Python3 support for opencog/atomspace

160 views
Skip to first unread message

Eyob

unread,
Dec 19, 2017, 3:49:53 PM12/19/17
to opencog
(1)

So we have an intern at icog labs who is interested in making python3 bindings for opencog/atomspace.

If anybody wants to give him guidance on how to go about it, please write him here or here https://github.com/opencog/atomspace/issues/501

His github username/profile  is   https://github.com/Yididya/

Hope he would be asking for pull requests soon :D

(2)

I was wondering does every part of opencog/atomspace need to be built using python3 or we can proceed part by part. E.g.

It seems i really like this import statement to work right now

from opencog.scheme_wrapper import scheme_eval

in python3 but the error reads " ImportError: No module named 'opencog' " when executed.

Its like at the moment we need python3 bindings only for a small number of import statements, for a project we are doing now currently.

Thanks!

Alexey Potapov

unread,
Dec 19, 2017, 6:04:48 PM12/19/17
to ope...@googlegroups.com
This sounds great! We just needed python3 binding for OpenCog to send an output of a DNN to PLN in one script without additional wrappers.

--
You received this message because you are subscribed to the Google Groups "opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opencog+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/fab8f5fb-0ca4-40e9-92d4-33a448708b8f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Linas Vepstas

unread,
Dec 20, 2017, 2:04:49 AM12/20/17
to opencog
> does every part of opencog/atomspace need to be built using python3

There's really only one part: the C code that interacts with cython. its call in cython/PythonEval.cc  I'll bet that if you can get that to work, everything else will work without any porting. That's my guess.

--linas

--
You received this message because you are subscribed to the Google Groups "opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opencog+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/fab8f5fb-0ca4-40e9-92d4-33a448708b8f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
"The problem is not that artificial intelligence will get too smart and take over the world," computer scientist Pedro Domingos writes, "the problem is that it's too stupid and already has."

Roman Treutlein

unread,
Jan 16, 2018, 4:47:40 AM1/16/18
to opencog
There already exists a python3 branch in the atomspace repo unfortunately I never got it to work. This also contains a lot of changes I did trying to get it to work but didn't help that might have to be reversed. And the branch is over 2 years old not sure how much the python interface has changed since then but might still be worth looking at it.

Linas Vepstas

unread,
Jan 17, 2018, 2:57:31 AM1/17/18
to opencog
HI Roman, could you repost this directly on
https://github.com/opencog/atomspace/issues/501 where it won't get
lost? --linas
> --
> You received this message because you are subscribed to the Google Groups
> "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to opencog+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> Visit this group at https://groups.google.com/group/opencog.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/opencog/82208a05-9809-47e3-a2fa-84c69ff156bd%40googlegroups.com.

Eyob

unread,
Feb 15, 2018, 1:29:21 PM2/15/18
to opencog
So it seems opencog/atomspace has finally been ported to python3. Thanks Linas for doing the work :D
Reply all
Reply to author
Forward
0 new messages