Hi David,

"Python’s popularity is driven in no small part by the vast number of specialized libraries available for it, particularly in the domain of artificial intelligence, where the Keras library <https://keras.io/> is a heavyweight among deep-learning developers: Keras provides an interface to the TensorFlow <https://www.tensorflow.org/>, CNTK <https://docs.microsoft.com/en-us/cognitive-toolkit/>, and Theano <http://www.deeplearning.net/software/theano/> deep-learning frameworks and tool kits.”

and:

"dramatic increase in computing power <https://spectrum.ieee.org/geek-life/hands-on/machine-learning-thats-light-en...> found in microcontrollers means that embedded versions of Python, such as CircuitPython <https://spectrum.ieee.org/at-work/tech-careers/scott-shawcroft-is-squeezing-...> and MicroPython <https://micropython.org/>, are becoming increasingly popular among makers.”

I’d love to be able to say, when people challenge me about Smalltalk, that we have those, or the logical equivalent of this.

The Python library catalog is indeed very appealing. Making Python libraries accessible in Smalltalk is an obvious and significant gain for Smalltalk. This is exactly the problem we worked on. We built the Python Bridge, available here: https://github.com/ObjectProfile/PythonBridge <https://github.com/ObjectProfile/PythonBridge> This Python Bridge can be specialized for particular Python libraries, in particular all the AI libraries. We did this for the Keras library: https://github.com/ObjectProfile/PythonBridge <https://github.com/ObjectProfile/PythonBridge> As benefit, you can use Keras (and therefore Google’s tensorflow) without even knowing that you are talking to a Python libraries. One of the cool aspect of our work, is that it works for both VisualWorks and Pharo, as it was demonstrated at ESUG a few weeks ago. Cheers, Alexandre

On Wed, Sep 11, 2019 at 10:41 AM Boris Shingarov <boris@shingarov.com> wrote:

Hi Alexandre,

I went to look at PythonBridge and it promises to be massively interesting! Definitely much more systematic than the crude and primitive PyStGateway I hacked together for the MachineArithmetic package.

However, I must be doing something stupid. As PythonBridge's "Getting Started" page tells me, I loaded the metacello:

Metacello new baseline: 'PythonBridge'; repository: 'github://ObjectProfile/PythonBridge/src'; load.

so now I have PythonBridge master and Python3Generator v2.0.0 in my Pharo 7 image. Ok nice.

However, PBApplication class does not implement #installPipenvEnvironment, so

(Smalltalk at: #PBApplication) installPipenvEnvironment

results in MNU. In fact there are no implementors of #installPipenvEnvironment.

How would you suggest to debug this?

I already got this problem. Normally you don't have to do anymore: (Smalltalk at: #PBApplication) installPipenvEnvironment Just run the tests after your load PythonBridge in your image. Regards, -- Serge Stinckwic h Int. Research Unit on Modelling/Simulation of Complex Systems (UMMISCO) Sorbonne University (SU) French National Research Institute for Sustainable Development (IRD) U niversity of Yaoundé I, Cameroon "Programs must be written for people to read, and only incidentally for machines to execute." https://twitter.com/SergeStinckwich

you have also the solution of vincent arenaga that smoothly let you manipulate Python object in Pharo and the inverse a bit like a reflective oriented corba layers…

On 11 Sep 2019, at 01:56, Alexandre Bergel <abergel@dcc.uchile.cl> wrote:

Hi David,

...

"Python’s popularity is driven in no small part by the vast number of specialized libraries available for it, particularly in the domain of artificial intelligence, where the Keras library <https://keras.io/> is a heavyweight among deep-learning developers: Keras provides an interface to the TensorFlow <https://www.tensorflow.org/>, CNTK <https://docs.microsoft.com/en-us/cognitive-toolkit/>, and Theano <http://www.deeplearning.net/software/theano/> deep-learning frameworks and tool kits.”

and:

"dramatic increase in computing power <https://spectrum.ieee.org/geek-life/hands-on/machine-learning-thats-light-en...> found in microcontrollers means that embedded versions of Python, such as CircuitPython <https://spectrum.ieee.org/at-work/tech-careers/scott-shawcroft-is-squeezing-...> and MicroPython <https://micropython.org/>, are becoming increasingly popular among makers.”

I’d love to be able to say, when people challenge me about Smalltalk, that we have those, or the logical equivalent of this.

The Python library catalog is indeed very appealing. Making Python libraries accessible in Smalltalk is an obvious and significant gain for Smalltalk. This is exactly the problem we worked on. We built the Python Bridge, available here:

https://github.com/ObjectProfile/PythonBridge <https://github.com/ObjectProfile/PythonBridge>

This Python Bridge can be specialized for particular Python libraries, in particular all the AI libraries. We did this for the Keras library:

https://github.com/ObjectProfile/PythonBridge <https://github.com/ObjectProfile/PythonBridge>

As benefit, you can use Keras (and therefore Google’s tensorflow) without even knowing that you are talking to a Python libraries.

One of the cool aspect of our work, is that it works for both VisualWorks and Pharo, as it was demonstrated at ESUG a few weeks ago.

Cheers, Alexandre _______________________________________________ Esug-list mailing list Esug-list@lists.esug.org http://lists.esug.org/mailman/listinfo/esug-list_lists.esug.org

-------------------------------------------- Stéphane Ducasse http://stephane.ducasse.free.fr http://www.synectique.eu / http://www.pharo.org 03 59 35 87 52 Assistant: Julie Jonas FAX 03 59 57 78 50 TEL 03 59 35 86 16 S. Ducasse - Inria 40, avenue Halley, Parc Scientifique de la Haute Borne, Bât.A, Park Plaza Villeneuve d'Ascq 59650 France