#beets

bpd doesn't work with ncmpcpp? :o

marcaddeo: it works on the Debian ncmpcpp

but IIRC it implements an old version of the protocol.

Hi all, I'm getting the following error message when trying to import a new album; Error: "caused by: <urlopen error [Errno -2] Name or service not known>" in release search with query {'release': u'monstercat 019 - endeavour', 'tracks': '32', 'artist': u'various artists'}

I've tried updating beets from 1.3.8 to 1.3.9 but I get another error; No distributions at all found for beets in /usr/local/lib/python2.7/dist-packages

Are these these two errors related?

Ignore me. My machine didn't have a valid DNS server. :(

It would be nice to pull in work-rels though.

Also for TLAN.

sampsyo: Ping?

Hi! I'm on a flaky connection (traveling) so be aware that I might disappear.

sampsyo: Fair enough.

Are https://github.com/sampsyo/beets/commit/63803a9... and R110 the ones causing the additional ftintitle printing?

( https://github.com/sampsyo/beets/issues/1172 )

It's not that commit per se, but you're on the right track.

It should be using `logging` rather than `print_`

Ah cool.

Is that a "global" thing? That all the ui_print_()'s should be logging()'s?

(Well, "global" for ftintitle anyway.)

Yes, basically, at least for any code that is not just a CLI command.

Here's a little orientation:

The `ft_in_title` function in that plugin does all the actual work.

I can't figure out if it's calling something different during import than during `beet ftintitle`.

Yep. I figured that much. :)

It is called in two places: inside a closure inside `commands` (i.e., the CLI command), and inside `imported` (i.e., as an import hook)

This is a common pattern in plugins (e.g., lyrics, lastgenre).

There, the typical solution is to add some parameter to the "do the actual work" function that tells it how verbose to be.

Right.

Alright. That should be doable.

The same thing should work fine here. You can choose between putting all the logging statements under `if`s, or passing in a logging level to be passed to `log.log`.

(This is so common that I'd like to add it to the plugin interface so that everybody doesn't have to reinvent it every time.)

Soon!

Soon!™

;)

:)

Is there something in the environment one could use to figure out whether it's running a command or an import? (Apart from parsing argv...)

Not currently, no. But that's exactly what I want to add. :)

Ie., does beet set a global variable "I am not importing stuff".

Heh. Alright then. :p

*now

Where would you place this? beets/plugins.py? Ie., the logging function?

Should it be a method of BeetsPlugin?

Well, what I'm imagining would actually be part of a slightly larger refactoring.

I want to provide every BeetsPlugin with its own `log` member, which would cascade to the global beets logger.

It would automatically prefix the plugin name, and it could also adjust its verbosity level depending on whether an import was running.

:)

Perhaps so.

Or maybe I should just hack ftintitle for now. :p

I would be happy either way. :)

So.

It seems like other plugins call log.info to just log stuff, and log.debug to actually print it?

Eh.

The other way around, obviously.

Yes, info is the default threshold. `-v` enables debug.

Cool.

Wasn't there a guide somewhere for how to set up virtualenv and stuff for beets?

Or am I mixing it up with mopidy now?

I probably am. :|

Freso: you could use virtualenv if you don't want beets installed globally

virtualenv <folder>, pip install beets any-other-dependencies

and ofc use ./folder/bin/activate to later activate that virtualenv

ThomW: I know.

Already did that.

Mopidy just has a special setup.py command to use rather than just install: https://docs.mopidy.com/en/develop/contributing...

`python2 setup.py test` seems to be broken here?

*.

Is there a way to `beet ls` everything that hasn't got a given value set?

am i incorrect in thinking ftintitle should fix this? http://pastie.org/private/polbgpcun25pq0tssutcg

marcaddeo: Define "fix"? What do you expect (or want) to happen?

well, i'd like the featured artist to be in the track title as opposed to being in the artist field

marcaddeo: Did you run ls again after ftintitle?

yes, it's the same

Try using -v?

Freso: http://pastie.org/private/fyppwkrklx5gdsx3blba

marcaddeo: What happens if you do a more general query?

such as?

E.g., have it run on "the roots"?

(Instead of "the roots feat talib phren".)

That looks like an `ls` output?

(Eh, `beet ls`)

oh did you want me to run ftintitle on 'the roots'

that makes sense.

:)

do you prefer with or without -v

I don't know if ftintitle maybe trips over itself because it's a single track.

lol

Just without.

https://github.com/sampsyo/beets/pull/1176 hasn't been merged yet, so ftintitle is plenty verbose without -v :p

same thing basically http://pastie.org/private/elaglf7ec8zi5lbxowaqg

i wonder if it's tripping up the regular expression because the artist is showing as:

The Roots feat. Talib Kweli / The Roots

Ah. Yeah, possibly.

File a report, probably.

so i'm pretty sure it's because the artist is in there twice

playing with the code

it seems the problem is with albumartist_split = artist.split(albumartist)

and then it assumes that there will only be two elements in the list

and only compares the last and first elements of the array

or, list

when in reality the middle element of the array albumartist_split[1] = feat. Talib Kweli /

the slash is a whole nother problem

wait maybe it's not

if there were a case that just looked for the first non blank element instead of left-hand side this would work, but there would still be an extraneous / at the end of the track title