SoundKloud

[eureka]

Why did you convert your repository to private? Nobody was spying on you...

[twobob]

evidently they were...

[eureka]

Originally Posted by twobob

evidently they were...

I can't beleive! Are you sure? But why? And why to raise barriers? Maybe they are like good guardian angels?

Here is my variant of code:

Spoiler:

#!/usr/bin/env python

# -*- encoding: utf-8 -*-

from subprocess import call
import soundcloud

client = soundcloud.Client(client_id='FILL_ME,
                           client_secret='FILL_ME_TOO')

query = raw_input("What do you want to search for?  ")
tracks = client.get('/tracks', q=query, limit=10, filter="streamable")

for idx, track in enumerate(tracks, start=1):
    print(u'[{0}] Title: {1}'.format(idx, track.title))
    print(u' ID: {0}'.format(track.id))
track_idx = int(input("Which track would you like to play?")) - 1

track = client.get('/tracks/{0}'.format(tracks[track_idx].id))
print("Artist: {0}".format(track.user.get(u'username')))
print("Duration (in milliseconds): {0}".format(track.duration))

stream_url = client.get(track.stream_url, allow_redirects=False).location
call(["mplayer", "{0}".format(stream_url)])

[twobob]

: )

Yep that would have been sensible but we were just trying to get stuff scratched together.

In retrospect a public facing version would make more sense when it is ready for release.

Thanks for your kind pointers.

[qlob]

Maybe we could have a devel release? We might need to remove the secret client id from the file though...

[eureka]

Originally Posted by qlob

Maybe we could have a devel release? We might need to remove the secret client id from the file though...

I suggest requiring a settings file from user, like soundcloud.ini:

[soundkloud]
id = USER_API_ID
secret = USER_API_SECRET

It's really easy to read ini files in Python with built-in ConfigParser.

Then you'll lose a neccessity to store private, sensitive information in repository.

[qlob]

Originally Posted by eureka

I suggest requiring a settings file from user, like soundcloud.ini:

[soundkloud]
id = USER_API_ID
secret = USER_API_SECRET

It's really easy to read ini files in Python with built-in ConfigParser.
Then you'll lose a neccessity to store private, sensitive information in repository.

.....hmmm good idea there...... twobob, what do you think about that?

[eureka]

Originally Posted by qlob

I almost fainted when I saw this thread...

This sounds like it could be pretty awesome.....

If you're willing to write it in python or ruby, I'd love to help as much as I can.

BTW, just out of curiousity, why not Go?

[twobob]

Yeah I'm happy with that.

Toying with using Kindlets for interface.

Because Python interface was implemented and is now done. : )

[twobob]

Hmm... So yeah. I got python to launch via a kindlet.
reused the CR3runner command structure for now.

Got some thread issues though. Kindlets don't like not being serviced (a very short fuse ) so the

call (["mplayer", "{0}".format(stream_url.location)])

makes it grumpy... is this a backgrounded call? can it be made one?

EDIT: subprocess.Popen didn't do the trick either... will keep digging


I did get the python bits to start spitting out music via a kindlet - the details are a bit messy right now, but it maybe a solution

[qlob]

try os.exec instead of the subprocess.call maybe?

[twobob]

okay . I thought one replaced the other. but I will

[twobob]

tracks = client2.get('/tracks', q = sys.argv[1], limit=10, filter="streamable")

i = 1
temp = {} #initalize dictionary

for track in tracks:
track.id, i)

temp.update({i:track.id}) #set dictionary key
i=i + 1
queryb = i-1

trackid = temp.get(queryb) #retrieve key

track = client2.get('/tracks/{0}'.format(trackid))

stream_url = client2.get(track.stream_url, allow_redirects=False)
subprocess.Popen(['mplayer','-identify',"{0}".format(stream_url.location)])

gives fairly solid results (95% of the time) as you can see I just use q = sys.argv[1] to get the search term in at runtime

[qlob]

ooooh looks fancy. how far along is the kindlet?

[twobob]

yeah.. hmm... not that far really. I spend much of the time wrestling with making it not die.

I can pass a few arguments to the kindlet via a static command.txt (this could be created dynamically of course) This could be Tracks / Sets / Search / Artist / Upload / Dashboard / Last played Track... etc. etc. or something

I can also pass an argument to the shell that call the python and an argument to the python from that so the cascade for information is slowly being put in place.

It's a overhead to invoke it via this method. I suspect there is an easier wasy but meh.
at least the thing seems to not want to crash all the time now

Will feedback when I have more... (some screenies maybe)