Jupyter-Drive (integration with google drive) beta.

[Matthias BUSSONNIER]

Hi all, 

You might all have seen that IPython 3.0 Beta was released not so long ago.

First, thanks for the feedback !

 

For those unaware, we are working on integration with google

drive as an extension. 

Thanks to Google and especially to Kester Tong for helping a lot. 

You can find the package on PyPI :

https://pypi.python.org/pypi/jupyterdrive

And of course on github :

https://github.com/jupyter/jupyter-drive/issues

You can consider this 0.9 version as a 1.0 Beta, 

there are still a few missing things like duplication of file, 

but the  core functionality will mostly be there. 

For now you also still need a **separate** IPython server to run with the google-drive

profile. Cf readme.md.

This version **does not** include real-time collaboration. It will need some refactor 

that will be done in IPython during the following year. So do expect next version 

of jupyterdrive to probably only be compatible with IPython 4.0. 

For those interested there is draft branch on github that start to implement a "mixed" content

manager, that would allow with the same IPython server to access both local hard drive content

and google drive one. 

As usual, back up your data, Jupyter-drive is much less battle tested than IPython itself.

-- 

Matthias

[Brian Granger]

Are we still planning on merging the branch that enabled "mixed" content before 3.0?

--
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/523B928E-4BC9-47DC-9514-4EB0F5CF3DDE%40gmail.com.
For more options, visit https://groups.google.com/d/optout.

--

Brian E. Granger
Cal Poly State University, San Luis Obispo
@ellisonbg on Twitter and GitHub
bgra...@calpoly.edu and elli...@gmail.com

[Matthias Bussonnier]

Quick from the airport.

No and it was not planned, it was "best effort", "would be great to have synced release". Jupyter drive is not Ipython. Release schedule is decoupled. 

Though release of IPython 3.0 is needed for me to actually release JDrive because it needs it, I can do whatever with JDrive.

If there are interest in mixed content 

And I have time, it might got in before 3.0. But if the use is low I'll focus on real time API. 

Envoyé de mon iPhone

To view this discussion on the web visit https://groups.google.com/d/msgid/jupyter/CAH4pYpRTtPDctpRUEpWvQwgbjbDjbz5W8mSDUhZFcYTXEE2Tyg%40mail.gmail.com.

[Brian Granger]

My only concern is that not having the full integration done that folks won't really use this much as they will have to create a separate profile and start a new server. If we can get this into 3.1, that would enable a lot more people to try this. As with all of our MVPs, we want to get as many people using it as possible to get feedback.

Think about it this way - we have a hypothesis: "lots of current notebook users would like to be able to store notebook on Google Drive."

For us to test that hypothesis, we need as much data as possible (even if it is qualitative feedback). If not many people try it out, we can't make any conclusions about what we have built...

Cheers,

Brian

To view this discussion on the web visit https://groups.google.com/d/msgid/jupyter/B963F274-56C5-4F71-B85A-D6F8217FDD04%40gmail.com.

For more options, visit https://groups.google.com/d/optout.

[Kester Tong]

Brian,

These are all really good points you make.  I'm going to see what's missing from the current mixed contents branch, so we can figure out if it's realistic to release with 3.0 or 3.1.  The only change in IPython needed is some kind of special folder icon for the Drive mount point (which is /gdrive right now).  But I'll try and figure out if any other changes on the IPython side are needed.

Kester

_______________________________________________
IPython-dev mailing list
IPyth...@scipy.org
http://mail.scipy.org/mailman/listinfo/ipython-dev

[Kester Tong]

After rebasing, the mixedcontents branch seems to work.  It seems realistic to me to have this by 3.0.  There may be some polishing needed, but the basic functionality is there now.

Kester

[Matthias Bussonnier]

Le 6 févr. 2015 à 18:38, Brian Granger <elli...@gmail.com> a écrit :

> My only concern is that not having the full integration done that folks won't really use this much as they will have to create a separate profile and start a new server. If we can get this into 3.1, that would enable a lot more people to try this. As with all of our MVPs, we want to get as many people using it as possible to get feedback.

I'll re-have a look then, but I still believe the schedule is unrelated to IPython.
We don't need to release a 3.1 for mixed content to work.

Le 6 févr. 2015 à 19:32, 'Kester Tong' via Project Jupyter <jup...@googlegroups.com> a écrit :

> After rebasing, the mixedcontents branch seems to work. It seems realistic to me to have this by 3.0. There may be some polishing needed, but the basic functionality is there now.

I'll have a look tomorrow.
I'll re-look at the implementation but I think we could even have an even
more generic mixed content manager than takes N contents manager and prefixes
and exposes each of them.

I suppose we could probably try with Scott (IIRC) postgres backend
as an exercise.

--
M

[Kester Tong]

I'll have a look tomorrow.
I'll re-look at the implementation but I think we could even have an even
more generic mixed content manager than takes N contents manager and prefixes
and exposes each of them.

I suppose we could probably try with Scott (IIRC) postgres backend
as an exercise.

I like this idea.  It would be slightly more general and consistent than what we currently use.  The implementation would be simpler since the first path component will always tell you which content manager to use.  Maybe the mixed content manager could be part of IPython?  We could test it out in Jupyter Drive first and move it later (still using regular contents manager by default).

Kester

 

--

M

[pablo...@gmail.com]

This version **does not** include real-time collaboration. It will need some refactor 

that will be done in IPython during the following year. So do expect next version 

of jupyterdrive to probably only be compatible with IPython 4.0. 

Can you explain what changes in IPython are required for real-time collaboration?
Will the OT happen on the IPython server?
Will it be tied to google-drive?

[Matthias Bussonnier]

Can you explain what changes in IPython are required for real-time collaboration?

We are discussing that, among other as

https://github.com/ipython/ipython/issues/7784

among other.

Will the OT happen on the IPython server?

OT happen on all sides, for the time being only on GDrive, 

and gapi javascript library. 

The computation of the permeation might only happen on the server. 

Will it be tied to google-drive?

In the first time yes to get it working. 

But we plan on making it pluggable, and do our best 

so that you can have a version that works on close network

and purely open-source software. Though we will need help for that. 

-- 

M

On Saturday, February 7, 2015 at 9:23:09 PM UTC+2, Kester Tong wrote:

I'll have a look tomorrow.
I'll re-look at the implementation but I think we could even have an even
more generic mixed content manager than takes N contents manager and prefixes
and exposes each of them.

I suppose we could probably try with Scott (IIRC) postgres backend
as an exercise.

I like this idea.  It would be slightly more general and consistent than what we currently use.  The implementation would be simpler since the first path component will always tell you which content manager to use.  Maybe the mixed content manager could be part of IPython?  We could test it out in Jupyter Drive first and move it later (still using regular contents manager by default).

Kester

 

--

M






_______________________________________________
IPython-dev mailing list
IPyth...@scipy.org
http://mail.scipy.org/mailman/listinfo/ipython-dev

--
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/d265e589-ee94-461b-b607-1591bcccd8a7%40googlegroups.com.