diff options
Diffstat (limited to 'commcrawler/scrapy.py')
| -rw-r--r-- | commcrawler/scrapy.py | 7 |
1 files changed, 3 insertions, 4 deletions
diff --git a/commcrawler/scrapy.py b/commcrawler/scrapy.py index 0967650..490142c 100644 --- a/commcrawler/scrapy.py +++ b/commcrawler/scrapy.py @@ -89,9 +89,6 @@ class DefaultSpider: if "src" not in attributes: continue src = attributes["src"] - is_a_real_src = src.startswith("http") or src.startswith("/") - if not src or not is_a_real_src: - continue current_domain = get_domain(src) if current_domain in YOUTUBE_DOMAINS: append_to_results(result, "youtube", src) @@ -196,7 +193,9 @@ class DefaultSpider: yield response.follow(link.url, self.parse) if not is_internal: current_domain = get_domain(url) - if current_domain in FACEBOOK_DOMAINS: + if not current_domain: + pass + elif current_domain in FACEBOOK_DOMAINS: append_to_results(result, "facebook", url) elif current_domain in TWITTER_DOMAINS: append_to_results(result, "twitter", url) |