Rhythmbox plugin: A method to sift and remove duplicate songs?

[mihai.ile]

Hello.
This is fantastic!

I am already working on the symlinks problem and I already got it working. just a little polish and I reply back with the solution.

[mihai.ile]

Ok, here it is the same plugin but now you can play the files in the duplicate list.

I also added a diff from your and my version to ease the integration.

Basically I create a hidden folder in ~/ with all symlinks and remove it at the exit (no need to have extra trash on people's filesystem as it is very easy and fast to recreate it next time; we need to search for duplicates again at startup so no big deal anyway)

[mihai.ile]

I also noted that the .rhythmbox-duplicates.ignore file grows with one byte every time you exit rhythmbox, better check it out before we grow into a 5GB .rhythmbox-duplicates.ignore file

[scrawl]

Thanks for your effort! I polished it up a bit so its now included. They were some problems though, the "Remove from library" and "Move to trash" didnt work anymore, because they relied on the location of the entry, but thats fixed now too. And by the way, the last bit is very much simpler than you did it: shutil.rmtree( os.path.expanduser("~/.rhythmbox-duplicates-symlinks/") ) and thats it

So, lets call it version 0.2. Here is the full changelog:

- Fixed: Due to a limitation in Rhythmbox, you can't playback files from the duplicate source
- Fixed: .rhythmbox-duplicates.ignore file slowly increases its file size
- Dont consider duplicate when the 2 tracks are from different albums
- When one of the duplicates is from an album, and the other is from an "Unknown" album, consider the latter one as duplicate
- Rescan for duplicates when an entry gets removed from the library
- Prefer files that have been added earlier when marking as duplicate


Download: http://scrawl.bplaced.net/duplicate-source.tar.gz

[mihai.ile]

Yes, it's working good right now.
I was wondering if we need to create the path structure from the original file inside rhythmbox-duplicates-symlinks folder, because for example if I want to see the properties of the file, to see where is located, it's impossible to tell. Maybe we should make the exact path of the original file (including it's name) but just inside our rhythmbox-duplicates-symlinks folder, so if needed one can always have a clue from where in his music library that duplicate is coming from.

Oh and we have to figure out how we can get information about the original file... because sometimes you just need to liste/check properties of both original and the duplicate to decide which to remove.

Another thing... I think it's great not to have a menu entry, the plugin should just work, and now that is really fast in looking for duplicates, it could run in background, just like the missing files source. If there is a duplicate show the source, if not... hide it (oh and with this hide/show thing... I noticed that at startup the source appears, then right after dissapears to appear again a little after)

I think a that quite high quality functionality in rhythmbox is comming from this plugin

[scrawl]

I was wondering if we need to create the path structure from the original file inside rhythmbox-duplicates-symlinks folder, because for example if I want to see the properties of the file, to see where is located, it's impossible to tell. Maybe we should make the exact path of the original file (including it's name) but just inside our rhythmbox-duplicates-symlinks folder, so if needed one can always have a clue from where in his music library that duplicate is coming from.

The solution is much simpler: Don't show the properties dialog of the entry in the duplicate source, but of the entry in the library source. /Edit: patched, it works now.

Oh and we have to figure out how we can get information about the original file... because sometimes you just need to liste/check properties of both original and the duplicate to decide which to remove.

Yeah, I was thinking of this too, but I'm not sure how this should be implemented. A very simple solution would be a "Show versions" button that just switches to the library source and enters song name and artist into the search bar.

(oh and with this hide/show thing... I noticed that at startup the source appears, then right after dissapears to appear again a little after)

Well, it happens like this, first, I create the source (show), then I tell Rhythmbox to hide it when its empty (hide), then the source fills up (show). I have no idea what we could do about it, but thats very minor, isnt it

By the way, there was another bug coming with the playback thing, marking as non-duplicate didnt work properly, but I've fixed it. If anyone runs across that problem just download it again from here: http://scrawl.bplaced.net/duplicate-source.tar.gz

And another thing, I will set up a launchpad page very soon where you can always get the latest source, submit code, view bugs, and translate the plugin (hooray!!).

[Mr Nemo]

Very cool, I'm trying this out now.

[mihai.ile]

Great, Do you think it would be acceptable to have an option for the plugin where you can check to show the original file in the duplicates list, next to its duplicates? this way we could have the original and it's duplicate(s) and you could easily choose to play all of them and remove the ones you don't want.

[scrawl]

Hey, thats actually a great idea I was really wondering how to implement that "show all versions" thingy, but this seems to be a very good solution.

[Mr Nemo]

How would I go about installing this? If its possible at this point.