Matched Pattern Returns Tuple Of Strings Instead Of A Match As A Whole

Question

I have a regular expression to fetch some links in HTML document.

((http://)(|up)(\.example\.com))*(/uploads/pp2p|/sites/default/files/[-_a-zA-Z0-9%/]+)\.(jpg|jpeg|gif|png)

What I am intending to match is, if the http part exists match it if not, don't. If up part exists match it if not, don't. If example.com exists match it if not, don't. The same about /uploads/pp2p and the other one, if exists match if not, don't. Finally, if it has one of the following image formats match it if not, don't. I expect to get a list of links like

links = ['http://up.example.com/uploads/pp2p/www.jpg', '/sites/default/files/.png', 'http://example.com/uploads/zzz.jpg']

And the elements in the link continue to be filled with different combinations. Anyway, I am getting results as a tuple like

[('', '', '', '', '/sites/default/files/favicon', 'png'), ('', '', '', '', '/sites/default/files/logo_2', 'png')]

I don't want to get a tuple, I want the match to be represented as a whole. Only a complete link in each list element. How can I avoid getting a tuple as a result of the Regex match?

I think you need to reorganise the regex to have a single capture group — jonrsharpe, May 27 '15 at 16:54
Try using non-capturing groups: http://stackoverflow.com/questions/3512471/non-capturing-group — flamusdiu, May 27 '15 at 16:57
@flamusdiu Is it possible to turn non-capturing-group off completely? — FreeMind, May 27 '15 at 17:17
Not sure what you mean? Non-capturing groups are not returned. Also, Can you post a example of the source file you are parsing with this? It would make it easier to hel you out. — flamusdiu, May 27 '15 at 17:25
@flamusdiu I mean I just don't want the smaller groups to be returned, I only need the whole match. I am looking for a way to turn off groups globally. — FreeMind, May 27 '15 at 17:32

flamusdiu · Accepted Answer · 2015-05-27T21:10:47.183

I am assuming you are getting images off a web page somewhere.

Here's a quick way to grab all the image src links using lxml.html:

from lxml.html import parse
import re

doc = parse('http://www.androidpolice.com').getroot()
links = []
img_list = []

for img in doc.cssselect('img'):
    links.append(img.get('src'))

for link in links:
    match = re.search(".*androidpolice\.com.*",link)
    if match:
        img_list.append(match.group(0))

for img in img_list:
    print(img)

Gives the output:

http://www.androidpolice.com/wp-content/themes/ap2/ap_resize/ap_resize.php?src=http%3A%2F%2Fwww.androidpolice.com%2Fwp-content%2Fuploads%2F2015%2F05%2Fnexus2cee_gamethumb_thumb1.png&h=128&zc=3
http://www.androidpolice.com/wp-content/themes/ap2/ap_resize/ap_resize.php?src=http%3A%2F%2Fwww.androidpolice.com%2Fwp-content%2Fuploads%2F2015%2F05%2Fnexus2cee_gamethumb_thumb1.png&w=150&h=75&f=8|8|8|8|8|8|8|8|8|8|8|8|8
http://www.androidpolice.com/wp-content/themes/ap2/ap_resize/ap_resize.php?src=http%3A%2F%2Fwww.androidpolice.com%2Fwp-content%2Fuploads%2F2014%2F06%2Fnexusae0_Google-Photos-icon-logo-150x150.png&h=128&zc=3

-----[output truncated]-----

Then you could do (?:%2F)([\w-]+\.(?:png|jpg)) to get all the image names (of course just an example). I.E. nexus2cee_gamethumb_thumb1.png

Updated Code

Changed it to search only for androidpolice.com in each link. You can find more on using re module at 6.2. re — Regular expression operations.

Pretty nice way. Is not there a way to do regex on SRC tags? restricting it to special src tags which start for instance by, `example.com` — FreeMind, May 27 '15 at 19:32

Matched Pattern Returns Tuple Of Strings Instead Of A Match As A Whole

1 Answers1