mirror of
https://github.com/RussellChamp/tilde-projects.git
synced 2026-10-28 00:22:34 +00:00
Add files I forgot
This commit is contained in:
1 parent
5f1bf146c3
commit
772995c0ff
29 files changed
+1807
No files matched your search
@@ -0,0 +1,31 @@
|
||||
#!/usr/bin/python3
|
||||
import requests
|
||||
from bs4 import BeautifulSoup
|
||||
import util.duckduckgo
|
||||
|
||||
def xkcd(query):
|
||||
res = util.duckduckgo.get_zci("site:xkcd.com " + query)
|
||||
title = ""
|
||||
try: # ddg returns a url for these searches. i want a title as well
|
||||
title = BeautifulSoup(requests.get(res).content, "html.parser").title.text
|
||||
except:
|
||||
pass # just swallow the error. maybe the result wasn't a url or something else bad happened
|
||||
return [(((title + " - ") if title else "") + res)]
|
||||
|
||||
|
||||
# never mind, blocked by ddg
|
||||
# def xkcd_links(query):
|
||||
# url = "https://duckduckgo.com/html/?q=site%3Axkcd.com+" + query.replace(' ', '+')
|
||||
# soup = BeautifulSoup(requests.get(url).content, 'html.parser')
|
||||
# items = soup.find_all("a", class_="result__a")
|
||||
# print items
|
||||
# items = list(filter(lambda i: i[0:8] == 'xkcd.com',
|
||||
# [i.find(class_="result__title").text.strip() for i in items]))
|
||||
# print items
|
||||
# def pretty_link(item):
|
||||
# url = item.find(class_="result__url").text.strip()
|
||||
# title = item.find(class_="result__title").text.strip()
|
||||
# return (title + ' - ' + url) if title else url
|
||||
#
|
||||
# links = map(lambda url: pretty_link(url), items)
|
||||
# return links
|
||||
Reference in new issue
Block a user