feat: 이벤트 로깅 시스템 추가 및 주요 컴포넌트 로깅 통합

This commit is contained in:
2025-08-28 12:24:55 +09:00
parent f0ddc5aebe
commit 4f3611e45d
9 changed files with 306 additions and 7 deletions

View File

@@ -3,6 +3,7 @@ from bs4 import BeautifulSoup
import json
import time
import os
from event_logger import get_logger
try:
from selenium import webdriver
@@ -64,7 +65,10 @@ class WebScraper:
"""
웹사이트에서 정보를 수집합니다.
"""
logger = get_logger()
try:
if logger:
logger.log_event("scrape_start", url=url)
if self.use_selenium and self.driver is not None:
self.driver.get(url)
time.sleep(self.delay)
@@ -88,9 +92,13 @@ class WebScraper:
'content': text_content[:5000],
'timestamp': time.time()
}
if logger:
logger.log_event("scrape_done", url=url, title=title, size=len(data['content']))
return data
except Exception as e:
print(f"스크래핑 실패: {url} - {e}")
if logger:
logger.log_event("scrape_error", url=url, error=str(e))
return None
def crawl_multiple_pages(self, start_urls, keywords=None):