Mailman 3 October 2016 - yt-dev

Forward pull request created e-mails to the yt-dev mailing list
by Nathan Goldbaum 28 Oct '16

28 Oct '16

Hi all, I've noticed recently a distinct downturn in the number of pull request reviews. That's not to implicate anyone and is not a value judgement on anyone's contribution. It simply reflects that people have other things to work and focus on. I happily acknowledge that doing code review regularly is a tall order. I'd like to make it so there's more of a culture of doing "drive-by" code review. I think one of the main reasons only a few people regularly do code review is that only a few people want to brave the e-mail onslaught and subscribe to all pull request notifications on the yt_analysis/yt repository. Instead, I'd like to propose a hopefully lower noise alternative: forward notifications about new pull requests to this mailing list. I think it's straightforward to add this via a webhook on the yt_analysis/yt repository. We already have similar notifications when an issue is created. This is good and it encourages more people to look at new issues and offer a helping hand if they know what to do. I think having a similar setup for pull requests will encourage more people to regularly pop in and take note of new code contributions since they'll be newly aware that they are happening. Let me know what you think, especially if you are against the proposal. -Nathan

6 5

[ANN] yt-3.3.2 released
by Nathan Goldbaum 28 Oct '16

28 Oct '16

The yt community is proud to announce the release of yt 3.3.2! yt (http://yt-project.org) is an open source, community developed toolkit for the analysis and visualization of volumetric data of all types, with a particular emphasis on astrophysical and nuclear engineering simulations. yt 3.3.2 is a bugfix release that includes a number of fixes and enhancements for issues discovered after the release of yt 3.3.1. We urge all yt users to update to the latest version. If your python installation is managed via conda, update following: $ conda update -c conda-forge yt If your python installation is managed via pip, update following: $ pip install -U yt Finally, if you manage your yt installation by running out of a clone of the yt mercurial repository, update following: $ yt update We would like to thank the following people for their contributions to this release: Corentin Cadiou Nathan Goldbaum Cameron Hummels Suoqing Ji Ben Keller Chang-Goo Kim Kacper Kowalik Alex Lindsay Andrew Myers John Regan Mark Richardson Rafael Ruggiero Hsi-Yu Schive Britton Smith Matthew Turk John Wise Michael Zingale John ZuHone A summary of changes in this release, along with the number of the pull request implementing the change follows below: * Ensure package data files are installed to the correct location (PR 2425) * Ensure particle profile date are interpreted in the correct units for Enzo data (PR 2423) * Work around an issue with reading FLASH 2.5 data using the latest release of h5py. (PR 2421) * Update docstrings for RockstarHaloFinder (PR 2416) * Convert the axis argument of the argmax function to a list if it is not one already (PR 2414) * Make the annotate_clear callback return an instance of the plot for improved user experience in the Jupyter notebook (PR 2413) * Explicitly use integer divion to avoid reading in incorrect grid data in python3 (PR 2412) * Suppress logging when searching for datasets in a simulation time series (PR 2409) * Update instructions on how to set the log level for the yt logger in the documentation (PR 2400) * Various bugfixes for the light ray functionality. These fixes will impact answers generated by the Trident code. (PR 2321, PR 2323, PR 2342, PR 2344, PR 2345, PR 2399) * Don't use /apjs macro in suggested citation (PR 2398) * Add new fields to the list of known fields in the Enzo frontend (PR 2397) * Update configuration for nose so that absolute paths to tests work correctly (PR 2395) * Add mention of ytree to the merger tree documentation (PR 2393) * Install netcdf by default in the install script (PR 2389) * Fix a units issue in the implementation of ds.box (PR 2387) * Disable a randomly failing test (PR 2384) * Fix zoom for the perspective lens (PR 2381) * Fix test failures in a 32 bit build of yt (PR 2380) * Avoid a zero division error in the volume renderer (PR 2379, PR 2388) * Don't take the absolute value of the redshift for the timestamp PlotWindow callback (PR 2373) * Don't try to read non-existent tipsy auxiliary files (PR 2370) * Fix a type issue that causes load_octree to crash (PR 2362) * Performance improvements for the get_vertex_centered_data function (PR 2341) * Performance improvements for the particle_trajectories analysis module (PR 2337) *The LaTeX representation of yt units with numeric multiplicative factors has been improved (PR 2310) * The annotate_cell_edges PlotWindow callback now produces output that is more uniform as a function of image resolution (PR 2307) * Fix a number of issues in the example skeleton frontend (PR 2357) * Mention the availability of demo.use.yt in the quickstart guide (PR 2358) * A realoaded ytdata dataset generated from a profile object can now be plotted with ProfilePlot or PhasePlot (PR 2357) * Make smoothed_covering_grid work correctly with 1D and 2D data. (PR 2354) * Fail more gracefully if yt loads a FLASH dataset with accompanying particle file and the particle file is not from the same simulation time as the plot file. (PR 2353) * Fix an issue with locally defined derived fields not generating correct colorbar labels (PR 2351) * Avoid mutating a YTArray passed into the YTArray initializer (PR 2347) * RAMSES fields and parameters with units of time are now read in correctly for cosmological simulations (PR 2346, PR 2374) * Print a nicer error message when trying to create a profile object that mixes particle and mesh fields (PR 2338) * Ensure the FixedResolutionBuffer object associated with PlotWindow plots is regenerated before it is used again if it is invalidated. (PR 2335) * Fix an issue that would cause incorrect units to be used if units are passed in at dataset load time with units_override (PR 2334, PR 2407) * Add gravitational_potential as a known field in the Athena frontend (PR 2333) * Make 'metallicity' a known field in the stream frontend (PR 2331) * Prevent a segmentation fault in the volume renderer by accounting for round-off error. (PR 2328) * Minor fix for the documentation of the Streamline functionality (PR 2317) * Spelling fixes in the PlotWindow documentation (PR 2313) * Add metadata for a number of new castro fields (PR 2312) * Fix an issue with converting units from the latest version of the Pint package to yt units (PR 2309) * Halo properties written out by rockstar-galaxies can now be read in by the rockstar frontend (PR 2306) * The git hash of the version of castro used to produce a castro dataset can now be queried (PR 2305) * Fix a bug in scene.show() that might cause volume rendering images to now show up inline in the notebook (PR 2304) * Various fixes for the install script (PR 2308, PR 2316, PR2311, PR 2327)

1 0

Issue #1294: yt cannot read some gizmo fields (yt_analysis/yt)
by Onur Catmabacak 28 Oct '16

28 Oct '16

New issue 1294: yt cannot read some gizmo fields https://bitbucket.org/yt_analysis/yt/issues/1294/yt-cannot-read-some-gizmo-… Onur Catmabacak: Hi I am a phd student UZH institution of computational sciences. I am using yt to read hdf5 files created with GIZMO code. I can read vector fields like coordinates or velocities but i am not able to read MagneticField and DivBcleaningFunctionGradPhi fields. It always gives me this error could not broadcast input array from shape (N,3) into shape (N) and no matter what i did (reshape, taking transpose of matrix or any thing you can find on google), i could not overcome the problem. Below you can find a initial condition file and my code and error as an attachment. http://www.tapir.caltech.edu/~phopkins/sims/briowu_ics.hdf5 Thank you

1 0

Issue #1293: trouble opening FITS data (yt_analysis/yt)
by emmanuel donate 24 Oct '16

24 Oct '16

New issue 1293: trouble opening FITS data https://bitbucket.org/yt_analysis/yt/issues/1293/trouble-opening-fits-data emmanuel donate: I went to open a FITS file from my computer and got an error : I typed this: ds = yt.load('NGC_3184_CUBE.FITS') and got this: YTOutputNotIdentified Traceback (most recent call last) <ipython-input-16-092c37fd13ed> in <module>() ----> 1 ds = yt.load('NGC_3184_CUBE.FITS') /Users/emmanueldonate/anaconda/envs/astro/lib/python3.5/site-packages/yt/convenience.py in load(*args, **kwargs) 74 mylog.error("None of the arguments provided to load() is a valid file") 75 mylog.error("Please check that you have used a correct path") ---> 76 raise YTOutputNotIdentified(args, kwargs) 77 for n, c in types_to_check.items(): 78 if n is None: continue YTOutputNotIdentified: Supplied ('NGC_3184_CUBE.FITS',) {}, but could not load! I am in the folder the fits file is in, i'm using jupyter nb

1 0

Issue #1292: Add the equivalency keyword argument to `PlotWindow.set_unit` (yt_analysis/yt)
by Hsi-Yu Schive 24 Oct '16

24 Oct '16

New issue 1292: Add the equivalency keyword argument to `PlotWindow.set_unit` https://bitbucket.org/yt_analysis/yt/issues/1292/add-the-equivalency-keywor… Hsi-Yu Schive: Currently set_unit() does not support converting units using the "equivalency" keyword. So one cannot convert, for example, a temperature field stored in kelvin to keV even though there is a well-defined conversion. As @ngoldbaum suggests, it might be convenient to support something like ``` #!python import yt ds = yt.load(…) plot = yt.SlicePlot(ds, 2, 'temperature') plot.set_unit('temperature', 'keV', equivalency='thermal') ```

1 0

Issue #1291: Simulation parameters in FLASH 2.5 datasets are parsed incorrectly. (yt_analysis/yt)
by Kacper Kowalik 19 Oct '16

19 Oct '16

New issue 1291: Simulation parameters in FLASH 2.5 datasets are parsed incorrectly. https://bitbucket.org/yt_analysis/yt/issues/1291/simulation-parameters-in-f… Kacper Kowalik: As reported by Philipp Girichidis on [yt-users](http://lists.spacepope.org/pipermail/yt-users-spacepope.org/2016-… yt currently fails to properly load FLASH2.5 datasets. Step to reproduce: 1. `curl -JO http://use.yt/upload/1106d6c4` 2. `python -c 'import yt; yt.load("DensTurbMag_hdf5_plt_cnt_0015")'` Current result: ``` yt : [INFO ] 2016-10-19 10:19:06,651 Parameters: current_time = 4.66839074012e-313 yt : [INFO ] 2016-10-19 10:19:06,652 Parameters: domain_dimensions = [1113971554 -402653184 1108822134] yt : [INFO ] 2016-10-19 10:19:06,652 Parameters: domain_left_edge = [ -2.00000000e+18 -2.00000000e+18 -2.00000000e+18] yt : [INFO ] 2016-10-19 10:19:06,652 Parameters: domain_right_edge = [ 2.00000000e+18 2.00000000e+18 2.00000000e+18] yt : [INFO ] 2016-10-19 10:19:06,652 Parameters: cosmological_simulation = 0.0 ``` Expected result: ``` yt : [INFO ] 2016-10-19 10:20:38,881 Parameters: current_time = 751000000000.0 yt : [INFO ] 2016-10-19 10:20:38,882 Parameters: domain_dimensions = [8 8 8] yt : [INFO ] 2016-10-19 10:20:38,882 Parameters: domain_left_edge = [ -2.00000000e+18 -2.00000000e+18 -2.00000000e+18] yt : [INFO ] 2016-10-19 10:20:38,882 Parameters: domain_right_edge = [ 2.00000000e+18 2.00000000e+18 2.00000000e+18] yt : [INFO ] 2016-10-19 10:20:38,883 Parameters: cosmological_simulation = 0.0 ```

1 0

One more set of eyes on a PR?
by Nathan Goldbaum 17 Oct '16

17 Oct '16

Hi all, Can I get one more person to look over and merge this PR: https://bitbucket.org/yt_analysis/yt/pull-requests/2387/dont-cast-ndarrays-… It fixes an issue that a real user ran into and it's been available to review for almost a month now. -Nathan

1 0

Lightray problem-solved
by Pengfei Chen 12 Oct '16

12 Oct '16

Hi all, I sent an email named "Lightray problem" yesterday. After updating yt to the latest tip, the problem is gone. Sorry for the confusion I caused. Thanks, Pengfei

1 0

pyXSIM version 1.1.0
by John Zuhone 11 Oct '16

11 Oct '16

Hi everyone, pyXSIM has been updated to version 1.1.0. This version contains a bugfix and some minor new features. Fixed a bug which did not use the correct file names for AtomDB tables when using TableApecModel. Refactored the absorption model handling into a new class. No user-facing changes have been made. Added special classes for the TBabs and wabs absorption models. De-emphasizing XSpec-based spectral models in favor of the table-based alternatives. For more information see the website: http://hea-www.cfa.harvard.edu/~jzuhone/pyxsim <http://hea-www.cfa.harvard.edu/~jzuhone/pyxsim> pyXSIM can be installed in a few different ways. The simplest way is via the conda package if you have the Anaconda Python Distribution: [~]$ conda install -c jzuhone pyxsim The second way to install pyXSIM is via pip. pip will attempt to download the dependencies and install them, if they are not already installed in your Python distribution: [~]$ pip install pyxsim Alternatively, to install into your Python distribution from source (http://github.com/jzuhone/pyxsim <http://github.com/jzuhone/pyxsim>): [~]$ python setup.py install Thanks for using pyXSIM, John ZuHone

1 0

Lightray problem
by Pengfei Chen 11 Oct '16

11 Oct '16

Hi all, I was trying to generate a long lightray using multiple data sets with yt3, but got a problem. The generated lightray was not continuous (Fig 1 attached). The parameters I use to generate lightray with yt3 are like: lr = LightRay("a.enzo", simulation_type="Enzo", near_redshift=2.0, far_redshift=4.0) But when I switch back to yt2, I got the right result (Fig 2). The parameters I use to generate lightray with yt2 are like: lr = LightRay("a.enzo", "Enzo", z_l, z_h) I use the same random seed and they generate the same segments, but the spectrum are so different. I suspect that it's still due to the unit issue ( https://bitbucket.org/yt_analysis/yt/issues/1258/potential-unit-error-in-co…). Hopefully someone could take a look. Thanks, Pengfei

1 0