def parse_news(self, response):
news = response.meta["news"]
if "window.location.replace" in response.body:
news["crawl_url"] = news["original_url"]
news["key"] = g_cache_key(news["crawl_url"])
yield self.g_news_request(news)
else:
redirects = response.request.meta.get("redirect_urls")
if redirects:
news["crawl_url"] = response.url
news["key"] = g_cache_key(news["crawl_url"])
body = response.body_as_unicode().encode("utf-8")
if news["crawl_url"].startswith("http://www.yidianzixun.com/"):
extractor = YiDianZiXunExtractor(body, response.url)
title, post_date, post_user, summary, content = extractor()
else:
try:
title, post_date, post_user, summary, tags, content = extract(
news["crawl_url"], document=body)
except Exception as e:
self.logger.warning(e.message +
" outer link: %s" % news["crawl_url"])
return
if content:
news["content"] = content
news["content_html"] = response.body
yield news
else:
self.logger.warning("content empty: %s" % news["crawl_url"])
def parse_item(self, response):
body = response.body_as_unicode().encode("utf-8")
extractor = GeneralExtractor(body)
title, post_date, post_user, summary, content = extractor(
self.title_param, self.post_date_param, self.post_user_param,
self.summary_param, self.content_param)
if not post_user:
post_user = self.crawl_source
news = get_default_news(
crawl_url=response.url,
key=g_cache_key(response.url),
title=title,
publish_time=post_date,
original_source=post_user,
original_url=response.url,
content=content,
crawl_source=self.crawl_source,
)
print("*" * 50)
print("url: %s" % response.url)
print("title: %s" % title)
print("post date: %s" % post_date)
print("post user: %s" % post_user)
print("summary: %s" % summary)
show(content)
print("\n\n")
def g_news_item(self, article, start_url="", meta=None):
news = NewsItem()
news["docid"] = article["docID"]
url_163 = article.get("url_163", None)
if url_163 is None: return None
news["crawl_url"] = self._g_crawl_url(url_163)
news["key"] = g_cache_key(news["crawl_url"])
if news_already_exists(news["key"]): return None
news["title"] = article["title"]
news["tags"] = list()
news["summary"] = article["summary"]
news["publish_time"] = str_from_timestamp(article["publish_time"] /
1000)
news["content"] = list()
position = article.get("position", "null,null,null,null").split(",")
news["province"] = position[0] if position[0] != "null" else None
news["city"] = position[1] if position[1] != "null" else None
news["district"] = position[2] if position[2] != "null" else None
news["love"] = 0
news["up"] = 0
news["down"] = 0
news["original_url"] = article.get("doc_url", "")
news["channel"] = article.get("channel", "/").split("/")[0]
news["category"] = article.get("category", "")
news["crawl_source"] = CRAWL_SOURCE
news["original_source"] = article.get("source", "")
if news['original_source'] == u'糗事百科':
return None
news["comment_url"] = self._g_comment_url(docid=news["docid"])
news["comment_queue"] = COMMENT_SPIDER_NAME + ":start_urls"
news["start_url"] = start_url
news["start_meta_info"] = meta
return news
def g_news_item(self, article, start_url="", meta=None):
if article.get("has_video"):
return None
docid = article["source_url"]
crawl_url = self._g_crawl_url(article)
key = g_cache_key(crawl_url)
if news_already_exists(key):
return None
news = get_default_news(
crawl_url=crawl_url,
key=key,
title=article["title"],
tags=article.get("keywords", "").split(","),
summary=article.get("abstract", ""),
publish_time=str_from_timestamp(article["publish_time"]),
love=article.get("favorite_count", 0),
up=article.get("digg_count", 0),
down=article.get("bury_count", 0),
original_url=article.get("url", ""),
original_source=article.get("source", ""),
crawl_source=CRAWL_SOURCE,
start_url=start_url,
start_meta_info=meta,
comment_url=self._g_comment_url(docid),
comment_queue=COMMENT_SPIDER_NAME + ":start_urls")
news["docid"] = news["comment_url"]
return news
def g_news_item(self, article, start_url="", meta=None):
if article["ctype"] not in ["news", "picture"]:
return None # fixme: only support news now
docid = article["docid"]
crawl_url = self._g_article_url(article.get("url"), docid)
if not crawl_url:
return None
key = g_cache_key(crawl_url)
if news_already_exists(key):
return None
news = get_default_news(crawl_url=crawl_url,
key=key,
title=article["title"],
summary=article.get("summary", ""),
publish_time=article["date"],
love=article.get("like", 0),
up=article.get("up", 0),
original_url=article.get("url", ""),
crawl_source=CRAWL_SOURCE,
original_source=article.get("source", ""),
start_url=start_url,
start_meta_info=meta,
comment_url=self._g_comment_url(docid),
comment_queue=COMMENT_SPIDER_NAME +
":start_urls")
news["docid"] = news["comment_url"]
return news
def g_news_item(self, article, start_url="", meta=None):
"""
生成 news item 对象
:param article: 包含新闻 url, title 字段的字典
:type article: dict
:param start_url: 抓取 meta info 的起始 url
:type start_url: str
:param meta: 附加配置信息
:type meta: dict
:return: 新闻 Item
:rtype: News.items.NewsItem | None
"""
crawl_url = article["url"]
key = g_cache_key(crawl_url)
if news_already_exists(key):
return None
news = get_default_news(title=article["title"],
crawl_url=crawl_url,
docid=crawl_url,
key=key,
crawl_source=self.crawl_source,
start_url=start_url,
summary=article.get("summary", ""),
start_meta_info=meta)
return news
def g_news_item(self, article, start_url="", meta=None):
docid = article['url']
crawl_url = self._g_article_url(article['url'])
if not crawl_url:
return None
key = g_cache_key(crawl_url)
if news_already_exists(key):
return None
news = get_default_news(crawl_url=crawl_url,
key=key,
title=article['title'],
docid=docid,
start_meta_info=meta,
crawl_source=CRAWL_SOURCE)
return news
def g_news_item(self, article, start_url="", meta=None):
crawl_url = article['crawl_url']
key = g_cache_key(crawl_url)
news = get_default_news(
crawl_url=crawl_url,
key=key,
title=article['title'],
love=article.get("like", 0),
up=article.get("up", 0),
original_url=article.get("url", ""),
crawl_source=CRAWL_SOURCE,
start_url=start_url,
start_meta_info=meta,
docid=crawl_url,
)
return news
def g_news_item(self, article, start_url="", meta=None):
crawl_url = article["crawl_url"]
comment_url = self._g_comment_url(crawl_url)
news = get_default_news(
crawl_url=crawl_url,
docid=comment_url,
key=g_cache_key(article["title"].encode("utf-8")),
crawl_source=CRAWL_SOURCE,
start_url=start_url,
summary=article["summary"],
publish_time=article["publish_time"],
title=article["title"],
start_meta_info=meta,
comment_url=comment_url,
comment_queue=COMMENT_SPIDER_NAME+":start_urls"
)
return None if news_already_exists(news["key"]) else news
def g_news_item(self, article, start_url="", meta=None):
news = NewsItem()
news["docid"] = article["docid"]
news["crawl_url"] = self._g_crawl_url(news['docid'])
news["key"] = g_cache_key(news["crawl_url"])
if news_already_exists(news["key"]): return None
news["title"] = article["title"]
if 'ltitle' in news and news['ltitle']:
news['summary'] = article['ltitle']
news["tags"] = list()
news["publish_time"] = article["ptime"]
news["content"] = list()
news["love"] = 0
news["up"] = 0
news["down"] = 0
news["crawl_source"] = CRAWL_SOURCE
news['original_url'] = ''
news["original_source"] = article.get('source', '')
# news["comment_url"] = self._g_comment_url(docid=news["docid"])
news["start_url"] = start_url
news["start_meta_info"] = meta
return news