yt hub | Projects

Enhance your renderings

samskillman — Thu, 26 Jan 2012 12:17:43 PST

This script takes in a list of png images, renormalized the colors, and outputs an enhanced version.

Example Fisheye Rendering

samskillman — Tue, 17 Jan 2012 12:00:08 PST

This is an example of how to use the MosaicFisheyeCamera to create a volume rendering to be projected on a dome.

ngoldbaum / diskdyn_yt / overview — Bitbucket

ngoldbaum — Tue, 15 Nov 2011 16:00:13 PST

This a collection of analysis scripts and helper routines I've written to analyze my FLASH2.5 data. This uses parallel_objects() to make frames and collect time series from thousands of simulation snapshots in parallel. I use yt to interface with the raw data and access image frames using FixedResolutionBuffer. The actual plots are done 'by hand' in matplotlib.

Fancy Multi-panel Plots

britton — Wed, 19 Oct 2011 19:00:43 PDT

This repository contains a framework for making nice looking multi-panel images from projection and slices saved to an hdf5 file. It also contains scripts for making projections and saving them to hdf5 files. This is very useful for working with very large simulation data.

Archive tools for NICS Kraken

britton — Fri, 30 Sep 2011 07:50:41 PDT

These are tools for archiving multiple simulation datasets on NICS Kraken, including a script for archiving a single data dump into multiple files. This is useful with extremely large datasets for which a single tar file is unreasonable. Also included is an example batch script for using the hpss queue for archiving.

LChier_to_standard

samskillman — Wed, 14 Sep 2011 12:49:07 PDT

Used to convert the .hierarchy file from the "Santa Fe Lightcone" simulation to the standard format read in by yt.

Paramdiff for diffing parameter files.

dcollins4096 — Mon, 12 Sep 2011 08:59:41 PDT

paramdiff.py
Formatted diff for unordered parameter files.

Order independent parameter files are a fact of life in computational
computing. Unfortunately, two parameter files that have significantly
different histories may have vastly different orders of elements, but differ
by only a single important parameter. paramdiff gives you the answer you want.

mydb - View your yt datasets online

sskory — Fri, 09 Sep 2011 09:52:38 PDT

This is simple and quickly written script which will allow you to view all your datasets that you've ever touched with yt, from anywhere. Given your AWS credentials, it will log you in and display the particulars of the datasets. You can perform searches with various filters to narrow down what is displayed.

toomre.py

ngoldbaum — Thu, 08 Sep 2011 22:35:48 PDT

Calculates the Toomre Q of a disk in an isolated galaxy simulation. This assumes that the angular momentum of the disk is aligned with the x axis.

yt.profiling_support

mturk — Fri, 02 Sep 2011 03:40:35 PDT

Profiling scripts and support for yt

until

jwise — Thu, 01 Sep 2011 05:08:01 PDT

This script reads all of the parameter files in this directory and its subdirectories to give the user a list of redshifts, code times, universe ages, and time until the final output.

nanoProgress

jwise — Thu, 01 Sep 2011 04:40:09 PDT

nanoProgress parses through a file with enzo's STDOUT for the timesteps on each AMR level. It starts with the last topgrid timestep. After collecting all of these data, it tells the user (a) the total code time elapsed on each level, (b) average timestep, (c) number of timesteps, and (d) total number of timesteps required to finish a topgrid timestep. It also gives the redshift, code time, topgrid dt, and cycle number of the last topgrid timestep.

resubmit

chummels — Mon, 29 Aug 2011 17:55:01 PDT

resubmit is a python script meant for use on NICS kraken with enzo jobs. If you create a submission file for your enzo jobs, rather than qsub'ing that submission file, run resubmit , and resubmit will monitor the job's progress. When the job runs out of time, resubmit will restart it but use the most recent data output as its start point. It will continue to do this until the sim is done, until the max_iterations are reached, or until you manually kill it. Very useful for large jobs that take much longer than kraken's 24-hour job limit.

TPFileConvert

devinsilvia — Wed, 24 Aug 2011 13:36:35 PDT

Since tracer particle data from Enzo simulations are written out into streaming binary files, I've written a C++ script, TPFileConvert.C, that parses the raw binary data and converts it into HDF5 files so that the tracer particle information can be more easily analyzed.

inits_sort - Fix particles for multi-grid Enzo simulations

sskory — Mon, 22 Aug 2011 18:03:00 PDT

For multi-grid (aka nested) Enzo simulations, this script will move particles between the initial conditions files in order to prevent their loss when ring is run.

Stand-alone parallel HOP

sskory — Mon, 22 Aug 2011 18:00:51 PDT

A stand-alone version of parallel HOP.

Enzo Workshop Fall 2011

mturk — Mon, 22 Aug 2011 09:24:28 PDT

This fall there will be an Enzo workshop at Columbia University from October 12-15. If you're interested in getting your hands dirty with Enzo, or learning about how to develop, please come by! Email a developer to add your name to the wiki page if you can't add it yourself.

cyan: a Collection of Yt scripts to Analyze eNzo datasets

mornkr — Sun, 21 Aug 2011 16:00:57 PDT

cyan is an example of an advanced usage of yt, a collection of yt scripts to analyze serial enzo datasets across different ensembles. The functionality includes: total mass, gas fraction by temperature, star formation rate, black hole accretion rate, density and temperature profile, probability distribution function, Kennicutt-Schmidt relation, and some other less advertised ones.

queue_stat - a better view of qstat on Kraken

britton — Sat, 20 Aug 2011 06:16:00 PDT

queue_stat will gather statistics on number of jobs and processors in various states (queued, running, held, complete), as well as cpu hours devoted to each of these (thanks to Cameron Hummels). It will also return information on users with the highest number of queued and running jobs and cpus. All job entries for the user, as well as other users given on the command line, are also given. Works well on NICS Kraken and other supercomputers that use qstat.

yt.volume_rendering

samskillman — Fri, 19 Aug 2011 17:37:44 PDT

This is an example of an advanced usage of the yt volume rendering Camera object. We use both rotation() and move_to() to create a series of images that first rotates the entire volume, then pan and zoom to a location, then rotate around that object. This can be used with the --parallel flag.