Re: [jupyter] Output cell history / diff?

2016-11-14 Thread Martin Sandve Alnæs 
Hi guys,
a sidenote to the git integration comment: we're working
on getting the new project nbdime ready for a first release
soon, which is a tool for notebook diff and merge including
git integration. You can watch it here:

https://github.com/jupyter/nbdime

Be aware it still needs a little polish and some design work.

Martin


On 28 October 2016 at 10:46, wwwald  wrote:

> Hello Yuvi,
>
> Just thinking out loud here - wouldn't it be possible to use git as the
> backend for managing the cell's history?
>
> I'm sure it would take quite some additional work and thinking. But I've
> been eager to find a decent solution to integrate git with our
> notebooks/JupyterHub setup for quite some time, I think this would be a
> great way to have fine-grained versioning, even at the cell level. And it
> would avoid notebooks blowing up in size.
>
> Maybe it's a stupid idea - I'm not very familiar with notebook internals
> yet. Let me know what you think.
>
> wwwald
>
> On Friday, October 28, 2016 at 3:57:42 AM UTC+2, Yuvi Panda wrote:
>>
>> I've extracted that code out into
>> https://github.com/yuvipanda/nbtimetravel. It doesn't have an UI to
>> explore yet - will add that next week, probably. I'm thinking of
>> mostly adding a slider for the whole notebook, and also one per-cell.
>> Ideas / patches welcome :D
>>
>> I'm very interested in having this be installed with students' working
>> on learning to code - I feel a lot of insight can be gained by
>> analyzing the evolution of code over time as such students are working
>> towards their assignments or whatever. Just graphing the shape of the
>> AST of the answers to the same question by different students over
>> time seems like it would provide insights...
>>
>> On Thu, Oct 27, 2016 at 12:34 PM, Yuvi Panda  wrote:
>> > Haha, we just built something like this last Friday for a workshop we
>> > did at UC Berkeley. I'm going to spend some more time polishing it and
>> > making it better before publishing it. It's called 'nbhistory' - I'll
>> > send a link as soon as I've a vague version up, and would love it for
>> > you to test it up and give feedback!
>> >
>> > On Thu, Oct 27, 2016 at 8:41 AM, Patrick Surry 
>> wrote:
>> >> One of the great things about Jupyter for data analysis is that it
>> preserves
>> >> the output of each cell within and between sessions as a "historical
>> >> record".
>> >>
>> >> I often find myself copying & pasting a cell to execute with small
>> changes
>> >> and then compare to a previous iteration, either while I'm exploring
>> >> interactively, or because I'm re-running a previous analysis with
>> updated
>> >> data.
>> >>
>> >> In those cases it'd be really handy if you could preserve a few prior
>> >> versions of the output cell, so you could flip back & forth to compare
>> and
>> >> perhaps highlight differences automagically.
>> >>
>> >> Has anyone experimented with anything like that?
>> >>
>> >> Cheers,
>> >> Patrick
>> >>
>> >> --
>> >> You received this message because you are subscribed to the Google
>> Groups
>> >> "Project Jupyter" group.
>> >> To unsubscribe from this group and stop receiving emails from it, send
>> an
>> >> email to jupyter+u...@googlegroups.com.
>> >> To post to this group, send email to jup...@googlegroups.com.
>> >> To view this discussion on the web visit
>> >> https://groups.google.com/d/msgid/jupyter/d32d8284-dd02-4f14
>> -96e9-705c5d3d9adc%40googlegroups.com.
>> >> For more options, visit https://groups.google.com/d/optout.
>> >
>> >
>> >
>> > --
>> > Yuvi Panda T
>> > http://yuvi.in/blog
>>
>>
>>
>> --
>> Yuvi Panda T
>> http://yuvi.in/blog
>>
> --
> You received this message because you are subscribed to the Google Groups
> "Project Jupyter" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to jupyter+unsubscr...@googlegroups.com.
> To post to this group, send email to jupyter@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/
> msgid/jupyter/0152028b-0ded-41de-bbc4-6f9c4c1d6b6b%40googlegroups.com
> 
> .
>
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"Project Jupyter" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to jupyter+unsubscr...@googlegroups.com.
To post to this group, send email to jupyter@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/jupyter/CADv7UFET%3Dxfp%3DYWOeqJaXCviKpqwU8vUyP4JtwnaNWVp-_C2ZQ%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Re: [jupyter] Output cell history / diff?

2016-10-28 Thread Patrick Surry 
Very cool!  

Looking forward to seeing your approach on UI, I was imagining just 
forward/back arrows under the "In [20]:" prompt in the margin to flip 
back/forth in the cell history, maybe highlight stuff that differs from the 
next version of the cell?

Not sure how a whole workbook slider would work - wouldn't each cell have 
potentially different amount of history?

Best,
Patrick

On Thursday, October 27, 2016 at 9:57:42 PM UTC-4, Yuvi Panda wrote:
>
> I've extracted that code out into 
> https://github.com/yuvipanda/nbtimetravel. It doesn't have an UI to 
> explore yet - will add that next week, probably. I'm thinking of 
> mostly adding a slider for the whole notebook, and also one per-cell. 
> Ideas / patches welcome :D 
>
> I'm very interested in having this be installed with students' working 
> on learning to code - I feel a lot of insight can be gained by 
> analyzing the evolution of code over time as such students are working 
> towards their assignments or whatever. Just graphing the shape of the 
> AST of the answers to the same question by different students over 
> time seems like it would provide insights... 
>
> On Thu, Oct 27, 2016 at 12:34 PM, Yuvi Panda  > wrote: 
> > Haha, we just built something like this last Friday for a workshop we 
> > did at UC Berkeley. I'm going to spend some more time polishing it and 
> > making it better before publishing it. It's called 'nbhistory' - I'll 
> > send a link as soon as I've a vague version up, and would love it for 
> > you to test it up and give feedback! 
> > 
> > On Thu, Oct 27, 2016 at 8:41 AM, Patrick Surry  > wrote: 
> >> One of the great things about Jupyter for data analysis is that it 
> preserves 
> >> the output of each cell within and between sessions as a "historical 
> >> record". 
> >> 
> >> I often find myself copying & pasting a cell to execute with small 
> changes 
> >> and then compare to a previous iteration, either while I'm exploring 
> >> interactively, or because I'm re-running a previous analysis with 
> updated 
> >> data. 
> >> 
> >> In those cases it'd be really handy if you could preserve a few prior 
> >> versions of the output cell, so you could flip back & forth to compare 
> and 
> >> perhaps highlight differences automagically. 
> >> 
> >> Has anyone experimented with anything like that? 
> >> 
> >> Cheers, 
> >> Patrick 
> >> 
> >> -- 
> >> You received this message because you are subscribed to the Google 
> Groups 
> >> "Project Jupyter" group. 
> >> To unsubscribe from this group and stop receiving emails from it, send 
> an 
> >> email to jupyter+u...@googlegroups.com . 
> >> To post to this group, send email to jup...@googlegroups.com 
> . 
> >> To view this discussion on the web visit 
> >> 
> https://groups.google.com/d/msgid/jupyter/d32d8284-dd02-4f14-96e9-705c5d3d9adc%40googlegroups.com.
>  
>
> >> For more options, visit https://groups.google.com/d/optout. 
> > 
> > 
> > 
> > -- 
> > Yuvi Panda T 
> > http://yuvi.in/blog 
>
>
>
> -- 
> Yuvi Panda T 
> http://yuvi.in/blog 
>

-- 
You received this message because you are subscribed to the Google Groups 
"Project Jupyter" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to jupyter+unsubscr...@googlegroups.com.
To post to this group, send email to jupyter@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/jupyter/ea969a17-54f0-4385-8a26-2474cbff992a%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [jupyter] Output cell history / diff?

2016-10-28 Thread wwwald 
Hello Yuvi,

Just thinking out loud here - wouldn't it be possible to use git as the 
backend for managing the cell's history?

I'm sure it would take quite some additional work and thinking. But I've 
been eager to find a decent solution to integrate git with our 
notebooks/JupyterHub setup for quite some time, I think this would be a 
great way to have fine-grained versioning, even at the cell level. And it 
would avoid notebooks blowing up in size.

Maybe it's a stupid idea - I'm not very familiar with notebook internals 
yet. Let me know what you think.

wwwald

On Friday, October 28, 2016 at 3:57:42 AM UTC+2, Yuvi Panda wrote:
>
> I've extracted that code out into 
> https://github.com/yuvipanda/nbtimetravel. It doesn't have an UI to 
> explore yet - will add that next week, probably. I'm thinking of 
> mostly adding a slider for the whole notebook, and also one per-cell. 
> Ideas / patches welcome :D 
>
> I'm very interested in having this be installed with students' working 
> on learning to code - I feel a lot of insight can be gained by 
> analyzing the evolution of code over time as such students are working 
> towards their assignments or whatever. Just graphing the shape of the 
> AST of the answers to the same question by different students over 
> time seems like it would provide insights... 
>
> On Thu, Oct 27, 2016 at 12:34 PM, Yuvi Panda  > wrote: 
> > Haha, we just built something like this last Friday for a workshop we 
> > did at UC Berkeley. I'm going to spend some more time polishing it and 
> > making it better before publishing it. It's called 'nbhistory' - I'll 
> > send a link as soon as I've a vague version up, and would love it for 
> > you to test it up and give feedback! 
> > 
> > On Thu, Oct 27, 2016 at 8:41 AM, Patrick Surry  > wrote: 
> >> One of the great things about Jupyter for data analysis is that it 
> preserves 
> >> the output of each cell within and between sessions as a "historical 
> >> record". 
> >> 
> >> I often find myself copying & pasting a cell to execute with small 
> changes 
> >> and then compare to a previous iteration, either while I'm exploring 
> >> interactively, or because I'm re-running a previous analysis with 
> updated 
> >> data. 
> >> 
> >> In those cases it'd be really handy if you could preserve a few prior 
> >> versions of the output cell, so you could flip back & forth to compare 
> and 
> >> perhaps highlight differences automagically. 
> >> 
> >> Has anyone experimented with anything like that? 
> >> 
> >> Cheers, 
> >> Patrick 
> >> 
> >> -- 
> >> You received this message because you are subscribed to the Google 
> Groups 
> >> "Project Jupyter" group. 
> >> To unsubscribe from this group and stop receiving emails from it, send 
> an 
> >> email to jupyter+u...@googlegroups.com . 
> >> To post to this group, send email to jup...@googlegroups.com 
> . 
> >> To view this discussion on the web visit 
> >> 
> https://groups.google.com/d/msgid/jupyter/d32d8284-dd02-4f14-96e9-705c5d3d9adc%40googlegroups.com.
>  
>
> >> For more options, visit https://groups.google.com/d/optout. 
> > 
> > 
> > 
> > -- 
> > Yuvi Panda T 
> > http://yuvi.in/blog 
>
>
>
> -- 
> Yuvi Panda T 
> http://yuvi.in/blog 
>

-- 
You received this message because you are subscribed to the Google Groups 
"Project Jupyter" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to jupyter+unsubscr...@googlegroups.com.
To post to this group, send email to jupyter@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/jupyter/0152028b-0ded-41de-bbc4-6f9c4c1d6b6b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [jupyter] Output cell history / diff?

2016-10-27 Thread Yuvi Panda 
Haha, we just built something like this last Friday for a workshop we
did at UC Berkeley. I'm going to spend some more time polishing it and
making it better before publishing it. It's called 'nbhistory' - I'll
send a link as soon as I've a vague version up, and would love it for
you to test it up and give feedback!

On Thu, Oct 27, 2016 at 8:41 AM, Patrick Surry  wrote:
> One of the great things about Jupyter for data analysis is that it preserves
> the output of each cell within and between sessions as a "historical
> record".
>
> I often find myself copying & pasting a cell to execute with small changes
> and then compare to a previous iteration, either while I'm exploring
> interactively, or because I'm re-running a previous analysis with updated
> data.
>
> In those cases it'd be really handy if you could preserve a few prior
> versions of the output cell, so you could flip back & forth to compare and
> perhaps highlight differences automagically.
>
> Has anyone experimented with anything like that?
>
> Cheers,
> Patrick
>
> --
> You received this message because you are subscribed to the Google Groups
> "Project Jupyter" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to jupyter+unsubscr...@googlegroups.com.
> To post to this group, send email to jupyter@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/jupyter/d32d8284-dd02-4f14-96e9-705c5d3d9adc%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



-- 
Yuvi Panda T
http://yuvi.in/blog

-- 
You received this message because you are subscribed to the Google Groups 
"Project Jupyter" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to jupyter+unsubscr...@googlegroups.com.
To post to this group, send email to jupyter@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/jupyter/CAFw%3DyShuWa8qsefL3Yxam42nnh3f%3Db192W_rcBqm-o%2B_SJb5LA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.