Kukhazikitsa Agentic AI mu Deterministic Test Automation ndi LangChain ndi DSPy

Kukhazikitsa Agentic AI mu Deterministic Test Automation ndi LangChain ndi DSPy

Chiyambi

Kuchulukirachulukira kwamakina apulogalamu kwapangitsa kuyesa makina kukhala gawo lofunikira kwambiri pakupanga mapulogalamu. Zoyeserera zachikale zoyeserera zimadalira njira zolembera, zomwe zimatha kukhala zolimba komanso zovuta kuzisamalira. M'zaka zaposachedwa, kutuluka kwa Agentic AI ndi Orchestration Tools kwatsegula mwayi watsopano wodzipangira mayeso ovuta. M'nkhaniyi, tiwona mphambano ya Agentic AI, Orchestration Tools, ndi Software Engineering kuti tigwiritse ntchito mayesero ovuta pogwiritsa ntchito LangChain ndi DSPy.

Agent AI ndi chiyani?

Agentik AI imatanthawuza mtundu wanzeru zopanga zomwe zimapangidwa kuti zichitepo kanthu m'malo mwa wogwiritsa ntchito kapena dongosolo. Machitidwe a Agetic AI amatha kupanga zisankho, kuphunzira kuchokera pazomwe akumana nazo, ndikusinthira kuzochitika zatsopano. Pankhani yoyeserera zokha, Agentic AI itha kugwiritsidwa ntchito kusinthiratu zochitika zovuta zoyeserera poyesa kuyanjana kwa ogwiritsa ntchito, kupanga zisankho motengera zomwe zayesedwa, ndikusintha kusintha kwa mayeso.

LangChain ndi chiyani?

LangChain ndi chida chotseguka cha Orchestration Tool chomwe chimalola opanga kupanga makina ovuta a AI pogwiritsa ntchito zilankhulo zosiyanasiyana zamapulogalamu, kuphatikiza Python, Rust, ndi Go. LangChain imapereka mapangidwe osinthika komanso osinthika omwe amathandizira opanga kuphatikizira magawo osiyanasiyana a AI, monga chilankhulo chachilengedwe (NLP) ndi masomphenya apakompyuta, kuti apange machitidwe apamwamba a AI.

DSPy ndi chiyani?

DSPy ndi laibulale ya Python yomwe imapereka zida zingapo zomangira ndi kutumiza machitidwe a AI. DSPy idapangidwa kuti izigwira ntchito mosasunthika ndi LangChain ndipo imapereka zinthu zingapo, kuphatikiza kukonza deta, kuphunzira pamakina, ndi kuphunzira mozama.

**Kukhazikitsa Agent AI mu Deterministic Test Automation **

Kuti tigwiritse ntchito Agentic AI mu deterministic test automation, tikhoza kugwiritsa ntchito LangChain ndi DSPy kupanga ndondomeko yoyesera yokha yomwe imayesa kuyanjana kwa ogwiritsa ntchito, kupanga zisankho potengera deta yoyesera, ndikusintha kusintha kwa mayesero.

Nachi chitsanzo chosavuta cha momwe tingagwiritsire ntchito LangChain ndi DSPy kuti tigwiritse ntchito mayeso:

import langchain
kuchokera ku langchain.chains kuitanitsa LLMChain
kuchokera ku langchain.llms lowetsani DSPyLLM


test_scenario = {
    "user_input": "login",
    "expected_output": "kulowa mwapambana"
}

# Tanthauzirani wothandizira wa Agetic AI
wothandizira = LLMChain (
    llm=DSPyLLM(),
    input_type="mawu",
    output_type = "mawu"
)

# Fotokozani ntchito yoyeserera yokha
def test_automation(test_scenario):
    # Yendetsani kuyanjana kwa ogwiritsa ntchito
    user_input = test_scenario["user_input"]
    # Pangani chisankho potengera zomwe mwayesa
    chisankho = agent.run(user_input)
    # Gwirizanani ndi kusintha koyeserera
    ngati chisankho == "kulowa bwino":
        sindikiza ("Mayeso adutsa")
    zina:
        sindikiza ("Mayeso alephera")

# Yendetsani ntchito yoyeserera yokha
test_automation(test_scenario)

Muchitsanzo ichi, tikufotokozera za mayeso omwe amaphatikizapo kuyerekezera kuyanjana kwa ogwiritsa ntchito, kupanga chigamulo potengera data yoyeserera, ndikusintha kusintha kwa mayeso. Kenako timatanthauzira wothandizira wa Agentic AI pogwiritsa ntchito LangChain ndi DSPy, ndikuigwiritsa ntchito kuti isinthe mawonekedwe a mayeso.

**Malamulo/Zidutswa Zapamwamba za 5-10 **

Nawa malamulo/zidutswa 5-10 zomwe zimagwiritsidwa ntchito kwambiri pa LangChain ndi DSPy:

1. langchain.LLMChain(llm=DSPyLLM()): Amapanga Agent AI wothandizira pogwiritsa ntchito LangChain ndi DSPy.
2. DSPyLLM(): Imayambitsa chitsanzo cha DSPy LLM (Large Language Model).
3. LLMChain(input_type = "zolemba", output_type = "zolemba"): Amapanga Agent AI agent yokhala ndi zolemba ndi mitundu yotulutsa.
4. agent.run(user_input): Imayendetsa wothandizila Agent AI ndi zolowetsa za ogwiritsa.
5. test_scenario = {"user_input": "login", "expected_output": "login yachita bwino"}: Imatanthawuza mayesero omwe ali ndi kulowetsamo kwa wogwiritsa ntchito ndi zomwe zikuyembekezeka.
6. langchain.chains.LLMChain(llm=DSPyLLM()): Amapanga Agentic AI agent pogwiritsa ntchito LangChain ndi DSPy.
7. DSPyLLM().predict(user_input): Imaneneratu za kutulutsa kwa DSPy LLM ndi kulowetsa kwa wogwiritsa ntchito.
8. LLMChain(input_type = "zolemba", output_type = "zolemba").run(user_input): Imayendetsa wothandizila Agentic AI ndi zolowetsa mawu ndi mitundu yotulutsa.
9. test_automation(test_scenario): Imayendetsa ntchito yoyeserera ndi kuyesa.
10. langchain.chains.LLMChain(llm=DSPyLLM()).run(user_input): Imayendetsa wothandizila wa Agetic AI ndikulowetsa kwa wogwiritsa.

**Kuwona **

Nachi chithunzithunzi cha wothandizira wa Agetic AI pogwiritsa ntchito Chart.js:

{
  "type": "bar",
  "data": {
    "labels": ["Zolowetsa Wogwiritsa", "Decision", "Adaptation"],
    "ma dataset": [
      {
        "label": "Agent AI Agent",
        "data": [10, 20, 30],
        "backgroundColor": "rgba(255, 99, 132, 0.2)",
        "borderColor": "rgba(255, 99, 132, 1)",
        "BorderWidth": 1
      }
    ]
  },
  "zosankha": {
    "mutu": {
      "kuwonetsa": zoona,
      "text": "Agent AI Agent"
    },
    "miyeso": {
      "yAxes": [
        {
          "nkhupakupa": {
            "beginAtZero": zoona
          }
        }
      ]
    }
  }
}

Chiwonetserochi chikuwonetsa magwiridwe antchito a wothandizira wa Agent AI poyerekezera kuyanjana kwa ogwiritsa ntchito, kupanga zisankho motengera data yoyeserera, ndikusintha kuti zigwirizane ndi zoyeserera.

Mapeto

M'nkhaniyi, tafufuza njira zodutsamo Agentic AI, Orchestration Tools, ndi Software Engineering kuti tigwiritse ntchito mayesero ovuta pogwiritsa ntchito LangChain ndi DSPy. Tidagwiritsa ntchito chitsanzo chosavuta cha Agentic AI poyesa makina oyeserera ndikupereka malamulo/zidutswa zingapo za LangChain ndi DSPy. Tidawonanso momwe agentic AI amathandizira pogwiritsa ntchito Chart.js. Pogwiritsa ntchito zida za Agentic AI ndi Orchestration, opanga amatha kupanga zida zapamwamba zoyeserera zomwe zimatengera kuyanjana kwa ogwiritsa ntchito, kupanga zisankho potengera mayeso a data, ndikusintha kuti zigwirizane ndi kusintha kwa mayeso.