** Aiwatar da AI Agent a cikin Gwajin Gwajin Aiki tare da LangChain da DSPy ***

** Gabatarwa ***

Ƙirƙirar tsarin software ya sanya aikin gwaji ya zama muhimmin al'amari na haɓaka software. Tsarin gwaji na al'ada na sarrafa kansa ya dogara da hanyoyin da aka rubuta, waɗanda za su iya zama gagara kuma mai wahalar kiyayewa. A cikin 'yan shekarun nan, fitowar Agentic AI da Orchestration Tools sun buɗe sabbin damar yin aiki da hadaddun yanayin gwaji. A cikin wannan labarin, za mu bincika haɗin gwiwar Agent AI, Kayan Aikin Orchestration, da Injiniya na Software don sarrafa hadaddun yanayin gwaji ta amfani da LangChain da DSPy.

** Menene Agent AI?**

Agent AI yana nufin nau'in hankali na wucin gadi wanda aka ƙera don ɗaukar ayyuka a madadin mai amfani ko tsarin. Tsarin AI na ainihi yana da ikon yanke shawara, koyo daga gogewa, da daidaitawa zuwa sabbin yanayi. A cikin mahallin gwajin sarrafa kansa, ana iya amfani da Agentic AI don sarrafa rikitattun yanayin gwaji ta hanyar daidaita ma'amalar mai amfani, yanke shawara dangane da bayanan gwaji, da daidaitawa ga canza yanayin gwaji.

** Menene LangChain?**

LangChain kayan aikin Orchestration ne mai buɗewa wanda ke ba masu haɓakawa damar gina hadaddun tsarin AI ta amfani da harsunan shirye-shirye iri-iri, gami da Python, Rust, da Go. LangChain yana ba da tsarin gine-gine mai sassauƙa da na yau da kullun wanda ke ba masu haɓaka damar haɗa nau'ikan abubuwan AI daban-daban, kamar sarrafa harshe na halitta (NLP) da hangen nesa na kwamfuta, don gina ƙaƙƙarfan tsarin AI.

Mene ne DSPy?

DSPy ɗakin karatu ne na Python wanda ke ba da saiti na kayan aikin gini da tura tsarin AI. An tsara DSPy don yin aiki tare da LangChain kuma yana ba da nau'i-nau'i iri-iri, ciki har da sarrafa bayanai, koyan na'ura, da ilmantarwa mai zurfi.

** Aiwatar da AI Agent a cikin Ƙaddamarwar Gwajin Aiki Automation**

Don aiwatar da AI Agentic AI a cikin ƙirar gwajin ƙididdigewa, za mu iya amfani da LangChain da DSPy don gina tsarin ƙirar gwajin gwaji wanda ke daidaita ma'amalar mai amfani, yin yanke shawara dangane da bayanan gwaji, kuma ya dace da canza yanayin gwaji.

Ga misali mai sauƙi na yadda za mu iya amfani da LangChain da DSPy don sarrafa yanayin gwaji: `` Python shigo da langchain daga langchain.chains shigo da LLMChain daga langchain.llms shigo da DSPyLLM

test_scenario = { "user_input": "login", "expected_output": "cikin nasara" }

ayyana wakilin AI Agent

wakili = LLMChain ( llm=DSPyLLM(), input_type = "rubutu", fitarwa_type = "rubutu" )

Ƙayyade aikin sarrafa kansa na gwaji

def test_automation(test_scenario): # Kwaikwayi hulɗar mai amfani user_input = test_scenario["user_input"] # Yi shawara bisa bayanan gwaji yanke shawara = wakili.run(user_input) # Daidaita da canza yanayin gwaji idan shawarar == "shigo yayi nasara": buga ("Gwaji ya wuce") wani: buga ("Gwajin ya kasa")

Gudanar da aikin gwaji ta atomatik

test_atomatik(test_scenario)

A cikin wannan misalin, mun ayyana yanayin gwaji wanda ya haɗa da yin kwaikwayon hulɗar mai amfani, yanke shawara bisa bayanan gwaji, da daidaitawa ga canza yanayin gwaji. Sai mu ayyana wani wakilin AI ta amfani da LangChain da DSPy, kuma muna amfani da shi don sarrafa yanayin gwajin.

** 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.LLMChain(llm=DSPyLLM()): Yana ƙirƙira wakili na AI ta amfani da LangChain da DSPy.
2. `DSPyLLM()`: Yana ƙaddamar da misali na DSPy LLM (Babban Harshe Model).
3. `LLMChain (input_type = "rubutu", fitarwa_type = "rubutu")`: Yana ƙirƙira wakili na AI tare da shigar da rubutu da nau'ikan fitarwa.
4. `agent.run(user_input): Yana gudanar da wakilin AI tare da shigar da mai amfani.
5. `test_scenario = {"user_input": "login", "expected_output": "layi nasara shiga"}`: Yana bayyana yanayin gwaji tare da shigarwar mai amfani da fitar da ake sa ran.
6. `langchain.chains.LLMChain(llm=DSPyLLM())`: Yana ƙirƙirar wakili na AI ta amfani da LangChain da DSPy.
7. `DSPyLLM().predict(user_input)`: Yana tsinkayar fitowar DSPy LLM tare da shigar mai amfani.
8. `LLMChain (input_type = "rubutu", fitarwa_type = "rubutu").run(user_input):': Yana gudanar da wakilin AI tare da shigarwar rubutu da nau'ikan fitarwa.
9. `test_automation(test_scenario)`: Yana gudanar da aikin sarrafa kansa na gwaji tare da yanayin gwaji.
10. `langchain.chains.LLMChain(llm=DSPyLLM()) .run(user_input)`: Yana gudanar da wakilin AI tare da shigar da mai amfani.

**Kallon gani**

Anan ga hangen nesa na wakilin AI ta amfani da Chart.js:
Json
{
  "type": "bar",
  "data": {
    "labels": ["Input User", "Decision", "Adaptation"],
    "datasets": [
      {
        "label": "Wakilin AI",
        "bayanai": [10, 20, 30],
        "backgroundColor": "rgba(255, 99, 132, 0.2)",
        "BorderColor": "rgba(255, 99, 132, 1)",
        "BorderWidth": 1
      }
    ]
  },
  "zaɓi": {
    "take": {
      "nuna": gaskiya,
      "text": "Agent AI Agent"
    },
    "ma'auni": {
      "yAxes": [
        {
          "zama": {
            "beginAtZero": gaskiya
          }
        }
      ]
    }
  }
}

Wannan hangen nesa yana nuna aikin wakilin AI na Agent a cikin simintin hulɗar masu amfani, yanke shawara dangane da bayanan gwaji, da daidaitawa ga canza yanayin gwaji.

Kammalawa

A cikin wannan labarin, mun binciko haɗin gwiwar Agentic AI, Kayan aikin Orchestration, da Injiniya na Software don sarrafa hadaddun yanayin gwaji ta amfani da LangChain da DSPy. Mun aiwatar da misali mai sauƙi na AI Agentic a cikin ƙirar gwajin ƙididdigewa kuma mun samar da kewayon umarni / snippets don LangChain da DSPy. Mun kuma hango aikin wakilin AI na Agent ta amfani da Chart.js. Ta hanyar amfani da AI da Kayayyakin Orchestration, masu haɓakawa za su iya gina ingantattun tsarin sarrafa kansa na gwaji waɗanda ke kwaikwaya mu'amalar mai amfani, yanke shawara dangane da bayanan gwaji, da daidaitawa ga canza yanayin gwaji.