正则表达式：匹配指定字符串之间的所有匹配项

Question

I'm dealing with a bunch of text files that refer to image filenames. These filenames were sanitized (made lowercase and whitespace replaced with hyphens) - but the text referring to them was not.

I need to transform strings like this:

(image: uploaded IMAGE.jpg caption: this is my caption)
(image: uploaded IMAGE copy.jpeg caption: this is my caption)
(image: IMG_6087.png caption: this is my caption)
(image: IMG_6087 copy.gif)
(image: IMG_9999_copy.jpg)
(image: somehow, a comma.jpg)
(image: other ridic'ulous characters!.jpg)

to:

(image: uploaded-image.jpg caption: this is my caption)
(image: uploaded-image-copy.jpeg caption: this is my caption)
(image: img_6087.png caption: this is my caption)
(image: img_6087-copy.gif)
(image: img_9999_copy.jpg)
(image: somehow-a-comma.jpg)
(image: other-ridiculous-characters.jpg)

These strings are parts of larger blocks of text, but are all on their own lines, like so:

This is not a short guide to write about art. Go in, out of the window, inside New York’s stars qualities, dreams and schemes. People are gathered together, brewing coffee — you have seen their faces? The artists in Manhattan.

(image: manhattan photo.jpg)

Drive till sunset and say goodbye to your body, because this is not a photograph. I saw sixteen americans, raised by wolves, probably lost in paradise city. I found your head — Do you still want it?

I'm using Sublime text and was planning on doing multiple Replace Alls:

1. strip whitespace
2. strip characters that are not alphanumeric or _ or -
3. make lowercase

But I can't manage to capture all instances of something between the two delimiters.

(?<=^\\(image: )[what do I do here??](?=\\.jpe?g|png|gif)

Answer 1

you can use non-greedy match-all .*?

so ^\\(image: (.*?\\.(:?jpe?g|png|gif)) to capture the filename including extension

Answer 2

You can grab the filenames with:

(?<=image:\s)([^.]++)(?=\.jpe?g|\.png|\.gif)

After that, the transformations depend on the language that you're working in. Add file extensions as you need them. Right now you support jpg , jpeg , png , and gif .

Answer 3

Here is a working way to do it in PHP

<?php
$string =
"This is not a short guide to write about art. Go in, out of the window, inside New York’s stars qualities, dreams and schemes. People are gathered together, brewing coffee — you have seen their faces? The artists in Manhattan.

(image: uploaded IMAGE.jpg caption: this is my caption)
This is not a short guide to write about art. Go in, out of the window, inside New York’s stars qualities, dreams and schemes. People are gathered together, brewing coffee — you have seen their faces? The artists in Manhattan.

(image: uploaded IMAGE copy.jpeg caption: this is my caption)
(image: IMG_6087.png caption: this is my caption)
(image: IMG_6087 copy.gif) blah blah
(image: IMG_9999_copy.jpg)
(image: somehow, a comma.jpg)
(image: other ridic'ulous characters!.jpg)";

echo preg_replace_callback('~(?<=\(image: )(.*?)\.(jpg|jpeg|png|gif)~', function($matches)
{
    return preg_replace('~\W~', '-', stripslashes(strtolower($matches[1]))) . ".$matches[2]";
}, $string);

?>

[EDIT] add regex explanation:

• (?<=image: ) : is a positive lookbehind - so checking the presence of 'image: ' but not capturing.
• (.*?) : captures everything before the image extension in a greedy way - so match as few text as possible.
• \\.(jpg|jpeg|png|gif) : will match . literally + one of the given extensions - and capturing the extension to reuse.
• ~ : is the delimiter, this choice just because it is very seldom used in strings and won't need to \\ the /
• \\W : is the opposite of \\w and it will match any non-alphanumeric character.

Will output (in view source):

This is not a short guide to write about art. Go in, out of the window, inside New York’s stars qualities, dreams and schemes. People are gathered together, brewing coffee — you have seen their faces? The artists in Manhattan.

(image: uploaded-image.jpg caption: this is my caption)
This is not a short guide to write about art. Go in, out of the window, inside New York’s stars qualities, dreams and schemes. People are gathered together, brewing coffee — you have seen their faces? The artists in Manhattan.

(image: uploaded-image-copy.jpeg caption: this is my caption)
(image: img_6087.png caption: this is my caption)
(image: img_6087-copy.gif) blah blah
(image: img_9999_copy.jpg)
(image: somehow--a-comma.jpg)
(image: other-ridic-ulous-characters-.jpg)

You can then fine-tune in the callback what character you want to transform into what, with str_replace() for instance.

hope it helps! ;)

Answer 4

Can you try Jetbrains webstrom front end IDE. Which provides lot of capabilities to achieve any regex operations in readable way. Select a text you want to split are check for delimiters or any white-spaces.

You will get it for 30 days trail version . Also will share you the regex query shortly.

Also checkout http://myregexp.com/ or some plugin to valid your regex queries

Online Regex editor