yt development - BitBucket, Task Queues, and Streamlines
@ Matthew Turk | Monday, Feb 28, 2011 | 4 minute read | Update at Monday, Feb 28, 2011

The major changes this week came mostly in the form of administrative shifts. However, SamS did some great work I’m going to hint at (he’ll post a blog entry later) and I started laying the ground work for something I’ve been excited about for a while, an MPI-aware task queue.


For the last couple months, yt has been struggling under the constraints of the hg server on its hosting plan. The issue was that particular files checked into the repository ( for one, which is now gone, and amr_utils.c, also gone, for another) took a while to transfer over some connections. During this transfer, the (shared) hosting provider on would kill the server process, resulting in an “abort” message given to the cloning user.

Basically, this was kind of awful, because it meant people couldn’t clone the yt repo reliably, and it also meant that the install script would fail in unpredictable ways (usually indicating a Forthon or error.) I’m kind of bummed out that I didn’t do something about this sooner; I suspect several people probably have tried to install yt and failed as a result of this. I added some workarounds that staged the download of yt over a couple pulls, which usually fixed it, but there was no reliable solution.

Enter BitBucket. A few of the developers had been using BitBucket for private projects, small repositories, and even (especially) papers that we’d been working on. For a while we’d been talking about moving yt there and trying to leverage the functionality it brings for Distributed Version Control Systems – like forking and pull requests, social coding, and on and on –and last week we hit the breaking point. So we created a new user (yt_analysis) and uploaded the yt repo, the documentation repo, and the cookbook, and we’re going to be conducting our development there. The old addresses should all still work – we have forwarded to the new location.

One of the coolest aspects of this is that anyone can now “fork” the yt repository. What this means is that you can then get your own private version, which you can then make changes to very easily, and then submit them back upstream. I’m really excited about this and I would encourage people to take advantage of it. I’ve rewritten the Developer Documentation to describe how to do this.

All in all, I think this will be a very positive move. BitBucket has a number of value adds, including the forking model, but we should also immediately see a dramatic increase in the reliability of the repository.


SamS has done some work implementing streamlines. Right now they operate by integrating using RK4 any set of vector fields, and then plotting their paths using Matplotlib’s mplot3d command. He’s working on some cool ways to colorize their values, and one of the things I am pushing for is to take any given streamline and convert it to an AMR1DData object. This would enable you to, for instance, follow a stream line in magnetic fields and calculate the density at every point along that streamline.

Once Sam’s comfortable with the feature as-is, he’s going to blog about it, so I won’t steal the thunder for his hard work here.

Task Queues

Building on the ideas behind the time series analysis I started work on the idea of a task queue that’s MPI aware. When this is finished being implemented, it will act as a mechanism for dispatching work, which will be fully integrated with time series analysis. Right now it’s not even close to being done, but a few pieces of the architecture have been implemented.

The idea here is that you will be able to launch a parallel yt job, but have it split itself into sub-parallel tasks. For instance, it you had 100 outputs of a medium-size simulation to analyze, you would write your time series code as usual – you would describe the actions you want taken, how to output them, etc etc. You would then launch yt with a “sub-parallel” option, saying that you wanted to split the total number of MPI jobs into jobs of size N – for instance you could launch a 64 processor yt job, telling it to split into sub-groupings of 4 processors each. Each output would then be distributed in a first come first serve fashion to each of the processor groups. When each group finished its job, it would ask for the next job available, and so on. When completed, the results would be collated and returned.

I’m excited about this, but right now it’s in its infancy. I’ve constructed the mechanisms to do this within a single process space, with no sub-delegation of MPI tasks. The process of implementing this and properly integrating it with time series analysis is going to be a long one, but I am setting it as a task for the next major release of yt. If you’re at all interested in this, drop me a line, and I’m happy to show you how to get started testing it out.

yt extension modules

yt has many extension packages to help you in your scientific workflow! Check these out, or create your own.


ytini is set of tools and tutorials for using yt as a tool inside the 3D visual effects software Houdini or a data pre-processor externally to Houdini.


Trident is a full-featured tool that projects arbitrary sightlines through astrophysical hydrodynamics simulations for generating mock spectral observations of the IGM and CGM.


pyXSIM is a Python package for simulating X-ray observations from astrophysical sources.


Analyze merger tree data from multiple sources. It’s yt for merger trees!


yt_idv is a package for interactive volume rendering with yt! It provides interactive visualization using OpenGL for datasets loaded in yt. It is written to provide both scripting and interactive access.


widgyts is a jupyter widgets extension for yt, backed by rust/webassembly to allow for browser-based, interactive exploration of data from yt.


yt_astro_analysis is the yt extension package for astrophysical analysis.

Make your own!!

Finally, check out our development docs on writing your own yt extensions!

Contributing to the Blog

Are you interested in contributing to the yt blog?

Check out our post on contributing to the blog for a guide!

We welcome contributions from all members of the yt community. Feel free to reach out if you need any help.

the yt data hub

The yt hub at has a ton of resources to check out, whether you have yt installed or not.

The collections host all sorts of data that can be loaded with yt. Some have been used in publications, and others are used as sample frontend data for yt. Maybe there’s data from your simulation software?

The rafts host the yt quickstart notebooks, where you can interact with yt in the browser, without needing to install it locally. Check out some of the other rafts too, like the widgyts release notebooks – a demo of the widgyts yt extension pacakge; or the notebooks from the CCA workshop – a user’s workshop on using yt.

Social Links