Re: Jupyter notebooks / Dynabook / JupyterTalk

Thank you Serge,

good to see that there is progress with Tensorflow and PolyMath.

I have problems installing Jupytertalk into a pristine 6.1 (I follow
the instructions)



    Metacello new
            baseline: 'JupyterTalk';
            repository: 'github://jmari/JupyterTalk';
            load:'all'

gives an error message

    'Could not resolve: BaselineOfJupyterTalk'

Is there something else which needs to be installed first?

--Hannes



On 7/28/18, Serge Stinckwich <[hidden email]> wrote:

Try:

Metacello new
        githubUser: 'jmari' project: 'JupyterTalk' commitish: 'master' path:
'repository';
        baseline: 'JupyterTalk';
        load



--
Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Thank you, Evan,  for an alternative load script.

Unfortunately it gives the same error message.

I use a recent, pristine Pharo 6.1 installation on a 64bit Ubuntu based Linux.


On 8/4/18, Evan Donahue <[hidden email]> wrote:

P.S. Some notes and links how Jupyter notebooks relate to the Dynabook idea here

http://wiki.squeak.org/squeak/1318

My aim is to use nteract as client (https://nteract.io/ ) together
with a Pharo based Smalltalk "kernel" (i.e. server acting as the
"scripting engine")

On 8/7/18, H. Hirzel <[hidden email]> wrote:

Somewhat related - but the Pharo equivalent (said naively - as I’ve only seen demos of Jupyter and not used it in anger myself) is gtDocumentor which looks to be the concept reimagined in Smalltalk. Apologies if you already know about it - but its so cool its worth checking out if you haven’t - https://github.com/feenkcom/gtoolkit-documenter (note - take note of the bottom comment and load all of glamourous toolkit to get the full impression).

Hannes Hirzel wrote
> P.S. Some notes and links how Jupyter notebooks relate to the Dynabook
> idea here
> http://wiki.squeak.org/squeak/1318

From the swiki: "A Jupyter Notebook is… similar to Active Essays kept in a
Smalltalk based Dynabook."

I've understood a key point of the Dynabook dream (and Smalltalk as
prototype Dynabook software) is that as you drill down to something facing
the end user, you keep the full power of the computer. For example, when
Alan Kay presents, he doesn't use Powerpoint, but Squeak. Hence, the slides
are each fully functioning World where one can e.g. bring up halos,
inspectors, class browsers.

Is that the case with Jupiter Notebooks? I assumed not since the notebooks
seem to be saved as JSON, unless maybe it's used for object serialization
like STON?



-----
Cheers,
Sean
--
Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Grafoscopio exists due to various limitations of Jupyter, particularly the lack of real support for objects in its core language - Python.  JSON is not equivalent to STON even in terms of storing JavaScript, it's mainly a data format.

On 8/7/18, 2:14 PM, "Pharo-users on behalf of Sean P. DeNigris" <[hidden email] on behalf of [hidden email]> wrote:

    Hannes Hirzel wrote
    > P.S. Some notes and links how Jupyter notebooks relate to the Dynabook
    > idea here
    > http://wiki.squeak.org/squeak/1318
   
    From the swiki: "A Jupyter Notebook is… similar to Active Essays kept in a
    Smalltalk based Dynabook."
   
    I've understood a key point of the Dynabook dream (and Smalltalk as
    prototype Dynabook software) is that as you drill down to something facing
    the end user, you keep the full power of the computer. For example, when
    Alan Kay presents, he doesn't use Powerpoint, but Squeak. Hence, the slides
    are each fully functioning World where one can e.g. bring up halos,
    inspectors, class browsers.
   
    Is that the case with Jupiter Notebooks? I assumed not since the notebooks
    seem to be saved as JSON, unless maybe it's used for object serialization
    like STON?
   
   
   
    -----
    Cheers,
    Sean
    --
    Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html
   
   

Jupyter notebooks are not limited to Python, there are more than 40
https://github.com/jupyter/jupyter/wiki/Jupyter-kernels

Pharo Smalltalk is one of them
https://github.com/jmari/JupyterTalk

I'd like to get it up and running....


On 8/7/18, Andrew Glynn <[hidden email]> wrote:

Hi,

Yes, Grafoscopio exist to overcome limitations of Jupyter as Andrew
says, its lack of a object model but also its overcomplicated
architecture[0]. I have said that in some way, Grafoscopio and
JupyterLab[1] are following opposite paths. The last started as an
interactive notebook and is trying to become an IDE (with "I" for
Interactive, not for Integrated) and the former started from the Pharo
IDE and is trying to provide interactive notebook capabilities.

[0]
http://mutabit.com/offray/static/blog/output/posts/grafoscopio-idea-and-initial-progress.html
[1] https://github.com/jupyterlab/jupyterlab

At some point is also a deconstruction over the idea of the Active Essay
proposed by Alan Kay and his team. I think that the problem with the
Dynabook is that is so rooted into the future that, when time passes and
the present occupies the place of the vision from the past, it usually
does with some subobtimal but popular technology (the web, Java, CORBA,
you name it). At some point I think that this is because, despite of the
Dynabook being from children from 6 to 100 years, most of the research
was done with children and teenagers and the adult world was occupied by
those other popular technologies and because the world is inherited by
the majority of adults to children we have the world that we have today.
I wonder what happen with the Dynabook kids? How are they now as adults?
It seems that the role of becoming and adult (or being already one) was
not deeply deployed by the Dynabook and there is a lot of room to do in
that place, for teachers (as DrGeo and Hilaire writings are showing),
for researchers, journalists, investigators, activists, as Grafoscopio
is trying to show.

At some point I would like better integration between Grafoscopio and
GTDocumenter, but is a project that advances slowly with a solo main
developer and with not wide use in the Pharo community (I think is
because the use of Markdown instead of Pillar, among other reasons) or
other commit contributions. In that sense is pretty similar to other
FLOSS projects that are coded by one or two developers. But for the last
quarter of this year the local community is going to make some kind of
anti-bootcamp and I hope to show some advances in Grafoscopio in the end
of the year as a result of this.

With Pharo 7 and the eventual integration of GTDocumenter in future
Pharo releases I think we could be approach/deconstruct the Dynabook
even more.

Cheers,

Offray

On 07/08/18 14:00, Andrew Glynn wrote:

Hi Hannes,
I did not find a nice way to solve the problem but you will be able to load
jupytertalk following these simple steps:
- Try to load it executing
Metacello new
        baseline: 'JupyterTalk';
        repository: 'github://jmari/JupyterTalk';
        load:'all'
- It will fail...open Iceberg, there must be a new repository "JupyterTalk"
in.
- Edit the repository (cmd+E) and add "repository" to the code subdirectory
field.
- Try to execute this again,
Metacello new
        baseline: 'JupyterTalk';
        repository: 'github://jmari/JupyterTalk';
        load:'all'
------
Befor installing Jupytertalk, you need ZeroMQ and Jupyter installed in your
system and you must create a the folder
'/usr/local/share/jupyter/kernels/pharo'.  and a file 'kernel.json' with the
contents like that...

'{
  "argv": [
    "/Applications/Pharo6.1-64_ZeroMQ.app/Contents/MacOS/Pharo",
   
"/Applications/Pharo6.1-64_ZeroMQ.app/Contents/Resources/Pharo6.1-64.image",
    "ipharo",
    "{connection_file}"
  ],
  "display_name": "Pharo Smalltalk",
  "language": "smalltalk"
}'
 
feel free to ask more questions, I'll be glad to help!



--
Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

That worked for me

Metacello new
        baseline: 'JupyterTalk';
        repository: 'github://jmari/JupyterTalk:master/repository';
        load:'all'

I will correct the github readme page.

Anyway I think adding the subdirectory after the project name work for me in previous versions of iceberg...I have to check it in my other projects...Thanks



--
Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Hello Jesús Marí and others

Thank you for the updated Metacello script. This brings some steps forward.


The use case:

1. I started with http://pharo.org/web/download

    curl https://get.pharo.org/64/ | bash


2. Open pharo with
    ./pharo-ui


3. JupyterTalk installation
I pasted


   Metacello new
        baseline: 'JupyterTalk';
        repository: 'github://jmari/JupyterTalk:master/repository';
        load:'all'

4. A problem came up, see screen shot JupyterTalk_Installation
    I hit 'proceed'


5. I wanted to save the image and did 'save and quit'. Then an
additional error popped up
    See screen shot 'Error_after_choosing_save_and_quit'

Help is appreciated.

Regards
Hannes





On 8/13/18, Jesús Marí <[hidden email]> wrote:

P.S. I see that the Metacello script installs a lot in one go.

I might be easier to identify the problem if the installation can be
done in steps, i.e. several load scripts in sequence.

On 8/15/18, H. Hirzel <[hidden email]> wrote:

Also useful would be a catalog entry.

On 8/15/18, H. Hirzel <[hidden email]> wrote:

H. Hirzel <[hidden email]> wrote:
> P.S. I see that the Metacello script installs a lot in one go.
>
> I might be easier to identify the problem if the installation can be
> done in steps, i.e. several load scripts in sequence.

You already have all the information you need. RT classes missing. Is this
supposed to be run in Moose only?

Stephan

On 8/15/18, Stephan Eggermont <[hidden email]> wrote:
> H. Hirzel <[hidden email]> wrote:
>> P.S. I see that the Metacello script installs a lot in one go.
>>
>> I might be easier to identify the problem if the installation can be
>> done in steps, i.e. several load scripts in sequence.
>
> You already have all the information you need. RT classes missing.
Good point. I was assuming the script pulls in Roassal in case it is needed.

HH

After loading Roassal2 through the Pharo 6.1 catalog and re-executing
the Metacello load script

  Metacello new
        baseline: 'JupyterTalk';
        repository: 'github://jmari/JupyterTalk:master/repository';
        load:'all'

All loaded fine.

I found 2 tests. (Result attached). One of them fails.


The question is now how to proceed.

On
https://github.com/jmari/JupyterTalk/blob/master/README.md

writing the installation procedure is noted on the 'ToDo' list.

I suggest that we just develop the procedure in this mail thread.

Most important question now is:

How do I start the Pharo Smalltalk Jupyter kernel?

HH.

On 8/15/18, H. Hirzel <[hidden email]> wrote:

Attachement shows the result of a successful installation into Pharo 6.1



On 8/15/18, H. Hirzel <[hidden email]> wrote: