Script for wget to log name of downloaded file? [SOLVED]

[audiodef]

Is it possible to write a script to record the filename of a file wget downloads with --trust-server name? For example, if

Code: Select all

wget --trust-server-name http://electro-music.com/forum/download.php?id=28985

downloads adventures_in_sound_117_110112_454.mp3, how can I get a script to capture the filename "adventures_in_sound_117_110112_454.mp3"?

[Quincy]

Perhaps I'm going for the wrong point, but will this do your job?

Code: Select all

wget --trust-server-name http://electro-music.com/forum/download.php?id=28985 2>&1 | grep -o -E "»».+««" | tr -d '»«'

It redirects wget's standard error to standard output, scans this for a line with the characters enclosing the file name (in my case/localization ""»»" and "««") and outputs only this to tr which removes the enclosing characters.

[khayyam]

audiodef ...

Its not entirely clear if you just want the file name, or if you want also to download the file. Anyhow, I'd probably approach it in a similar way to Quincy, though using '-o log' to be sure that we actually get to the 'save', and not simply take whatever comes from stdout.

Code: Select all

$ wget --trust-server-names=on -o log \
    http://electro-music.com/forum/download.php?id=28985 && \
    awk -F"'" '/saved/{print $2}' log

best ... khay

[audiodef]

Thanks, guys. This is the kind of thing I'm looking for. I only need to get the filename. I like what Quincy did - it outputs the filename, and I like khayyam's code that puts it in a file. I tried adding > downloaded_file to Quincy's code, but > doesn't seem to work with wget. What's the best way to do that?

[khayyam]

I only need to get the filename.

audiodef ... in which case both Quincy and my suggestions are off target as both will wget the file. To provide the filename, without downloading the file, '--spider' should be used.

I like what Quincy did - it outputs the filename, and I like khayyam's code that puts it in a file. I tried adding > downloaded_file to Quincy's code, but > doesn't seem to work with wget. What's the best way to do that?

Well, the following will print the filename to stdout, stdout can then be redirected to a file if need be.

Code: Select all

$ wget --trust-server-names=on --spider -o log \
    http://electro-music.com/forum/download.php?id=28985 && \
    awk -F"/" '/Location/{print substr($NF, 1, length($NF) - 12)}' log

best ... khay

[Quincy]

I have to correct my code a bit because I checked again and it would output the name twice when the download finishes (I didn't load the full file last time). Therefore a better solution (with actually downloading the file) would be to just use the short output of wget and then redirect as needed. But this does unfortunately not work with spider mode:

Code: Select all

wget -nv --trust-server-namehttp://electro-music.com/forum/download.php?id=28985 2>&1 | grep -o -E "\".+\"" | tr -d '"' > downloaded_file

[audiodef]

Thanks again, khay and Quincy. This is exactly what I was looking for!