Regex Expression to remove "autoplay" parameter in url - regex

Question

I'm trying to match the url https://youtube.com/embed/id and its parameters i.e ?start=10&autoplay=1, but I need the autoplay parameter removed or set to 0.
These are some example urls and what I want the results to look like:
http://www.youtube.com/embed/JW5meKfy3fY?autoplay=1
I want to remove the autoplay parameter and its value:
http://www.youtube.com/embed/JW5meKfy3fY
2nd example
http://www.youtube.com/embed/JW5meKfy3fY?start=10&autoplay=1
results should be
http://www.youtube.com/embed/JW5meKfy3fY?start=10
I have tried (https?:\/\/www.youtube.com\/embed\/[a-zA-Z0-9\\-_]+)(\?[^\t\n\f\r \"']*)(\bautoplay=[01]\b&?) and replace with $1$2, but it matches with a trailing ? and & in example 1 and 2 respectively. Also, it doesn't match at all for a url like
http://www.youtube.com/embed/JW5meKfy3fY
I have the regex and examples on here
NB:
The string I am working on contains HTML with one or more youtube urls in it, so I don't think I can easily use go's net/url package to parse the url.

Answer 1

You're asking for a regex but I think you'd be better off using Go's "net/url" package. Something like this:
import "net/url"
//...
u, _ := url.Parse("http://www.youtube.com/embed/JW5meKfy3fY?start=10&autoplay=1")
q := u.Query()
q.Del("autoplay")
u.RawQuery = q.Encode()
clean_url_string = u.String()
In real life you'd want to handle errors from u.Parse of course.

Answer 2

Here's a solution that ensures a valid page URI. Simply match this and only return capture group 1 and 3.
Edit: The pattern is not elegant but it ensures no stale ampersands stay. The previous solution was more elegant and albeit wouldn't break anything, isn't worth the tradeoff imo.
Pattern
(https?:\/\/www\.youtube\.com\/embed\/[^?]+\?.*)(&autoplay=[01]|autoplay=[01]&?)(.*)
See the demo here.

Answer 3

As the OP has linked to a regex tester that employs the the PCRE (PHP) engine I offer a PCRE-compatible solution. The one token I've used in the regular expression below that is not widely supported in other regex engines is \K (though it is supported by Perl, Ruby, Python's PyPI regex module, R with Perl=TRUE and possibly other engines.
\K causes the regex engine to reset the beginning of the match to the current location in the string and to discard any previously-matched characters in the match it returns (if there is one).
With one caveat you can replace matches of the following regular expression with empty strings.
(?x) # assert 'extended'/'free spacing' mode
\bhttps?:\/\/www.youtube.com\/embed\/
# match literal
(?=.*autoplay=[01]) # positive lookahead asserts 'autoplay='
# followed by '1' or '2' appears later in
# the string
[a-zA-Z0-9\\_-]+ # match 1+ of the chars in the char class
[^\t\n\f\r \"']* # match 0+ chars other than those in the
# char class
(?<![?&]) # negative lookbehind asserts that previous
# char was neither '?' nor '&'
(?: # begin non-capture group
(?=\?) # use positive lookahead to assert next char
# is a '?'
(?: # begin a non-capture group
(?=.*autoplay=[01]&)
# positive lookahead asserts 'autoplay='
# followed by '1' or '2', then '&' appears
# later in the string
\? # match '?'
)? # end non-capture group and make it optional
\K # reset start of match to current location
# and discard all previously-matched chars
\?? # optionally match '?'
autoplay=[01]&? # match 'autoplay=' followed by '1' or '2',
# optionally followed by '&'
| # or
(?=&) # positive lookahead asserts next char is '&'
\K # reset start of match to current location
# and discard all previously-matched chars
&autoplay=[01]&? # match '&autoplay=' followed by '1' or '2',
# optionally followed by '&'
) # end non-capture group
The one limitation is that it fails to match all instances of .autoplay=.. if more than one such substring appears in the string.
I wrote this expression with the x flag, called extended or free spacing mode, to be able to make it self-documenting.
Start your engine!