How do I parse a string to find a URL?

[Jayson Williams]

Is there a command in Ruby that will accept a string, and spit out a
URL that is contained in the string? I think I remember reading about
something that would do this, but I cant recall.

 

[Jano Svitok]

Is there a command in Ruby that will accept a string, and spit out a
URL that is contained in the string? I think I remember reading about
something that would do this, but I cant recall.

URI::extract

http://ruby-doc.org/core/classes/URI.html#M004839

 

[Jayson Williams]

Outstanding!
Thanks

URI::extract

http://ruby-doc.org/core/classes/URI.html#M004839

 

[Daniel DeLorme]

URI::extract

http://ruby-doc.org/core/classes/URI.html#M004839

Wow, I didn't know about that, very nice. But it has a few weaknesses:=> ["behold:", "http://www.xyz.com."]
(notice the period at the end of xyz.com)

Daniel

 

[franco]

http://ruby-doc.org/core/classes/URI.html#M004839

Wow, I didn't know about that, very nice. But it has a few weaknesses:=> ["behold:", "http://www.xyz.com."]
(notice the period at the end of xyz.com)

Daniel

not a weakness, in that string 'behold:' is a valid uri, it has a
scheme with a scheme delimeter (":"). "www.abc.com" is not an
unambiguous uri, no scheme present.

http://en.wikipedia.org/wiki/Uniform_Resource_Identifier

 

[franco]

http://ruby-doc.org/core/classes/URI.html#M004839

Wow, I didn't know about that, very nice. But it has a few weaknesses:=> ["behold:", "http://www.xyz.com."]
(notice the period at the end of xyz.com)

Daniel

also the period is legal,

 

[Daniel DeLorme]

not a weakness, in that string 'behold:' is a valid uri, it has a
scheme with a scheme delimeter (":"). "www.abc.com" is not an
unambiguous uri, no scheme present.

Is it a valid uri if nothing is present after the scheme? Anyway, I know
that the results are technically valid but they are less than useful if
you want, say, to extract and "linkify" urls that users might have
written inside a message. (which is what I assumed the OP wanted but I
might have been mistaken)

Daniel

 

[franco]

Is it a valid uri if nothing is present after the scheme? Anyway, I know
that the results are technically valid but they are less than useful if
you want, say, to extract and "linkify" urls that users might have
written inside a message. (which is what I assumed the OP wanted but I
might have been mistaken)

Daniel

you could just select the ones with a scheme scpecific part? or screen
scrape for //a/@href to get all hyperreferenced anchors (links).