elsanctum/the-great-escape
2
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

add extra post & user info

Browse Source
This commit is contained in:
Adrian Malacoda 2016-11-27 00:48:55 -06:00
parent 71a4f8c5a4
commit 5bcb6e8884
1 changed files with 10 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
tge/scrapers/yuku.py
View File

@ -65,7 +65,9 @@ def scrape_thread (url):
# returns the rest of the thread's contents instead of just that post. # returns the rest of the thread's contents instead of just that post.
# So we need to pick out only the first (username/signature/postbody) # So we need to pick out only the first (username/signature/postbody)
# to get around this. # to get around this.
date_element = post_entry.find(".date").eq(0)
post_content_container = post_entry.find(".post-content-container").eq(0) post_content_container = post_entry.find(".post-content-container").eq(0)
user_header = post_entry.find("header").eq(0)
signature = post_content_container.find(".signature").eq(0) signature = post_content_container.find(".signature").eq(0)
post_content_container.remove(".signature") post_content_container.remove(".signature")
@ -75,7 +77,14 @@ def scrape_thread (url):
signature = None signature = None
thread.children.append(Post( thread.children.append(Post(
author=User(name=post_entry.find("header > p > a").eq(0).text(), signature=signature), author=User(
name=user_header.find("p > a").eq(0).text(),
avatar=user_header.find("img[alt='avatar']").attr.src,
title=user_header.find(".auto-title").text(),
subtitle=user_header.find(".custom_title").text(),
signature=signature
),
timestamp=date_element.text(),
body=post_content_container.html().strip() body=post_content_container.html().strip()
)) ))