain, re hloka ho latela mehato ena:

  1. Pokello ea Lintlha: Bokella lintlha tse amehang ho tsoa mehloling e fapaneng, joalo ka litlaleho, metrics, le lifaele tsa tlhophiso.
  2. Tlhahiso-pele ea Boitsebiso: Etsa esale pele lintlha tse bokeletsoeng ho li lokisetsa bakeng sa tlhahlobo le mohlala.
  3. Thupelo ea Mohlala: Koetlisa mefuta ea AI le ML u sebelisa lintlha tse seng li hlophisitsoe ho bona le ho noha liketsahalo.
  4. Ho Phahamisa Meetso: Kenya mefuta e koetlisitsoeng ka har'a lipeipi tsa DevOps u sebelisa DSPy le LangChain.
  5. Phuputso ea Diketsahalo: Sebelisa mehlala e fanoeng ho fuputsa liketsahalo le ho fumana sesosa sa sesosa.
  6. Qeto ea Ketsahalo: Sebelisa lintlha tse fumanoeng liphuputsong tsa liketsahalo ho rarolla diketsahalo esale pele.

Mohlala oa Khoutu: Ho Sebelisa Mohlala o Bonolo oa RAG ka DSPy le LangChain

Mona ke mohlala o bonolo oa khoutu o bonts'ang mokhoa oa ho kenya ts'ebetsong mofuta oa RAG (Root Cause Analysis) o sebelisa DSPy le LangChain:

kenya dspy
ho tsoa ho langchain ho kenya LLMChain

# Kenya data e fetisitsoeng pele
data = pd.read_csv("data.csv")

# Hlalosa mofuta oa RAG
rag_model = dspy.RAGModel(
    input_features=["log_level", "error_message"],
    lihlahisoa_tsa_litšobotsi=""root_cause"]
)

# Koetlisa mohlala oa RAG
rag_model.train(data)

# Kenya mofuta oa RAG
llm_chain = LLMChain(
    mohlala = "davinci:002",
    max_length=2048
)
rag_model.deploy(llm_chain)

# Sebelisa mofuta oa RAG o kentsoeng ho etsa lipatlisiso ka liketsahalo
ketsahalo_data = pd.DataFrame({
    "log_level": ["ERROR"],
    "error_message": ["Ho na le phoso e etsahetseng"]
})
root_cause = rag_model.predict(incident_data)
hatisa(root_cause)

Mohlala ona oa khoutu o bontša mokhoa oa ho kenya ts'ebetsong mokhoa o bonolo oa RAG o sebelisa DSPy le LangChain. Mohlala oa RAG o nka boemo ba log le molaetsa oa liphoso e le likarolo tsa ho kenya le ho bolela esale pele sesosa sa ketsahalo.

Top 5-10 Litaelo / Linotsoana tse Sebelisitsoeng ka ho Fetisisa

Mona ke tse 5-10 tse sebelisoang ka ho fetesisa litaelo/likotoana tsa DSPy le LangChain:

  1. dspy.RAGModel(input_features, output_features): Hlalosa mofuta oa RAG o nang le likarolo tse itseng tsa ho kenya le ho tsoa.
  2. dspy.RAGModel.train(data): Koetlisa mofuta oa RAG u sebelisa lintlha tse fanoeng.
  3. dspy.RAGModel.deploy(llm_chain): Kenya mofuta oa RAG o koetlisitsoeng o sebelisa LangChain LLMChain.
  4. llm_chain = LLMChain(mohlala, max_length): Theha LangChain LLMChain e nang le mohlala o boletsoeng le bolelele bo boholo.
  5. rag_model.predict(data): Sebelisa mofuta oa RAG o sebelisitsoeng ho bolela esale pele sesosa sa ketsahalo.
  6. dspy.DataPreprocessor(input_data): Hlalosa phaephe ea ho lokisa data ka DSPy.
  7. dspy.ModelTrainer(mohlala, data): Koetlisa mohlala oa AI kapa ML u sebelisa DSPy.
  8. dspy.ModelDeployer(mohlala, llm_chain): Kenya mofuta o koetlisitsoeng o sebelisa DSPy le LangChain.
  9. langchain.LLMChain.predict(input_data): Sebelisa LangChain LLMChain ho bolela esale pele tlhahiso ea mohlala.
  10. dspy.RAGModel.evaluate(data): Lekola tshebetso ya mofuta wa RAG o sebedisa DSPy.

Ho bona pono

Mohlala ke ona oa pono e sebelisang Chart.js:

{
  "type": "bar",
  "data": {
    "labels": ["Mofuta oa Ketsahalo A", "Mofuta oa Ketsahalo ea B", "Mofuta oa Ketsahalo C"],
    "lintlha tsa boitsebiso": [
      {
        "label": "Motso Lebaka A",
        "data": [10, 15, 7],
        "backgroundColor": "rgba(255, 99, 132, 0.2)",
        "borderColor": "rgba(255, 99, 132, 1)",
        "borderWidth": 1
      },
      {
        "label": "Sesosa sa Motso B",
        "data": [8, 12, 9],
        "backgroundColor": "rgba(54, 162, 235, 0.2)",
        "borderColor": "rgba(54, 162, 235, 1)",
        "borderWidth": 1
      },
      {
        "label": "Motso Sesosa C",
        "data": [14, 18, 11],
        "backgroundColor": "rgba(255, 206, 86, 0.2)",
        "borderColor": "rgba(255, 206, 86, 1)",
        "borderWidth": 1
      }
    ]
  },
  "likgetho": {
    "sehlooho": {
      "bontša": 'nete,
      "text": "Tlhahlobo ea Lebaka la Motso"
    },
    "makala": {
      "yAxes": [
        {
          "li-ticks": {
            "beginAtZero": 'nete
          }
        }
      ]
    }
  }
}

Setšoantšo sena se bontša kabo ea lisosa tsa mefuta e fapaneng ea liketsahalo.

Qetello

Sengoliloeng sena, re ile ra hlahloba mokhoa oa ho kenya tšebetsong lipeipi tsa AI-powered DevOps re sebelisa DSPy le LangChain, re tsepamisitse maikutlo lipatlisisong le tharollo ea liketsahalo tsa AIOps le RAG, le ho hokahana le AWS DevOps Agent bakeng sa lipatlisiso tsa liketsahalo tse ikemetseng. Re bontšitse mohlala o bonolo oa khoutu o sebelisang mohlala oa RAG o sebelisang DSPy le LangChain, mme o fane ka lethathamo la litaelo / li-snippets tse ka holimo tsa 5-10 tse sebelisoang haholo bakeng sa DSPy le LangChain. Hape re kenyelelitse mohlala oa pono ka ho sebelisa Chart.js. Ka ho latela mehato e hlalositsoeng sehloohong sena, bahlahisi ba ka haha ​​​​le ho sebelisa liphaephe tsa AI-powered DevOps tse lemohang le ho rarolla liketsahalo ka potlako, ho ntlafatsa katleho le botšepehi ba lisebelisoa tsa bona tsa software.

Have Questions?

Let's Talk

Contact me on LinkedIn