123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306 |
- import asyncio
- import uuid
- from typing import List, Dict, Optional, Any
- import aiohttp
- from core.models.video_item import VideoItem
- from core.utils.helpers import generate_titles
- from core.utils.spider_config import SpiderConfig
- from core.utils.extractors import safe_extract, extract_fields
- from core.utils.log.logger_manager import LoggerManager
- from core.utils.template_resolver import resolve_request_body_template
- from services.async_mysql_service import AsyncMysqlService
- from services.pipeline import PiaoQuanPipeline
- from core.base.async_request_client import AsyncRequestClient
- from services.async_mq_producer import AsyncMQProducer
- class BaseSpider:
- """
- 通用爬虫基类,支持:
- - 依赖请求参数动态替换(cursor 或其它参数)
- - 支持单请求和依赖请求的分页抓取
- - 统一日志、MQ推送、异常捕获、异步请求
- 子类只需根据业务重写少量方法,如 process_video/process_item。
- """
- def __init__(self, rule_dict: Dict, user_list: List, env: str = "prod"):
- self.env = env
- self.user_list = user_list
- self.rule_dict = rule_dict
- self.class_name = self.__class__.__name__.lower()
- # 通过小写子类名读取配置
- self.platform_config = SpiderConfig.get_platform_config(classname=self.class_name)
- if not self.platform_config:
- raise ValueError(f"找不到对应配置: {self.class_name}")
- self.platform = self.platform_config.platform
- self.mode = self.platform_config.mode
- self.logger = LoggerManager.get_logger(platform=self.platform, mode=self.mode)
- self.aliyun_log = LoggerManager.get_aliyun_logger(platform=self.platform, mode=self.mode)
- self.mq_producer = AsyncMQProducer(topic_name="topic_crawler_etl_prod_v2", platform=self.platform, mode=self.mode)
- self.method = self.platform_config.method.upper()
- self.url = self.platform_config.url
- self.headers = self.platform_config.headers or {}
- self.request_body_template = self.platform_config.request_body or {}
- self.response_parse = self.platform_config.response_parse or {}
- self.next_cursor_path = self.response_parse.get("next_cursor")
- self.data_path = self.response_parse.get("data_path")
- self.field_map = self.response_parse.get("fields", {})
- self.loop_times = self.platform_config.loop_times or 100
- self.loop_interval = self.platform_config.loop_interval or 5
- self.logger.info(f"开始{self.platform}爬取,最大循环次数{self.loop_times},循环间隔{self.loop_interval}s")
- self.feishu_sheetid = self.platform_config.feishu_sheetid
- self.db_service = AsyncMysqlService(platform=self.platform, mode=self.mode)
- self.request_client = AsyncRequestClient(logger=self.logger, aliyun_log=self.aliyun_log)
- self.timeout = self.platform_config.request_timeout or 30
- self.max_retries = self.platform_config.max_retries or 3
- # 当前分页游标,默认空字符串,支持动态替换request_body中任何字段(如cursor)
- self.dynamic_params = {key: "" for key in self.request_body_template.keys()}
- # 允许子类重写,支持多游标等复杂情况
- self.current_cursor = ""
- self.download_cnt = 0
- self.limit_flag = False
- async def run(self):
- """ 爬虫主流程 """
- await self.before_run()
- total_success, total_fail = 0, 0
- async with aiohttp.ClientSession(timeout=aiohttp.ClientTimeout(total=self.timeout)) as session:
- for loop_index in range(self.loop_times):
- if not await self.is_video_count_sufficient():
- self.logger.info(f"视频抓取数量已达上限,提前结束")
- return
- succ, fail = await self.run_single_loop(session)
- total_success += succ
- total_fail += fail
- await self._wait_for_next_loop(loop_index + 1)
- self.logger.info(f"爬虫完成 成功:{total_success} 失败:{total_fail}")
- async def run_single_loop(self, session) -> (int, int):
- """
- 单轮请求与处理
- """
- success_count, fail_count = 0, 0
- try:
- # 爬取数据
- videos = await self.crawl_data(session)
- if not videos:
- self.logger.info(f"无数据返回,停止本轮")
- return success_count, fail_count
- for video in videos:
- # 依赖接口请求
- video_obj = await self.fetch_dependent_data(video)
- res = await self.process_and_push_video(video_obj)
- if res:
- success_count += 1
- else:
- fail_count += 1
- if self.limit_flag:
- break
- await self.after_run()
- except Exception as e:
- self.logger.exception(f"运行异常: {e}")
- return success_count, fail_count
- async def fetch_dependent_data(self, video: Dict) -> Dict:
- """
- 可在子类重写以实现依赖请求,用返回结果补充原有 video。
- 默认不做处理。
- """
- return video
- async def crawl_data(self, session) -> Optional[List[Dict]]:
- """
- 请求接口,自动渲染动态参数,自动更新游标
- 支持单请求和多请求(分页)逻辑。
- """
- # 动态渲染请求体
- # resolved_body = self._render_request_body()
- # 发送请求
- response = await self.request_client.request(
- session=session,
- method=self.method,
- url=self.url,
- headers=self.headers,
- json= self.dynamic_params
- )
- if not response:
- self.logger.error(f"响应为空")
- return []
- # 更新游标(支持动态参数更新)
- if self.next_cursor_path:
- next_cursor = safe_extract(response, self.next_cursor_path) or ""
- self._update_cursor(next_cursor)
- # 解析数据列表
- data_list = safe_extract(response, self.data_path)
- if not data_list:
- self.logger.info(f"未获取到有效数据")
- return []
- return data_list
- def _render_request_body(self) -> Dict:
- """
- 用当前动态参数渲染请求体模板,支持多参数动态替换
- """
- body = {}
- for k, v in self.request_body_template.items():
- if isinstance(v, str) and v.startswith("{{") and v.endswith("}}"):
- key = v.strip("{} ")
- body[k] = self.dynamic_params.get(key, "")
- else:
- body[k] = v
- return body
- def _update_cursor(self, cursor_value: str):
- """
- 更新分页游标并动态参数,方便下一次请求使用
- """
- self.current_cursor = cursor_value
- # 如果配置的游标字段在请求体中,更新动态参数
- if "cursor" in self.dynamic_params:
- self.dynamic_params["cursor"] = cursor_value
- async def process_and_push_video(self, video: Dict[str, Any]) -> bool:
- """
- 数据处理完整流程(字段映射 -> 校验 -> 推送)
- 子类可重写 process_video 或 filter_data 来定制处理和校验逻辑
- """
- try:
- video_obj = await self.process_video(video)
- if not video_obj:
- return False
- if not await self.filter_data(video_obj):
- return False
- await self.integrated_video_handling(video_obj)
- pushed = await self.push_to_etl(video_obj)
- # 达到下载上限,停止继续抓取
- if self.rule_dict.get("videos_cnt", {}).get("min") and \
- self.download_cnt >= self.rule_dict["videos_cnt"]["min"]:
- self.limit_flag = True
- if pushed:
- self.download_cnt += 1
- return pushed
- except Exception as e:
- self.logger.exception(f"视频处理异常: {e}")
- return False
- async def process_video(self, video: Dict) -> Optional[Dict]:
- """
- 统一字段抽取及 VideoItem 初始化
- 子类可重写或扩展以定制字段映射、过滤等
- """
- self.logger.debug(f"处理视频数据: {video.get('title', '无标题')}")
- publish_user = None
- if self.user_list:
- import random
- publish_user = random.choice(self.user_list)
- else:
- publish_user = {"uid": "default", "nick_name": "default_user"}
- item_kwargs = extract_fields(video, self.field_map, logger=self.logger,aliyun_log=self.aliyun_log)
- item_kwargs.update({
- "user_id": publish_user.get("uid"),
- "user_name": publish_user.get("nick_name"),
- "platform": self.platform,
- "strategy": self.mode
- })
- try:
- item = VideoItem(**item_kwargs)
- video_dict = await item.produce_item()
- if not video_dict:
- self.logger.warning(f"VideoItem 校验失败")
- return None
- return video_dict
- except Exception as e:
- self.logger.error(f"VideoItem 初始化失败: {e}")
- return None
- async def filter_data(self, video: Dict) -> bool:
- """
- 数据校验过滤,默认使用 PiaoQuanPipeline
- 子类可重写此方法实现自定义过滤
- """
- pipeline = PiaoQuanPipeline(
- platform=self.platform,
- mode=self.mode,
- rule_dict=self.rule_dict,
- env=self.env,
- item=video,
- trace_id=self.platform + str(uuid.uuid1())
- )
- return await pipeline.process_item()
- async def integrated_video_handling(self, video: Dict) -> None:
- """
- 钩子函数:可在此实现自动生成标题或其他业务逻辑
- """
- await generate_titles(self.feishu_sheetid, video)
- async def push_to_etl(self, video: Dict) -> bool:
- """
- 推送消息到消息队列ETL
- """
- try:
- await self.mq_producer.send_msg(video)
- self.logger.info(f"成功推送视频至ETL")
- return True
- except Exception as e:
- self.logger.exception(f"推送ETL失败: {e}")
- return False
- async def is_video_count_sufficient(self) -> bool:
- """
- 判断当天抓取的视频是否已达到上限,达到则停止继续抓取
- """
- max_count = self.rule_dict.get("videos_cnt", {}).get("min", 0)
- if max_count <= 0:
- return True
- async with AsyncMysqlService(self.platform, self.mode) as mysql:
- current_count = await mysql.get_today_videos()
- if current_count >= max_count:
- self.logger.info(f"{self.platform} 今日视频数量已达上限: {current_count}")
- self.aliyun_log.logging(code="1011", message="视频数量达到当日最大值", data=f"<今日视频数量>{current_count}")
- return False
- return True
- async def _wait_for_next_loop(self, current_loop: int) -> None:
- """等待下次循环"""
- if current_loop < self.loop_times and self.loop_interval > 0:
- self.logger.info(f"等待 {self.loop_interval} 秒后进行下一次请求")
- await asyncio.sleep(self.loop_interval)
- async def before_run(self):
- """运行前预处理钩子,子类可覆盖"""
- pass
- async def after_run(self):
- """运行后处理钩子,子类可覆盖"""
- pass
|