Mailman 3 September 2013 - yt-dev

Our last Conda blocker: HDF5
by Matthew Turk 20 Sep '13

20 Sep '13

John ZuHone mentioned in IRC today that this is the last major blocker to using Conda: https://github.com/ContinuumIO/anaconda-issues/issues/18 If anyone has any ideas, they'd be greatly appreciated. Kacper has the yt buildbots set up so we should be able to upload directly to binstar some nightly builds, too.

1 0

YTEP-0018: Changing dict-like access to StaticOutputs
by Matthew Turk 19 Sep '13

19 Sep '13

Hi all, I've issued a pull request for a YTEP that describes a graduated change to how StaticOutput objects work. This shouldn't be taken lightly, as I think it will be relatively disruptive. https://bitbucket.org/yt_analysis/ytep/pull-request/24/ytep-0018-reducing-d… Comments appreciated. -Matt

1 0

Issue #662: Use pip to manage pure python dependencies in the install script (yt_analysis/yt)
by Nathan Goldbaum 18 Sep '13

18 Sep '13

New issue 662: Use pip to manage pure python dependencies in the install script https://bitbucket.org/yt_analysis/yt/issue/662/use-pip-to-manage-pure-pytho… Nathan Goldbaum: I believe that was the cause of Mark Krumholz's issue in this thread: http://lists.spacepope.org/pipermail/yt-users-spacepope.org/2013-September/… This is a somewhat dramatic change compared to what we currently do in the install script, but I think it's necessary. Another option would be to make `yt update --all` a bit smarter or have it `pip uninstall` outdated python packages.

1 0

Issue #661: Add 'gotchas' subsections to the docs (yt_analysis/yt)
by Nathan Goldbaum 18 Sep '13

18 Sep '13

New issue 661: Add 'gotchas' subsections to the docs https://bitbucket.org/yt_analysis/yt/issue/661/add-gotchas-subsections-to-t… Nathan Goldbaum: Where appropriate, we should add subsections to existing doc pages with listings of common 'gotchas'. See for example the sympy docs: http://docs.sympy.org/dev/gotchas.html One example: when someone tries to plot a field that by default has logarithmic colorbar scaling, but with less than an order of magnitude in dynamic range, there will either be one or zero tick marks, making it very difficult to judge the scale. The solution is to turn off field logging but that's not obvious to a new user, so it should be in the docs. I suspect there are many other places in yt where similar gotchas exist.

1 0

Issue #660: Support for interactive MPL backends (yt_analysis/yt)
by Sam Skillman 18 Sep '13

18 Sep '13

New issue 660: Support for interactive MPL backends https://bitbucket.org/yt_analysis/yt/issue/660/support-for-interactive-mpl-… Sam Skillman: We should have better support for operations such as SlicePlot.show() to display in an interactive backend, much like what the PlotCollectionInteractive had. This will need to be done by detecting the backend, determining if it has interactive capability, and reporting useful error messages if not. If "sl" is a SlicePlot instance, the following would be nice to have: ``` #!python sl.show() # Pops up a window for each field. sl.[set_zlim, set_width, set...] # Modifies the existing window(s)/creates new one(s). ```

1 0

Issue #659: Enzo frontend does not respect NumberOfGhostZones (yt_analysis/yt)
by J.S. Oishi 14 Sep '13

14 Sep '13

New issue 659: Enzo frontend does not respect NumberOfGhostZones https://bitbucket.org/yt_analysis/yt/issue/659/enzo-frontend-does-not-respe… J.S. Oishi: If you have WriteGhostZones = 1 and NumberOfGhostZones != 3, then when you retrieve data, you will get the wrong size grid. This is because 3:-3 is hardcoded here: https://bitbucket.org/yt_analysis/yt/src/81c2f6f5f983e90ec58e4d1d9da48a7a4a… Responsible: jsoishi

1 0

Finalizing the license switch
by Matthew Turk 12 Sep '13

12 Sep '13

Hi everyone, I'm writing about the licensing situation. A while back we had a discussion about transitioning to BSD 3-clause. Here is the progress on that: * I have received (offline or online) consent from everyone listed on the spreadsheet. * The ARTIO headers in yt-3.0 will be licensed under the LGPL, which we will note in a file at the top level. * Kacper has moved in the main yt repository all of the external dependencies to an extern/ directory. This also isolates some of the LGPL code. * The rockstar wrapping code will need to be removed before relicensing can occur. (I believe -- Rockstar itself is GPL.) I have created a fork of Rockstar's source code that will bundle the yt interface. * Everyone has consented to the shared copyright model, similar to what IPython does. * One or two additional people have contributed since my initial spreadsheet was made, and I am tracking them down presently. Unless I hear any objections, I will issue pull requests to both 2.x and 3.0 that: * Remove the Rockstar wrapping code * Removes the GPLv3 license, replacing with 3-clause BSD * Removes all attributions that include affiliations and replaces with a collective copyright header, similar to what IPython does: http://ipython.org/ipython-doc/dev/about/license_and_copyright.html I am still not sure if we want authors listed in individual files, but I am certain we do not want the full author/affiliation/years/copyright. * Include the copyright information at the top level of the source repository * Notes that the ExtJS code itself is GPL, but we do not "link" against it directly in the code shipped in the yt source distribution, instead using a <script> tag to include it in an HTML file. I will also update the website and documentation to note the change in license and write a blog post summarizing the discussion. We may also want to issue a point release just for the change in license, although I would prefer to issue a 2.6 at some point with the license change. I'm going to begin preparing this potentially as early as today, but the pull request will not need to be accepted immediately; it can wait, and perhaps even be rejected depending on discussion. Best wishes, Matt

1 3

Issue #658: Update documentation for command-line arguments of yt (yt_analysis/yt)
by Cameron Hummels 11 Sep '13

11 Sep '13

New issue 658: Update documentation for command-line arguments of yt https://bitbucket.org/yt_analysis/yt/issue/658/update-documentation-for-com… Cameron Hummels: There are some arguments that are missing from the documentation: upload_notebook is among these.

1 0

Issue #657: Out-of-date docstring for AMRCoveringGridBase (yt_analysis/yt)
by Andrew Myers 11 Sep '13

11 Sep '13

New issue 657: Out-of-date docstring for AMRCoveringGridBase https://bitbucket.org/yt_analysis/yt/issue/657/out-of-date-docstring-for Andrew Myers: The example in the docs for AMRCoveringGridBase is out-of-date; it no longer accepts a right_edge argument. Responsible: atmyers

1 0

yt-fido/test downtime (09/12)
by Kacper Kowalik 11 Sep '13

11 Sep '13

Hi everybody, there's going to be a scheduled downtime of all services related to automated tests due to electrical grid maintenance @ TCfA All services will be unavailable starting at: Thursday, September 12, 2013 at 07:00 CEST (01:00 EDT, 22:00-1 PDT) lasting up to 6h. Hopefully everything will be back online before most of you wake up. Cheers, Kacper

1 0