You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
kmftools/bookscraper/scraper/tasks/download_tasks.py

216 lines
6.7 KiB

# ============================================================
# File: scraper/tasks/download_tasks.py
# Purpose: Download chapter HTML with global concurrency,
# retry/backoff logic, 429 support, and abort-awareness.
#
# Logging:
# - timestamp + book_id in message
# - logbus.publisher → console
# - ui_log.push_ui → Redis GUI
# ============================================================
from celery_app import celery_app
from scraper.utils import get_save_path
from scraper.abort import abort_requested, chapter_started, mark_chapter_started
from scraper.progress import (
inc_completed,
inc_chapter_done,
inc_chapter_download_skipped,
)
from logbus.publisher import log
from scraper.ui_log import push_ui
import requests
import redis
import os
import time
from datetime import datetime
print(">>> [IMPORT] download_tasks.py loaded")
# -----------------------------------------------------------
# TIMESTAMPED LOG WRAPPER
# -----------------------------------------------------------
def log_msg(book_id: str, message: str):
ts = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
full = f"{ts} [{book_id}] {message}"
log(full)
push_ui(full)
# -----------------------------------------------------------
# ENV CONFIG
# -----------------------------------------------------------
MAX_RETRIES = int(os.getenv("DOWNLOAD_MAX_RETRIES", "7"))
BASE_DELAY = int(os.getenv("DOWNLOAD_BASE_DELAY", "2"))
BACKOFF = int(os.getenv("DOWNLOAD_BACKOFF_MULTIPLIER", "2"))
DELAY_429 = int(os.getenv("DOWNLOAD_429_DELAY", "10"))
MAX_CONCURRENCY = int(os.getenv("DOWNLOAD_MAX_GLOBAL_CONCURRENCY", "1"))
GLOBAL_DELAY = int(os.getenv("DOWNLOAD_GLOBAL_MIN_DELAY", "1"))
REDIS_URL = os.getenv("REDIS_BROKER", "redis://redis:6379/0")
redis_client = redis.Redis.from_url(REDIS_URL)
SEM_KEY = "download:active"
DELAY_KEY = "download:delay_lock"
# -----------------------------------------------------------
# DELAY + CONCURRENCY HELPERS
# -----------------------------------------------------------
def wait_for_global_delay():
if GLOBAL_DELAY <= 0:
return
while redis_client.exists(DELAY_KEY):
time.sleep(0.1)
def set_global_delay():
if GLOBAL_DELAY <= 0:
return
redis_client.set(DELAY_KEY, "1", nx=True, ex=GLOBAL_DELAY)
def acquire_global_slot(max_slots: int, retry_delay: float = 0.5):
while True:
current = redis_client.incr(SEM_KEY)
if current <= max_slots:
return
redis_client.decr(SEM_KEY)
time.sleep(retry_delay)
def release_global_slot():
redis_client.decr(SEM_KEY)
# ============================================================
# CELERY TASK — NEW SIGNATURE WITH chapter_dict + book_meta
# ============================================================
@celery_app.task(bind=True, queue="download", ignore_result=False)
def download_chapter(self, book_id: str, chapter_dict: dict, book_meta: dict):
"""
New unified chapter model:
chapter_dict = {
"num": int,
"url": str,
"title": str,
"volume_path": str
}
book_meta is propagated through the pipeline for parse/save.
"""
chapter_num = chapter_dict.get("num")
chapter_url = chapter_dict.get("url")
chapter_title = chapter_dict.get("title") or f"Chapter {chapter_num}"
volume_path = chapter_dict.get("volume_path")
# -----------------------------------------------------------
# ABORT BEFORE START
# -----------------------------------------------------------
if abort_requested(book_id) and not chapter_started(book_id, chapter_num):
msg = f"[ABORT] Skip chapter {chapter_num} (abort active, not started)"
log_msg(book_id, msg)
inc_chapter_download_skipped(book_id)
return {
"book_id": book_id,
"chapter": chapter_dict,
"html": None,
"skipped": True,
"path": None,
"abort": True,
"book_meta": book_meta,
}
# Mark chapter as started
mark_chapter_started(book_id, chapter_num)
# -----------------------------------------------------------
# SKIP IF FILE ALREADY EXISTS
# -----------------------------------------------------------
save_path = get_save_path(chapter_num, volume_path)
if os.path.exists(save_path):
log_msg(book_id, f"[DL] SKIP {chapter_num} ({chapter_title}) → {save_path}")
return {
"book_id": book_id,
"chapter": chapter_dict,
"html": None,
"skipped": True,
"path": save_path,
"book_meta": book_meta,
}
# -----------------------------------------------------------
# GLOBAL + SYNC DELAY
# -----------------------------------------------------------
if GLOBAL_DELAY > 0:
time.sleep(GLOBAL_DELAY)
wait_for_global_delay()
acquire_global_slot(MAX_CONCURRENCY)
# log_msg(book_id, f"[DL] ACQUIRED SLOT for chapter {chapter_num}")
# -----------------------------------------------------------
# HTTP DOWNLOAD
# -----------------------------------------------------------
try:
log_msg(
book_id,
f"[DL] Downloading {chapter_num} ({chapter_title}): {chapter_url}",
)
resp = requests.get(
chapter_url,
headers={"User-Agent": "Mozilla/5.0"},
timeout=20,
)
resp.raise_for_status()
resp.encoding = resp.apparent_encoding or "gb2312"
html = resp.text
log_msg(book_id, f"[DL] OK {chapter_num}: {len(html)} bytes")
return {
"book_id": book_id,
"chapter": chapter_dict,
"html": html,
"skipped": False,
"path": save_path,
"book_meta": book_meta,
}
except Exception as exc:
attempt = self.request.retries
delay = BASE_DELAY * (BACKOFF**attempt)
# Specific 429 handler
if getattr(getattr(exc, "response", None), "status_code", None) == 429:
log_msg(
book_id,
f"[DL] 429 {chapter_num} → WAIT {DELAY_429}s "
f"(attempt {attempt}/{MAX_RETRIES})",
)
time.sleep(DELAY_429)
set_global_delay()
raise self.retry(exc=exc, countdown=0, max_retries=MAX_RETRIES)
# Normal retry
log_msg(
book_id,
f"[DL] ERROR {chapter_num}: {exc} → retry in {delay}s "
f"(attempt {attempt}/{MAX_RETRIES})",
)
raise self.retry(exc=exc, countdown=delay, max_retries=MAX_RETRIES)
finally:
set_global_delay()
release_global_slot()
# log_msg(book_id, f"[DL] RELEASED SLOT for chapter {chapter_num}")