asyncio.gather għall-Ipproċċar f'Lotti
F'din il-paġna
Orkestrazzjoni ta’ Talbiet Paralleli
asyncio.gather hija l-funzjoni ewlenija biex tħaddem ħafna coroutines flimkien u tiġbor ir-riżultati kollha. Meta tingħaqad ma’ semaphore, tippermettilek tipproċessa lotti ta’ talbiet b’mod effiċjenti filwaqt li tirrispetta l-limiti tal-API.
Sintassi Bażika
import asyncio
requests = ["Question 1", "Question 2", "Question 3"]
tasks = [process_request(r) for r in requests]
responses = await asyncio.gather(*tasks)
asyncio.gather tieħu numru varjabbli ta’ coroutines (għalhekk l-użu ta’ *tasks biex tiftaħ il-lista) u tirritorna lista ta’ riżultati bl-istess ordni bħall-kompiti sottomessi. Dan huwa punt importanti: anke jekk il-mistoqsija 3 tispiċċa qabel il-mistoqsija 1, responses[0] dejjem ikun fih ir-riżultat għall-mistoqsija 1.
Mudell Sħiħ: Lott b’Semaphore
Hawnhekk hawn il-mudell li se tuża l-aktar spiss fil-produzzjoni:
import os
from asyncio import Semaphore
from openai import AsyncOpenAI
import httpx
client = AsyncOpenAI(
api_key=os.getenv("XAI_API_KEY"),
base_url="https://api.x.ai/v1",
timeout=httpx.Timeout(3600.0)
)
async def process_batch(prompts: list[str], max_concurrent: int = 5) -> list[str]:
sem = Semaphore(max_concurrent)
async def single_request(prompt: str) -> str:
async with sem:
response = await client.chat.completions.create(
model="grok-4.5",
messages=[{"role": "user", "content": prompt}]
)
return response.choices[0].message.content
tasks = [single_request(p) for p in prompts]
return await asyncio.gather(*tasks)
Dan il-mudell jinkapsula s-semaphore fil-funzjoni tal-lott. Għandek biss bżonn issejjaħ process_batch(prompts) u tirrikupra l-lista ta’ risposti.
Ġestjoni tal-Iżbalji f’Lott
B’mod predefinit, jekk coroutine waħda tħajjar eċċezzjoni, asyncio.gather tikkanċella kollox u tippropaga l-iżball. Il-parametru return_exceptions=True ibiddel dan l-imġiba:
async def safe_batch(prompts: list[str]) -> list[str | Exception]:
sem = Semaphore(5)
async def single_request(prompt: str) -> str:
async with sem:
response = await client.chat.completions.create(
model="grok-4.5",
messages=[{"role": "user", "content": prompt}]
)
return response.choices[0].message.content
tasks = [single_request(p) for p in prompts]
results = await asyncio.gather(*tasks, return_exceptions=True)
return results
# Użu
results = await safe_batch(mes_prompts)
for i, result in enumerate(results):
if isinstance(result, Exception):
print(f"Error on prompt {i}: {result}")
else:
print(f"Result {i}: {result[:100]}...")
B’return_exceptions=True, l-iżbalji jiġu maqbuda bħala valuri fil-lista tar-riżultati. Imbagħad tista’ tidentifika u terġa’ tipprova t-talbiet li fallew.
Qsim f’Sottolotti (Chunking)
Għal volumi kbar ħafna, huwa aħjar li taqsam f’sottolotti minflok ma tniedi eluf ta’ kompiti fl-istess ħin:
async def chunked_batch(
prompts: list[str],
chunk_size: int = 50,
max_concurrent: int = 5
) -> list[str]:
all_results = []
for i in range(0, len(prompts), chunk_size):
chunk = prompts[i:i + chunk_size]
print(f"Processing batch {i//chunk_size + 1} ({len(chunk)} prompts)")
results = await process_batch(chunk, max_concurrent)
all_results.extend(results)
return all_results
Il-”chunking” joffri diversi vantaġġi: tista’ turi l-progress, issalva riżultati intermedjarji, u tillimita l-użu tal-memorja. Jekk jifalli sottolott wieħed, ma titlefx ir-riżultati mis-sottolotti preċedenti.
Traċċar tal-Progress
Għal lotti kbar, indikatur tal-progress huwa utli:
import asyncio
from asyncio import Semaphore
async def batch_with_progress(prompts: list[str], max_concurrent: int = 5) -> list[str]:
sem = Semaphore(max_concurrent)
completed = 0
total = len(prompts)
async def single_request(prompt: str) -> str:
nonlocal completed
async with sem:
response = await client.chat.completions.create(
model="grok-4.5",
messages=[{"role": "user", "content": prompt}]
)
completed += 1
print(f"\r{completed}/{total} completed", end="", flush=True)
return response.choices[0].message.content
tasks = [single_request(p) for p in prompts]
results = await asyncio.gather(*tasks)
print() # linja ġdida wara l-progress
return results
Punti Ewlenin
asyncio.gather(*tasks)iħaddem coroutines paralleli u jirritorna r-riżultati fl-ordni oriġinalireturn_exceptions=Truejipprevjeni li żball wieħed ifalli l-lott kollu- Il-”chunking” jippermettilek tittratta volumi kbar ħafna mingħajr ma tissatura l-memorja
- Dejjem għaqqad
gatherma’ semaphore biex tirrispetta l-limiti tar-rata - Żid traċċar tal-progress għal proċessi li jieħdu żmien twil