These are chat archives for ipython/ipython

5th
Dec 2014
Jonathan Frederic
@jdfreder
Dec 05 2014 00:44
@jasongrout no, just waiting to see if anyone else has input, otherwise I'll merge soon.
Kyle Kelley
@rgbkrk
Dec 05 2014 01:03
@jasongrout That's really great to see and hear about the SageMath initiatives for enhancing textbooks and materials in the classroom
It's right on the lines of what O'Reilly is wanting to do, integrated with their authorship platform
epifanio
@epifanio
Dec 05 2014 02:31
@takluyver @minrk I’m using the ipytohn js extension (to, gist, tags) thay work as expected. If i add the new “citep” extension (to enable citation in markdown from zotero) it doesn’t show up in the notebook (i checked the js console log print the statement: citep loaded).
If i comment the (toc, gist, tags) extension, then citation extension show up correctly.
I tried on a clean profile on browser with clean cache and private/secure window
epifanio
@epifanio
Dec 05 2014 02:38
don’t know how to debug it, just wondering if you encoutered already this behavior
Thomas Kluyver
@takluyver
Dec 05 2014 02:53
citep? Do you mean cite2c?
Vlad
@VDuda
Dec 05 2014 03:56
Hello; just wanted to put out this general Q: "Is there an app/site that allows me to spin up an ipython notebook, without me needing to know much about server management?"
Kyle Kelley
@rgbkrk
Dec 05 2014 04:00
If you want persistence, right now https://wakari.io is your best option
If you only need it for a fleeting moment, https://tmpnb.org (for now)
^ @VDuda
epifanio
@epifanio
Dec 05 2014 05:42
@takluyver , sorry, yes i meant cite2c
Andreas Klostermann
@akloster
Dec 05 2014 14:16
any news on Pydata NYC videos?
@VDuda - sagemathcloud is also an excellent option. I think it has a lot more than wakari.io right now (snapshots, collaborative editing in IPython notebooks, public sharing of notebooks, etc.): https://cloud.sagemath.com/
@akloster: they've been up for a while.
oh, sorry, I'm talking about the pydata NYC meetup. I'm not sure about the pydata conference
@jdfreder: just a friendly ping about #7093 (can you tell I'm excited about it going in?) :)
Jonathan Frederic
@jdfreder
Dec 05 2014 16:31
done :)
Jason Grout
@jasongrout
Dec 05 2014 16:31
thanks!
Andreas Klostermann
@akloster
Dec 05 2014 16:36
I meant the pydata conf that just finished...
Kyle Kelley
@rgbkrk
Dec 05 2014 17:13
@jasongrout @VDuda Gah, I feel like such a jerk. I totally forgot SageMathCloud. I have trouble shaking the Sage bit from my mind and remembering that you have IPython too.
Thomas Kluyver
@takluyver
Dec 05 2014 19:22
Just come across peru - like git submodules, but VCS independent
looks interesting
Jason Grout
@jasongrout
Dec 05 2014 19:25
@jdfreder: I notice that we now have warning messages about widgets being experimental
What I noticed, though, was that if you are running in the notebook, the warning messages are printed to the web server console, not the notebook :)
maybe that's intentional, but it sure seems odd.
Jonathan Frederic
@jdfreder
Dec 05 2014 19:26
That wasn't my intention
Jason Grout
@jasongrout
Dec 05 2014 19:27
@rgbkrk - yeah, maybe the name should be shortened or broadened. Perhaps MathCloud. Or maybe even "The Cloud"
Fernando Perez
@fperez
Dec 05 2014 21:25
@minrk, thanks! The write() api changed a bit, and I didn't realize that (I was updating old code). Agreed that the error msg could be clearer.
Jason Grout
@jasongrout
Dec 05 2014 21:34
is there a system place for nbextensions? If we're installing a package that has javascript files to be installed into an nbextension, it would be really nice if it could be installed at the system level when the package was installed
instead of having to have a user run install_nbextension
For example, for my pythreejs package for webgl graphics, it would be much better to just do pip install pythreejs, compared to pip install pythreejs, then make sure that you start up ipython and run import pythreejs; pythreejs.install_nbextension.
Thomas Kluyver
@takluyver
Dec 05 2014 21:42
I seem to remember a PR adding a systemwide place to install nbextensions. I forget where, though.
For stuff with a Python part, the other option is to copy the JS files into place at runtime.
Jason Grout
@jasongrout
Dec 05 2014 22:11
right, that's how you do it, I guess.
but that just seems bad, especially if you have a lot things to copy. In my case, I'm copying three.js when I'm developing. I don't want to pay that cost needlessly every time.
Thomas Kluyver
@takluyver
Dec 05 2014 22:13
install_nbextension() won't actually copy the files every time
it will compare the modified time and only copy if they're newer
Jason Grout
@jasongrout
Dec 05 2014 22:17
right; I guess we've had that conversation too. It still seems wasteful.
Especially if we can install once to a system directory, instead of having every user install to their personal directory
Thomas Kluyver
@takluyver
Dec 05 2014 22:22
Yep. If we can work out a solid scheme for how to deal with system, environment and user locations on different platforms, I think we should add other nbextension locations. I'm trying to work out something similar for kernelspecs.
Jason Grout
@jasongrout
Dec 05 2014 22:47
ah, I thought you already had something worked out for kernelspecs. I was hoping to just piggy-back on whatever design decisions you made there.
Thomas Kluyver
@takluyver
Dec 05 2014 22:48
We have something for systemwide installation worked out, but I'm still thinking how it should interact with e.g. conda environments
and I'd prefer not to copy it for nbextensions until we've worked that out
I'll try to keep thinking about it.
Jason Grout
@jasongrout
Dec 05 2014 22:49
sure. Let's work out one way of doing it all the way, and then consider other things
Nathan Goldbaum
@ngoldbaum
Dec 05 2014 23:32
Hi all, i'm working on the yt docs build right now. Every once in a while IPython crashes during the docs build with OperationalError: database is locked
which seems to be due to a lock timeout expiring in the IPython history db
is there a way to just completely turn off access to the history db?
this is all automated builds, so we don't care if we lose the history
we also have c.HistoryManager.enabled = False in the ipython config
Nathan Goldbaum
@ngoldbaum
Dec 05 2014 23:41
ah, found "HistoryAccessor.enabled = False"'
Thomas Kluyver
@takluyver
Dec 05 2014 23:41
@ngoldbaum - from the code, it looks like HistoryManager.enabled = False should work
HistoryManager subclasses HistoryAccessor, so setting either should be equivalent
not sure why it would fail
Nathan Goldbaum
@ngoldbaum
Dec 05 2014 23:42
unfortunately I don't have access to the build machine
we're trying again with both options set