catdog2/urlbot-native-trex
1
0
Fork 0
mirror of http://aero2k.de/t/repos/urlbot-native.git synced 2017-09-06 15:25:38 +02:00
Code Issues Releases Activity

recognize image/* as content-type

Browse Source
This commit is contained in:
urlbot
2014-07-27 12:21:32 +02:00
parent 0f45024f92
commit 5ee9628edd
1 changed files with 20 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
eventlooper.py
View File

@@ -39,18 +39,22 @@ def fetch_page(url):
response = urllib.urlopen(url) response = urllib.urlopen(url)
html = response.read(BUFSIZ) # ignore more than BUFSIZ html = response.read(BUFSIZ) # ignore more than BUFSIZ
response.close() response.close()
return html return (html, response.headers)
except IOError as e: except IOError as e:
logger('warn', 'failed: ' + e.errno) logger('warn', 'failed: ' + e.errno)
def extract_title(url): def extract_title(url):
logger('info', 'extracting title from ' + url) logger('info', 'extracting title from ' + url)
html = fetch_page(url) (html, headers) = fetch_page(url)
if html: if html:
if 'content-type' in headers:
if 'text/' != headers['content-type'][:len('text/')]:
return (1, headers['content-type'])
result = re.match(r'.*?<title.*?>(.*?)</title>.*?', html, re.S|re.M) result = re.match(r'.*?<title.*?>(.*?)</title>.*?', html, re.S|re.M)
if result: if result:
return result.groups()[0] return (0, result.groups()[0])
def chat_write(message, prefix='/say '): def chat_write(message, prefix='/say '):
if debug_enabled(): if debug_enabled():
@@ -89,12 +93,21 @@ def extract_url(data):
if ratelimit_exceeded(): if ratelimit_exceeded():
return False return False
title = extract_title(r) (status, title) = extract_title(r)
if title: if 0 == status:
message = 'Title: %s: %s' % (title.strip(), e(r)) if title:
message = 'Title: %s: %s' % (title.strip(), e(r))
else:
message = 'some error occurred when fetching %s' % e(r)
else: else:
message = 'some error occurred when fetching %s' % e(r) # of course it's fake, but it looks interesting at least
char = """,._-+=\|/*`~"'"""
message = 'No text but %s, 1-bit ASCII art preview: [%c] %s' %(
e(title),
char[int(time.time() % len(char))],
e(r)
)
message = message.replace('\n', '\\n') message = message.replace('\n', '\\n')