[x3d-public] Possible extensions to x3d.py (subclassing, etc)

[John Carlson]

Michalis, are you still willing to do a transfer from sourceforge to
GitHub? I’m not familiar with that.

Don, I have copied over my x3d.py versions.  I really can’t go backwards
every time you want me to.  I make stupid mistakes copying files around and
I have so many versions of the x3d.py stylesheet, I don’t know which is
which.

But there are plenty of Python examples in the archive which could be
modified to output XML  and the XML could be validated with tovrmlx3d.  I
have already show output from validating the output of a modified
JoeKick.py.  That is a is a great place to start, if JoeKick.py has not
been replaced already.

If you want a different X3dToPython.xslt to enable this, I have one.

No one wants to run Python that hasn’t been thoroughly inspected.
Obviously, I run a risk if I do so.  I have also not inspected all of
x3d.py to my satisfaction.

I recommend running the python in a properly secured sandbox, possibly some
kind of virtual machine, Kubernetes or Docker container.  People in your
position are more likely to have access and experience with this sort of
thing, and have colleagues, students or former students that can help.  I
stepped out of that environment in 2012, basically live in the middle of
corn fields, so I have little experience and contacts.  Hint:  none of my
former colleagues want to discuss such things because I worked at a top
secret weapons laboratory—plus they would rather be retired that deal with
tech.

You have the golden stylesheet and the golden x3d.py.  I don’t want to flip
to yours, because that would be backpedaling for me and possibly cause me a
lot of confusion.  I’m just trying to catch everybody up to where I am.
What i could do is create a separate repository for x3d.py et al, so people
could fork, branch, etc. on GitHub.  This would require XML schema, X3DUOM,
etc. etc.—too much.  I’m only one guy.  I can add collaborators,
organizations, etc, but I’ve limited my GitHub experience on purpose.  I
don’t want to become the Python, XML or XSLT star when I dislike them.  I’m
the John for Java* guy, with experience with Perl.  I use Python, because I
don’t like DOM in JavaScript for pulling information out of XML, seems
stunted, or I am unfamiliar with how to do it.  I do know querySelector,
maybe I should work on reading X3DUOM with that?

If there’s a more recent version of x3d.py than 4.0.64.4 (?) that would be
useful information.

In other words, if I were to test my version of x3d.py the only issues that
would show up would be containerField issues, I’m fairly sure.  I can
report on those.  These are not visible in Python code, and Vince and
Andreas already are on track with that.  I will chime in after the holiday.

Any subclasses of x3d.py would override the order of output in the XML,
 VRML and JSON output.  They would not be generated from X3DUOM.  I have no
clue how to introduce custom field ordering in X3DUOM, and I would prefer
to to it in the x3d.py stylesheet, but it’s mind-bending.  With some better
documentation about xsl:sort sequences, I might get somewhere, but I might
step on some other node’s ordering, so a full test suite is required, which
I’m willing to work on, but I need advice on how you capture ant output for
ant/python as you do.

If there aren’t issues with how others expect fields to be ordered in
x3d.py, can we proceed?


John

On Tue, Oct 31, 2023 at 4:27 PM Brutzman, Donald (Don) (CIV) <
brutzman at nps.edu> wrote:

> John, you are free to experimentally subclass or inject as you wish but
> there is no guarantee that such work will be maintainable.
>
>
>
> The way to upgrade x3d.py is to
>
>
>
>    - Show an erroneous output, or
>    - Provide example code showing a broadly useful utility method of some
>    sort.
>
>
>
> Such corrections and improvements can then get autogenerated from X3DUOM
> into x3d.py and then regression tested against X3D Examples.  If
> successfully working they can become part of a future x3d.py release, which
> is entirely open source.
>
>
>
>    - Python package x3d
>    - https://pypi.org/project/x3d
>
>
> Thanks in advance for all improvements to X3D and x3d.py python library.
>
>
>
> all the best, Don
>
> --
>
> Don Brutzman  Naval Postgraduate School, Code USW/Br
> brutzman at nps.edu
>
> Watkins 270,  MOVES Institute, Monterey CA 93943-5000 USA
> +1.831.656.2149
>
> X3D graphics, virtual worlds, navy robotics
> https://faculty.nps.edu/brutzman
>
>
>
> *From:* x3d-public <x3d-public-bounces at web3d.org> *On Behalf Of *John
> Carlson
> *Sent:* Tuesday, October 31, 2023 1:53 PM
> *To:* GPU Group <gpugroup at gmail.com>; Joe D Williams <
> joedwil at earthlink.net>
> *Cc:* X3D Graphics public mailing list <x3d-public at web3d.org>
> *Subject:* [x3d-public] Possible extensions to x3d.py (subclassing, etc)
>
>
>
> I’m going to take the approach that x3d.py cannot be modified, and instead
> use subclasses of x3d.py classes to provide order of fields as necessary
> for the Blender X3DV exporter. In particular, HAnimHumanoid XML, VRML and
> JSON methods will be overridden to provide containerField, field ordering,
> etc.
>
>
>
> x3d.py might take a dependency injection approach in the future, such that
> one or more delegates can be replaced.  I believe dependency injection is
> preferred to inheritance or monkey-patching, but X3D is pretty heavy into
> inheritance, or at least interfaces.  Any dependency injection should make
> any changes to output as desired.  I do not agree with totally replacing
> methods, as that is a maintenance nightmare.
>
>
>
> I apologize for all the emails over the last many days.  My mind is not as
> nimble as it could be.   I do not know if I could have arrived at a
> agreeable solution without all the emails.
>
>
>
> Now you know why technical interviews terrify me.
>
>
>
> Thanks everyone for being teddy bears.
>
>
>
> John
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20231031/389b9838/attachment-0001.html>