Search and replace using Regular Expressions in Sublime Text: replace text before search text

Question

Update 2:

I swear I checked on YouTube, and I was able to go to a video with &t=h5m10s and it would work as 0:05:10. If you need this to be 5m10s or 0h5m10s, you'll need to do two different expressions (since Sublime Text won't let you do complex callback functions on your regex replace).

To replace h:mm:ss:

href="(.*?)"(.*?)>(\d*):(\d{2}):(\d{2})

href="$1&t=$3h$4m$5s"$2>$3:$4:$5

To replace m:ss:

href="(.*?)"(.*?)>(\d{1,2}):(\d{2})(?!:\d)

href="$1&t=$3m$4s"$2>$3:$4

The only unusual syntax that needs explanation is (?!:\d). The ?! syntax is for a "negative lookahead". This means that the match group will not be followed by an additional :\d, otherwise this would match hours/minutes/seconds.

Update:

Here is an updated expression for matching optional hours. I double checked and you are able to use &t=h##m##s (i.e., a blank hour if it is not matched). This means we can do one search/replace and just have a blank capture for the hour if it does not exist (you'll see what I mean in a second).

Match this:

href="(.*?)"(.*?)>((?:(\d*):)?(\d{1,2}):(\d{2}))

And replace with:

href="$1&t=$4h$5m$6s"$2>$3

Things got a little more complicated here. Lets take a closer look at the part that changes:

(            # start an overarching capture group
  (\d*):?    # capture 0+ digits followed by an optional : (hours)
  (\d{1,2}): # capture 1-2 digits followed by a : (minutes)
  (\d{2})    # capture the last 2 digits (seconds)
)            # end the capture group

Now you may wonder why I used that overarching capture group. This is because we need a way to keep the contents of the anchor link the same. If we didn't use this, we would have to replace it with hour:minute:second and hour may be blank (meaning 3:45 would be replaced with :3:45).

You can match this:

href="(.*?)"(.*?)>(\d)+:(\d{2})

And replace it with

href="$1&t=$3m$4s"$2>$3:$4

How this works is it first finds any anchor's href, while capturing the contents in the first capture group. Then we capture everything up to the > in a second capture group. Next we capture 1+ digit (third capture), followed by :, and finally capturing the last 2 digits (fourth capture).

We essentially replace this with all the same content, but make sure to append &t=$3m$4s to the end of the link. You can see a working example on Regex101.