This commit is contained in:
doms9 2025-10-30 15:38:34 -04:00
parent 566f5c2c80
commit 00000d980a
12 changed files with 337 additions and 547 deletions

View file

@ -1,4 +1,3 @@
import asyncio
import re
from functools import partial
from pathlib import Path
@ -6,7 +5,7 @@ from typing import Any
from urllib.parse import urljoin
import httpx
from playwright.async_api import BrowserContext, async_playwright
from playwright.async_api import async_playwright
from .utils import Cache, Time, get_logger, leagues, network
@ -50,62 +49,6 @@ async def refresh_api_cache(
return data
async def process_event(
url: str,
url_num: int,
context: BrowserContext,
) -> str | None:
page = await context.new_page()
captured: list[str] = []
got_one = asyncio.Event()
handler = partial(network.capture_req, captured=captured, got_one=got_one)
page.on("request", handler)
try:
await page.goto(
url,
wait_until="domcontentloaded",
timeout=15_000,
)
wait_task = asyncio.create_task(got_one.wait())
try:
await asyncio.wait_for(wait_task, timeout=10)
except asyncio.TimeoutError:
log.warning(f"URL {url_num}) Timed out waiting for M3U8.")
return
finally:
if not wait_task.done():
wait_task.cancel()
try:
await wait_task
except asyncio.CancelledError:
pass
if captured:
log.info(f"URL {url_num}) Captured M3U8")
return captured[-1]
log.warning(f"URL {url_num}) No M3U8 captured after waiting.")
return
except Exception as e:
log.warning(f"URL {url_num}) Exception while processing: {e}")
return
finally:
page.remove_listener("request", handler)
await page.close()
async def get_events(
client: httpx.AsyncClient,
base_url: str,
@ -211,10 +154,11 @@ async def scrape(client: httpx.AsyncClient) -> None:
for i, ev in enumerate(events, start=1):
handler = partial(
process_event,
network.process_event,
url=ev["link"],
url_num=i,
context=context,
log=log,
)
url = await network.safe_process(