Shell Programming and Scripting

I've been attempting to use curl and sed to allow for downloading a file from a dynamically generated URL. I've been able to retrieve and save the HTML of the page that does the dynamic generation of the download URL using curl but I'm very new to sed and I seem to be stuck at this part.

HTML:

Code:

...Irrelevant code. Please <a href="/download.cgi?key=4589438949843895465657">download this file.</a> ...Irrelevant code...

I'm looking to then download the file using curl like this:

Code:

curl --location -O http://example.com/download.cgi?key=4589438949843895465657

Any suggestions?

First thing I'd do is

Code:

wget -k "http://site.com/thing-containing-url.html" -O urlfile.html

the -k will cause it to convert a local link into http://...

Then you can

Code:

URL=`egrep -o 'http://[^?"]*key=[^"]*' urlfile.html`
echo "Got URL $URL"
wget "$URL"