These are chat archives for ipython/ipython

19th
Mar 2015
athelopuce
@athelopuce
Mar 19 2015 12:48
On a notebook I would like to see wich cells need to be run again (after a content modification). For that I have try to change in my custom.css the render of some class. For 'text-cell' we can used the 'unrendered* css class, but not for 'code-cell'. Is there a solution ? If not could we had the feature. Thanks
mgunter6
@mgunter6
Mar 19 2015 15:18
Good morning (or whenever you are). I have am using a MultiKernelManager ("mkm" below) to, well, manage kernels. I need to perform some async executions against those kernels. What's the best approach? I see that I can call "mkm.get_kernel(kernel_id).client()" to create a blocking client, but blocking is no good. I see that I can call "mkm.connect_shell(kernel_id)" to get a ZMQStream for the shell, but then I'd have to build my own messages (which should be fairly painless, but...).
(Ultimately need IPython 2 and IPython 3 solutions.)
Sylvain Corlay
@SylvainCorlay
Mar 19 2015 16:15
Hello. Do you guys have an idea on when you are going to start merging PRs marked for 4.0
Sylvain Corlay
@SylvainCorlay
Mar 19 2015 16:22
(the traits stuff?)
K.-Michael Aye
@michaelaye
Mar 19 2015 17:11
wow, thank you for adding the auto encapsulating of marked text with quotes on both sides. Was missing that since long in the editing capabilities in the notebook. Hope it’s not there for long and just forgot to try it? ;)
Jason Grout
@jasongrout
Mar 19 2015 17:35
@minrk - when I have passwords set up on the notebook server, is there a good way to authenticate to the /api/contents/ api?
It seems that /api/contents/ forwards to the login page, even if I've already authenticated and have a cookie set from a previous manual interaction with the page.
Jason Grout
@jasongrout
Mar 19 2015 17:42
Never mind; I see that jquery's .getJSON needs extra arguments to send cookies...
Sylvain Corlay
@SylvainCorlay
Mar 19 2015 18:35
Even if all our experiments with phosphor will probably have to be re-done after the repo split, we probably need to have a common playground for our experiments
Jason Grout
@jasongrout
Mar 19 2015 18:35
@Carreau - in your note about HoloViews, you said it seemed to rely on widgets. I looked at the code a bit, and it doesn't seem to.
(your note on the dev meeting notes for next week)
Sylvain Corlay
@SylvainCorlay
Mar 19 2015 18:35
@KesterTong @jasongrout @ellisonbg @sccolbert
Even if all our experiments with phosphor will probably have to be re-done after the repo split, we probably need to have a common playground for our experiments.
hum weird gitter bug
Matthias Bussonnier
@Carreau
Mar 19 2015 18:36
that's why I said seem.
I only had a quick glance.
let me recheck.
Sylvain Corlay
@SylvainCorlay
Mar 19 2015 18:37
How about we agree on a new upstream branch / or a given fork as a collaborative thing for our experiments
Jason Grout
@jasongrout
Mar 19 2015 18:37
+1 @SylvainCorlay
@Carreau - right, so they don't throw an exception if the widgets don't exist
Matthias Bussonnier
@Carreau
Mar 19 2015 18:38
Yes, sure. Experiment can be in a common place.
Sure, probably I just saw "Widgets", and example looked like widget;
it's "compatible with widgets" then.
Jason Grout
@jasongrout
Mar 19 2015 18:59
It looks like they do something with widgets, but it also looks like they don't intrinsically depend on widgets.
Jonathan Frederic
@jdfreder
Mar 19 2015 19:00
@Carreau s/seen/saw
Matthias Bussonnier
@Carreau
Mar 19 2015 19:01
?
Oh, ok.
Jonathan Frederic
@jdfreder
Mar 19 2015 19:01
Don't worry I haven't forgot about our tutorial proposal!
I'm hoping I can get to it tonight.
Matthias Bussonnier
@Carreau
Mar 19 2015 19:01
I just seen ? or I've just seen ?
Jonathan Frederic
@jdfreder
Mar 19 2015 19:02
just seen that : -> I just saw that:
Thomas Kluyver
@takluyver
Mar 19 2015 19:02
"I have seen", but "I saw"
Jason Grout
@jasongrout
Mar 19 2015 19:03
So I've just seen is okay
(because it is I have (just) seen)
Matthias Bussonnier
@Carreau
Mar 19 2015 19:04
Gniaaa...get lost, bad at multitasking, will re-read later.
Jonathan Frederic
@jdfreder
Mar 19 2015 19:04
:D
Matthias Bussonnier
@Carreau
Mar 19 2015 19:04
Ou je peux juste parler en français, comme ça c'est plus facile pour moi.
(Can just speak french, easier for me)
Thomas Kluyver
@takluyver
Mar 19 2015 19:04
pas de problème
Jason Grout
@jasongrout
Mar 19 2015 19:04
we need a bot that hooks to Google Translate...
Matthias Bussonnier
@Carreau
Mar 19 2015 19:06
shoudl be in danger zone
Min RK
@minrk
Mar 19 2015 21:45
@SylvainCorlay I think we should start merging 4.0 PRs now. @/all any objection?
Matthias Bussonnier
@Carreau
Mar 19 2015 21:53
No objections.
Thomas Kluyver
@takluyver
Mar 19 2015 21:53
No objections from me. Can you sort out the 3.x branch? I don't know exactly what you did with it?
Matthias Bussonnier
@Carreau
Mar 19 2015 21:53
Should make announce on ML?
That master might brake soon ?
Sylvain Corlay
@SylvainCorlay
Mar 19 2015 21:54
houra!
Thomas Kluyver
@takluyver
Mar 19 2015 21:54
I'll draft an announcement
athelopuce
@athelopuce
Mar 19 2015 22:11
I try again now that many guys are herre : On a notebook I would like to see wich cells need to be run again (after a content modification). For that I have try to change in my custom.css the render of some class. For 'text-cell' we can used the 'unrendered' css class, but not the desirated effect for 'code-cell'. Is there a solution ? If not could we had the feature (a class for unrun cell ?). Thanks
Thomas Kluyver
@takluyver
Mar 19 2015 22:19
@athelopuce that's an interesting idea, but I think it's probably quite complex
do you consider all cells below the modified one 'need to be run again', or are you thinking of analysing the code to determine which ones it could affect?
athelopuce
@athelopuce
Mar 19 2015 22:25
In fact the best will be to determine which cells are affected. But I'm agree it will be to complex. I ask for the simplest solution : have a possibility to just identify the cell where I'have modify content (code, or markdown) and not run. After I can decide to propaguate to the above cells if necessary.
Matthias Bussonnier
@Carreau
Mar 19 2015 22:28
Like ipython/ipython#6030 ?
Thomas Kluyver
@takluyver
Mar 19 2015 22:33
yes, we've thought about that, and it probably does make sense. We got stuck on how to display two different kinds of 'modified since' state - modified since last executed, and modified since last saved/
athelopuce
@athelopuce
Mar 19 2015 22:33
@Carreau: NO. When you have modify a code cell and not run it.
Thomas Kluyver
@takluyver
Mar 19 2015 22:37
btw @/all - draft email to the ML about the repo split here: http://pastebin.com/cEPMGp1x
let me know if you want any changes, or I'll send it in a bit