Siddharth Bhat

A weblog of sorts

github linkedin email rss
Gsoc Update - Weeks 3 & 4
Jul 8, 2015
2 minutes read

I’m sitting at home with the flu and a runny nose, so this is as good a time as any to write another blog post!

I managed to get a decent amount of work done over the past two weeks. I think I can start working on plotting from next week, since most of the Scenegraph update I talked about last time is done. What’s left is a polish and performance improvements, which will be done iteratively, and (I’m hoping) in parallel with the plotting API.

Scenegraph Overhaul

I’ve spent most of my time porting over a lot of the visuals from the old system to the new rewrite. It was an interesting exercise to learn the new architecture and to figure out how everything fits. Much of it was a direct rewrite, but some of it was challenging.

I’m stuck on a few things right now, chief among them being porting lighting of MeshVisuals. I suspect that there is a bug in the shader code / normals calculation, but I’ve not been able to isolate it properly.


Vispy also has an experimental webGL backend that’s going to be used with IPython. I’ve been checking it out, and it’s a really interesting project.

Vispy uses a custom domain specific “language” (it’s not a language in the Turing completeness sense of the word. It’s more of a spec / internal representation (IR) kind of like the LLVM IR) that is designed to represent openGL operations, which is knows as GLIR. This provides a really neat way to specify openGL commands in a nicely serialize-able format.

Vispy (both the python version and the JavaScript version) has an object-oriented abstraction of OpenGL called “gloo”. Now, gloo has been implemented in vispy.js. I plan on implementing GLIR on top of gloo in the coming weeks. That should be a fun exercise (both to learn gloo properly and to implement it in JavaScript)

Odds and Ends

The usual “fix what annoys you” is around this time too.

I added an option to the test suite that lets one test docstrings. This feature existed before, but required one to manually run the python file. It’s just a tad easier with this change

There was also a “bug” in the installer( that silently failed to install in development mode (python develop) if setuptools wasn’t present. I just added a warning so that this would be reported.

That’s it for this time!

Back to posts

comments powered by Disqus