Ukuphumeza i-Agentic AI kwi-Deterministic Test Automation kunye neLangChain kunye ne-DSPy

Ukuphumeza i-Agentic AI kwi-Deterministic Test Automation ngeLangChain kunye ne-DSPy

Intshayelelo

Ukuntsonkotha okwandayo kweenkqubo zesoftware kwenze uvavanyo oluzenzekelayo lube ngumba obalulekileyo wophuhliso lwesoftware. Izikhokelo zovavanyo oluzenzekelayo zovavanyo lwemveli zixhomekeke kwiindlela ezibhalwe phantsi, ezinokuthi zibe brittle kwaye kube nzima ukuzigcina. Kwiminyaka yakutshanje, ukuvela kwe-Agentic AI kunye neZixhobo zeOrchestration kuye kwavula amathuba amatsha okuzenzekelayo iimeko zovavanyo ezinzima. Kweli nqaku, siza kuphonononga ukuhlangana kwe-Agentic AI, iZixhobo zeOkhestration, kunye nobuNjineli beSoftware ukwenza ngokuzenzekelayo iimeko zovavanyo olunzima usebenzisa iLangChain kunye neDSPy.

Yintoni iAgent AI?

I-Agent AI ibhekisa kuhlobo lobukrelekrele bokwenziwa obuyilelwe ukuthatha amanyathelo egameni lomsebenzisi okanye inkqubo. Iinkqubo ze-AI ze-Agent ziyakwazi ukwenza izigqibo, ukufunda kumava, kunye nokuziqhelanisa neemeko ezintsha. Kwimeko yovavanyo oluzenzekelayo, i-Agentic AI ingasetyenziselwa ukuzenzekelayo iimeko zovavanyo olunzima ngokufanisa ukusebenzisana komsebenzisi, ukwenza izigqibo ngokusekelwe kwidatha yovavanyo, kunye nokulungelelanisa ukutshintsha iimeko zovavanyo.

Yintoni iLangChain?

I-LangChain sisixhobo esivulekileyo se-Orchestration Tool esivumela abaphuhlisi ukuba bakhe iinkqubo ze-AI eziyinkimbinkimbi besebenzisa iilwimi ezahlukeneyo zeprogram, kuquka i-Python, i-Rust, kunye ne-Go. I-LangChain inikeza i-architecture eguquguqukayo kunye ne-modular eyenza abaphuhlisi badibanise amacandelo e-AI ahlukeneyo, afana nokulungiswa kolwimi lwendalo (NLP) kunye nombono wekhompyutheni, ukwakha iinkqubo ze-AI eziyinkimbinkimbi.

Yintoni i-DSPy?

I-DSPy lithala leencwadi lePython elibonelela ngeseti yezixhobo zokwakha kunye nokuhambisa iinkqubo ze-AI. I-DSPy yenzelwe ukusebenza ngokugqibeleleyo kunye neLangChain kwaye inikezela uluhlu lweempawu, kubandakanywa ukucutshungulwa kwedatha, ukufundwa komatshini, kunye nokufunda okunzulu.

**Ukuphumeza i-Agentic AI kwi-Deterministic Test Automation **

Ukuphumeza i-Agentic AI kwi-automation test deterministic, sinokusebenzisa i-LangChain kunye ne-DSPy ukwakha isakhelo sokuvavanya esizenzekelayo esilinganisa ukusebenzisana komsebenzisi, senze izigqibo ezisekelwe kwidatha yovavanyo, kunye nokulungelelanisa ukutshintsha iimeko zokuvavanya.

Nanku umzekelo olula wendlela esinokusebenzisa ngayo iLangChain kunye ne-DSPy ukwenza ngokuzenzekelayo imeko yovavanyo:

ngenisa i-langchain
ukusuka langchain.chains ukungenisa LLMChain
ukusuka langchain.llms ngenisa DSPyLLM


test_scenario = {
    "user_input": "login",
    "expected_output": "ukungena kuphumelele"
}

# Chaza iarhente ye-AI ye-Agent
iarhente = LLMChain(
    llm=DSPyLLM(),
    input_type="text",
    imveliso_uhlobo = "okubhaliweyo"
)

# Chaza umsebenzi wovavanyo oluzenzekelayo
def test_automation(test_scenario):
    # Xelisa intsebenziswano yomsebenzisi
    user_input = test_scenario["user_input"]
    # Yenza isigqibo ngokusekwe kwidatha yovavanyo
    isigqibo = agent.run(user_input)
    # Ziqhelanise nokutshintsha iimeko zovavanyo
    ukuba isigqibo == "ukungena kuphumelele":
        print("Uvavanyo luphumelele")
    enye into:
        print("Uvavanyo aluphumelelanga")

# Qhuba umsebenzi wovavanyo oluzenzekelayo
test_automation(test_scenario)

Kulo mzekelo, sichaza imeko yovavanyo ebandakanya ukulinganisa ukusebenzisana komsebenzisi, ukwenza isigqibo ngokusekelwe kwidatha yovavanyo, kunye nokulungelelanisa ukutshintsha iimeko zovavanyo. Emva koko sichaza i-arhente ye-Agentic AI usebenzisa i-LangChain kunye ne-DSPy, kwaye siyisebenzise ukuzenzekelayo imeko yokuvavanya.

Ephezulu 5-10 Eyeyona Isetyenziswayo Imiyalelo/iziqwengana

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

langchain.LLMChain(llm=DSPyLLM()): Yenza i-arhente ye-Agentic AI isebenzisa iLangChain kunye ne-DSPy.
DSPyLLM(): Iqalisa i-DSPy LLM (iModeli yoLwimi oMkhulu) umzekelo.
LLMChain(input_type = "text", output_type = "text"): Yenza i-arhente ye-Agent ye-AI ngegalelo lombhalo kunye neentlobo zemveliso.
i-arhente.run(igalelo_lomsebenzisi): Iqhuba i-arhente ye-AI ye-Agent ngegalelo lomsebenzisi.
test_scenario = {"user_input": "login", "expected_output": "login ngempumelelo"}: Ichaza imeko yovavanyo ngegalelo lomsebenzisi kunye nesiphumo esilindelekileyo.
langchain.chains.LLMChain(llm=DSPyLLM()): Yenza i-arhente ye-Agentic AI isebenzisa iLangChain kunye neDSPy.
DSPyLLM().predict(igalelo_lomsebenzisi): Iqikelela imveliso ye-DSPy LLM ngegalelo lomsebenzisi.
LLMChain(input_type = "text", output_type = "text").run(igalelo_lomsebenzisi): Isebenzisa i-arhente yeAgentic ngegalelo lombhalo kunye neentlobo zemveliso.
uvavanyo_oluzenzekelayo(uvavanyo_imeko): Iqhuba umsebenzi wovavanyo oluzenzekelayo kunye nemeko yovavanyo.
langchain.chains.LLMChain(llm=DSPyLLM()). qhuba(igalelo_lomsebenzisi): Iqhuba i-arhente ye-Agentic ngegalelo lomsebenzisi.

Umbono

Nantsi imbonakalo yearhente ye-AI ye-Agent usebenzisa iChart.js:

{
  "type": "bar",
  "idatha": {
    "labels": ["Igalelo lomsebenzisi", "Isigqibo", "Uhlengahlengiso"],
    "iiseti zedatha": [
      {
        "ileyibheli": "Iarhente ye-AI ye-Agent",
        "idatha": [10, 20, 30],
        "backgroundColor": "rgba(255, 99, 132, 0.2)",
        "borderColor": "rgba(255, 99, 132, 1)",
        "BorderWidth": 1
      }
    ]
  },
  "iinketho": {
    "isihloko": {
      "bonisa": yinyani,
      "umbhalo": "Iarhente ye-AI ye-Agent"
    },
    "izikali": {
      "yAxes": [
        {
          "amanqaku": {
            "beginAtZero": yinyani
          }
        }
      ]
    }
  }
}

Oku kubonwa kubonisa ukusebenza kwe-arhente ye-Agentic AI ekufaniseni ukusebenzisana komsebenzisi, ukwenza izigqibo ngokusekelwe kwidatha yovavanyo, kunye nokulungelelanisa ukutshintsha iimeko zovavanyo.

Isiphelo

Kule nqaku, sihlolisise i-intersection ye-Agentic AI, i-Orchestration Tools, kunye ne-Software Engineering ukwenza ngokuzenzekelayo iimeko zokuvavanya eziyinkimbinkimbi usebenzisa i-LangChain kunye ne-DSPy. Siphumeze umzekelo olula we-Agentic AI kwi-automation test deterministic kwaye sinike uluhlu lwemiyalelo / iziqwenga zeLangChain kunye ne-DSPy. Siphinde sabona ukusebenza kwearhente ye-Agentic sisebenzisa iChart.js. Ngokusebenzisa i-Agentic AI kunye neZixhobo ze-Orchestration, abaphuhlisi banokwakha izikhokelo zokuvavanya ezizenzekelayo ezilinganisa ukusebenzisana komsebenzisi, ukwenza izigqibo ezisekelwe kwidatha yovavanyo, kunye nokuziqhelanisa nokutshintsha kweemeko zovavanyo.

Ukuphumeza i-Agentic AI kwi-Deterministic Test Automation kunye neLangChain kunye ne-DSPy

Have Questions?