[10:23:28] <wikibugs>	 10Quarry: Define in a single place the pseudoname of unnamed queries - https://phabricator.wikimedia.org/T197029 (10zhuyifei1999) How about a `@property` in the [[https://github.com/wikimedia/analytics-quarry-web/blob/master/quarry/web/models/query.py|Query]] object?
[17:51:07] <ottomata>	 hey researchers!
[17:51:09] <ottomata>	 qq for you
[17:51:23] <ottomata>	 i'm looking at notebooks + spark stuff just beginning to try to figure out how to do it well
[17:51:29] <ottomata>	 and i came across python anaconda 
[17:51:31] <ottomata>	 are you familiar with it?
[17:51:43] <ottomata>	 if that was installed across analytics cluster and in swap notebooks
[17:51:47] <ottomata>	 would that be a helpful thing?
[17:51:49] <harej>	 anaconda is a good alternative to figuring out the baffling chain of dependencies for sci-kit and such
[17:52:08] <ottomata>	 aye, so far for cluster type jobs, we use deb packages to install
[17:52:10] <harej>	 basically it's an all-in-one python runtime environment
[17:52:12] <ottomata>	 with whatever exists in debian
[17:52:13] <ottomata>	 or, build our own
[17:52:15] <ottomata>	 which can be nasty too
[17:52:35] <ottomata>	 wondering if a .deb of anaconda for general python (and R?) installed across the whole cluster would be more useful?
[17:52:45] <ottomata>	 than just python3 + whtaever random .debs people ask for
[17:52:48] <halfak>	 Is anaconda freely licensed?
[17:52:51] <ottomata>	 i think it is...
[17:52:53] <ottomata>	 checking'
[17:52:59] <harej>	 It would probably be less work for everyone involved
[17:53:07] <halfak>	 I guess I'll still want to pip some stuff. 
[17:53:15] <harej>	 anaconda supports pip!
[17:53:17] <halfak>	 Are you thinking of the .deb dependencies of various libs?
[17:53:39] <halfak>	 If anaconda supports pip, what does anaconda do?
[17:54:38] <ottomata>	 i am not yet sure!
[17:54:39] <harej>	 anaconda is python + binary dependencies
[17:55:33] <harej>	 you use it in lieu of a regular python runtime if you have nasty binary dependencies that are difficult to install
[17:55:40] <ottomata>	 "Because Conda introduces a new packaging format, you cannot use pip and Conda interchangeably; pip cannot install the Conda package format. You can use the two tools side by side (by installing pip with conda install pip) but they do not interoperate either.
[17:55:40] <ottomata>	 "
[17:56:12] <harej>	 basically I ignore Conda and use pip
[17:56:43] <ottomata>	 so you use the python from anaconda
[17:56:50] <ottomata>	 but install python deps into the (conda?) venv with pip?
[17:56:51] <harej>	 yes
[17:57:43] <ottomata>	 anyway, ok google tangent over.  back to jupyter, just interesting, hadn't really heard of it before
[17:57:43] <harej>	 at least that's what I do; I'm sure others have come up with different workflows