How to combine Ruby regexp conditions - ruby-on-rails

Question

I need to check if a string is valid image url.
I want to check beginning of string and end of string as follows:
Must start with http(s):
Must end by .jpg|.png|.gif|.jpeg
So far I have:
(https?:)
I can't seem to indicate beginning of string \A, combine patterns, and test end of string.
Test strings:
"http://image.com/a.jpg"
"https://image.com/a.jpg"
"ssh://image.com/a.jpg"
"http://image.com/a.jpeg"
"https://image.com/a.png"
"ssh://image.com/a.jpeg"
Please see http://rubular.com/r/PqERRim5RQ
Using Ruby 2.5

Answer 1

Using your very own demo, you could use
^https?:\/\/.*(?:\.jpg|\.png|\.gif|\.jpeg)$
See the modified demo.
One could even simplify it to:
^https?:\/\/.*\.(?:jpe?g|png|gif)$
See a demo for the latter as well.
This basically uses anchors (^ and $) on both sides, indicating the start/end of the string. Additionally, please remember that you need to escape the dot (\.) if you want to have ..
There's quite some ambiguity going on in the comments section, so let me clarify this:
^ - is meant for the start of a string
(or a line in multiline mode, but in Ruby strings are always in multiline mode)
$ - is meant for the end of a string / line
\A - is the very start of a string (irrespective of multilines)
\z - is the very end of a string (irrespective of multilines)

Answer 2

You may use
reg = %r{\Ahttps?://.*\.(?:png|gif|jpe?g)\z}
The point is:
When testing at online regex testers, you are testing a single multiline string, but in real life, you will validate lines as separate strings. So, in those testers, use ^ and $ and in real code, use \A and \z.
To match a string rather than a line you need \A and \z anchors
Use %r{pat} syntax if you have many / in your pattern, it is cleaner.
Online Ruby test:
urls = ['http://image.com/a.jpg',
'https://image.com/a.jpg',
'ssh://image.com/a.jpg',
'http://image.com/a.jpeg',
'https://image.com/a.png',
'ssh://image.com/a.jpeg']
reg = %r{\Ahttps?://.*\.(?:png|gif|jpe?g)\z}
urls.each { |url|
puts "#{url}: #{(reg =~ url) == 0}"
}
Output:
http://image.com/a.jpg: true
https://image.com/a.jpg: true
ssh://image.com/a.jpg: false
http://image.com/a.jpeg: true
https://image.com/a.png: true
ssh://image.com/a.jpeg: false

Answer 3

The answers here are quite good, but if you wanted to avoid using a complicated regex and communicate your intent more clearly to a reader, you could let URI and File do the heavy lifting for you.
(And since you're using 2.5, let's use #match? instead of other regex-matching methods.)
def valid_url?(url)
# Let URI parse the URL.
uri = URI.parse(url)
# Is the scheme http or https, and does the extension match expected formats?
uri.scheme.match?(/https?/i) && File.extname(uri.path).match?(/(png|jpe?g|gif)/i)
rescue URI::InvalidURIError
# If it's an invalid URL, URI will throw this error.
# We'll return `false`, because a URL that can't be parsed by URI isn't valid.
false
end
urls.map { |url| [url, valid_url?(url)] }
#=> Results in:
'http://image.com/a.jpg', true
'https://image.com/a.jpg', true
'ssh://image.com/a.jpg', false
'http://image.com/a.jpeg', true
'https://image.com/a.png', true
'ssh://image.com/a.jpeg', false
'https://image.com/a.tif', false
'http://t.co.uk/proposal.docx', false
'not a url', false