Mailman 3 July 2014 - yt-dev

Online Docs Cookbook problems
by Cameron Hummels 21 Jul '14

21 Jul '14

Looking at the yt-3.0 docs that are online, it appears that something has gone wrong in the build process. I see that Kacper has re-enabled the cookbook build process, but now it seems that every recipe is also producing a funky density plot from a random AMR pf: http://yt-project.org/docs/dev-3.0/_images/simple_slice__AMRGridData_Slice_… It *does* look cool, but I think we should figure out what is producing it. Here is one of the cookbook pages where it shows up for every recipe: http://yt-project.org/docs/dev-3.0/cookbook/simple_plots.html -- Cameron Hummels Postdoctoral Researcher Steward Observatory University of Arizona http://chummels.org

2 2

Issue #863: mach_number field has units "code_length" (yt_analysis/yt)
by Cameron Hummels 20 Jul '14

20 Jul '14

New issue 863: mach_number field has units "code_length" https://bitbucket.org/yt_analysis/yt/issue/863/mach_number-field-has-units-… Cameron Hummels: I realize that mach_number is unlikely to work with units of "mach" given that it is a dynamical unit, but it should have a dimensionless unit as opposed to "code_length". i don't know how to set dimensionless units.

1 0

Hangout next week
by Matthew Turk 19 Jul '14

19 Jul '14

Hi all, Anybody up for a hangout/sprint either Friday or next week? Looks to me like there's almost nothing besides docs and existing outstanding PRs before we can release 3.0, so it'd be good to have a sprint where we just went over the docs and PRs and got it in a workable state. Trello has been fastidiously updated by various people, so if you're looking for things that need doing: https://trello.com/b/Y5XV4Hod/yt-3-0 -Matt

5 9

Change to defaults of ProjectionPlot and OffAxisProjectionPlot
by Cameron Hummels 19 Jul '14

19 Jul '14

Hey everyone, I've created a pull request which changes the defaults of the ProjectionPlot and OffAxisProjectionPlot, although I'm looking for feedback from the community. Right now, when you create a projection, say for "Density", it labels the colorbar with "Density" and then gives its projected units (instead of g/cm^3, it gives g/cm^2). My PR is simply to change the default label to be "Projected <field>" in this case "Projected Density (g/cm^2)". It will do this in the case of non-weighted projections. I think this is the expected behavior and more accurate than the former behavior, but I'm open to discussion from the rest of the dev community. In addition, it might be worthwhile to change the defaults on weighted-projections (e.g. density-weighted temperature projection), to give it and appropriate label as well, but I'm less convinced of this change. Perhaps something like "<weight_field>-Weighted <field> (units)" ? There is also a PR awaiting approval by John Regan that deals with this behavior which will allow users to easily specify whatever label they want for the colorbar, but I thought having a sensible default was appropriate as well. Anyway, what do people think about these potential changes? Cameron -- Cameron Hummels Postdoctoral Researcher Steward Observatory University of Arizona http://chummels.org

7 16

Issue #862: Removal of halo_profiler has broken halo_merger_tree (yt_analysis/yt)
by Kacper Kowalik 19 Jul '14

19 Jul '14

New issue 862: Removal of halo_profiler has broken halo_merger_tree https://bitbucket.org/yt_analysis/yt/issue/862/removal-of-halo_profiler-has… Kacper Kowalik: After pull request #1009 importing `halo_merger_tree` fails with: ``` #!text >>> import yt.analysis_modules.halo_merger_tree.api Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/tmp/yt/yt/analysis_modules/halo_merger_tree/api.py", line 16, in <module> from .merger_tree import \ File "/tmp/yt/yt/analysis_modules/halo_merger_tree/merger_tree.py", line 25, in <module> from yt.analysis_modules.halo_profiler.multi_halo_profiler import \ ImportError: No module named multi_halo_profiler ```

1 0

Issue #861: Plotting random particle data produces broadcasting error (yt_analysis/yt)
by Nathan Goldbaum 19 Jul '14

19 Jul '14

New issue 861: Plotting random particle data produces broadcasting error https://bitbucket.org/yt_analysis/yt/issue/861/plotting-random-particle-dat… Nathan Goldbaum: This traceback comes from executing the last cell of the loading generic particle data notebook in the docs: ``` #!python traceback /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/visualization/plot_window.pyc in SlicePlot(ds, normal, fields, axis, *args, **kwargs) 1883 del kwargs['north_vector'] 1884 -> 1885 return AxisAlignedSlicePlot(ds, normal, fields, *args, **kwargs) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/visualization/plot_window.pyc in __init__(self, ds, axis, fields, center, width, axes_unit, origin, fontsize, field_parameters, window_size, aspect) 1089 slc = ds.slice(axis, center[axis], 1090 field_parameters = field_parameters, center=center) -> 1091 slc.get_data(fields) 1092 PWViewerMPL.__init__(self, slc, bounds, origin=origin, 1093 fontsize=fontsize, fields=fields, /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in get_data(self, fields) 633 634 fields_to_generate += gen_fluids + gen_particles --> 635 self._generate_fields(fields_to_generate) 636 for field in self.field_data.keys(): 637 if field not in ofields: /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in _generate_fields(self, fields_to_generate) 653 fi = self.ds._get_field_info(*field) 654 try: --> 655 fd = self._generate_field(field) 656 if type(fd) == np.ndarray: 657 fd = self.ds.arr(fd, fi.units) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in _generate_field(self, field) 255 tr = self._generate_particle_field(field) 256 else: --> 257 tr = self._generate_fluid_field(field) 258 if tr is None: 259 raise YTCouldNotGenerateField(field, self.ds) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in _generate_fluid_field(self, field) 273 finfo.check_available(gen_obj) 274 except NeedsGridType as ngt_exception: --> 275 rv = self._generate_spatial_fluid(field, ngt_exception.ghost_zones) 276 else: 277 rv = finfo(gen_obj) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in _generate_spatial_fluid(self, field, ngz) 289 o = self._current_chunk.objs[0] 290 with o._activate_cache(): --> 291 ind += o.select(self.selector, self[field], rv, ind) 292 else: 293 chunks = self.index._chunk(self, "spatial", ngz = ngz) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in __getitem__(self, key) 218 return self.field_data[f] 219 else: --> 220 self.get_data(f) 221 # fi.units is the unit expression string. We depend on the registry 222 # hanging off the dataset to define this unit object. /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in get_data(self, fields) 633 634 fields_to_generate += gen_fluids + gen_particles --> 635 self._generate_fields(fields_to_generate) 636 for field in self.field_data.keys(): 637 if field not in ofields: /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in _generate_fields(self, fields_to_generate) 653 fi = self.ds._get_field_info(*field) 654 try: --> 655 fd = self._generate_field(field) 656 if type(fd) == np.ndarray: 657 fd = self.ds.arr(fd, fi.units) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in _generate_field(self, field) 255 tr = self._generate_particle_field(field) 256 else: --> 257 tr = self._generate_fluid_field(field) 258 if tr is None: 259 raise YTCouldNotGenerateField(field, self.ds) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in _generate_fluid_field(self, field) 275 rv = self._generate_spatial_fluid(field, ngt_exception.ghost_zones) 276 else: --> 277 rv = finfo(gen_obj) 278 return rv 279 /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/fields/derived_field.pyc in __call__(self, data) 182 "for %s" % (self.name,)) 183 with self.unit_registry(data): --> 184 dd = self._function(self, data) 185 for field_name in data.keys(): 186 if field_name not in original_fields: /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/fields/particle_fields.pyc in particle_cic(field, data) 114 115 def particle_cic(field, data): --> 116 pos = data[ptype, coord_name] 117 d = data.deposit(pos, [data[ptype, mass_name]], method = "cic") 118 d = data.apply_units(d, data[ptype, mass_name].units) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/octree_subset.pyc in __getitem__(self, key) 74 75 def __getitem__(self, key): ---> 76 tr = super(OctreeSubset, self).__getitem__(key) 77 try: 78 fields = self._determine_fields(key) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in __getitem__(self, key) 218 return self.field_data[f] 219 else: --> 220 self.get_data(f) 221 # fi.units is the unit expression string. We depend on the registry 222 # hanging off the dataset to define this unit object. /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/data_objects/data_containers.pyc in get_data(self, fields) 627 628 read_particles, gen_particles = self.index._read_particle_fields( --> 629 particles, self, self._current_chunk) 630 for f, v in read_particles.items(): 631 self.field_data[f] = self.ds.arr(v, input_units = finfos[f].units) /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/geometry/geometry_handler.pyc in _read_particle_fields(self, fields, dobj, chunk) 231 self._chunk_io(dobj, cache = False), 232 selector, --> 233 fields_to_read) 234 for field in fields_to_read: 235 ftype, fname = field /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/utilities/io_handler.pyc in _read_particle_selection(self, chunks, selector, fields) 172 #mylog.debug("Filling %s from %s to %s with %s", 173 # field_f, my_ind, my_ind+vals.shape[0], field_r) --> 174 rv[field_f][my_ind:my_ind + vals.shape[0],...] = vals 175 ind[field_f] += vals.shape[0] 176 # Now we need to truncate all our fields, since we allow for ValueError: could not broadcast input array from shape (5000000) into shape (521728) ```

1 0

Issue #860: One of the animation recipes in the docs build raises the following error: (yt_analysis/yt)
by Nathan Goldbaum 19 Jul '14

19 Jul '14

New issue 860: One of the animation recipes in the docs build raises the following error: https://bitbucket.org/yt_analysis/yt/issue/860/one-of-the-animation-recipes… Nathan Goldbaum: --------------------------------------------------------------------------- AttributeError Traceback (most recent call last) <ipython-input-4-e3b9f89be543> in <module>() 12 13 # call the animator. blit=True means only re-draw the parts that have changed. ---> 14 anim = animation.FuncAnimation(fig, animate, frames=44, interval=200, blit=False) 15 16 # call our new function to display the animation /home/xarth/yt-x86_64/lib/python2.7/site-packages/matplotlib-1.3.0-py2.7-linux-x86_64.egg/matplotlib/animation.pyc in __init__(self, fig, func, frames, init_func, fargs, save_count, **kwargs) 1008 self._save_seq = [] 1009 -> 1010 TimedAnimation.__init__(self, fig, **kwargs) 1011 1012 # Need to reset the saved seq, since right now it will contain data /home/xarth/yt-x86_64/lib/python2.7/site-packages/matplotlib-1.3.0-py2.7-linux-x86_64.egg/matplotlib/animation.pyc in __init__(self, fig, interval, repeat_delay, repeat, event_source, *args, **kwargs) 862 863 Animation.__init__(self, fig, event_source=event_source, --> 864 *args, **kwargs) 865 866 def _step(self, *args): /home/xarth/yt-x86_64/lib/python2.7/site-packages/matplotlib-1.3.0-py2.7-linux-x86_64.egg/matplotlib/animation.pyc in __init__(self, fig, event_source, blit) 543 544 # Clear the initial frame --> 545 self._init_draw() 546 547 # Instead of starting the event source now, we connect to the figure's /home/xarth/yt-x86_64/lib/python2.7/site-packages/matplotlib-1.3.0-py2.7-linux-x86_64.egg/matplotlib/animation.pyc in _init_draw(self) 1033 # artists. 1034 if self._init_func is None: -> 1035 self._draw_frame(next(self.new_frame_seq())) 1036 else: 1037 self._drawn_artists = self._init_func() /home/xarth/yt-x86_64/lib/python2.7/site-packages/matplotlib-1.3.0-py2.7-linux-x86_64.egg/matplotlib/animation.pyc in _draw_frame(self, framedata) 1047 # Call the func with framedata and args. If blitting is desired, 1048 # func needs to return a sequence of any artists that were modified. -> 1049 self._drawn_artists = self._func(framedata, *self._args) <ipython-input-4-e3b9f89be543> in animate(i) 9 def animate(i): 10 ds = load('Enzo_64/DD%04i/data%04i' % (i,i)) ---> 11 prj._switch_ds(ds) 12 13 # call the animator. blit=True means only re-draw the parts that have changed. /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/visualization/plot_container.pyc in _switch_ds(self, new_ds, data_source) 282 self._data_valid = self._plot_valid = False 283 self._recreate_frb() --> 284 self._setup_plots() 285 286 def __getitem__(self, item): /var/lib/jenkins/jobs/yt-docs-3.0/workspace/yt/visualization/plot_window.pyc in _setup_plots(self) 844 hinv = False 845 for i, un in enumerate((unit_x, unit_y)): --> 846 if hasattr(self.ds.coordinates, "default_unit_label"): 847 axax = getattr(self.ds.coordinates, "%s_axis" % ("xy"[i]))[axis_index] 848 un = self.ds.coordinates.default_unit_label[axax] AttributeError: 'YTQuadTreeProj' object has no attribute 'coordinates' Responsible: ngoldbaum

1 0

Long live "Dataset"
by Matthew Turk 19 Jul '14

19 Jul '14

Hi all, A big pull request, pushed by Nathan and accepted by Britton, has globally changed "pf" (an enzo-ism for "parameter file") to "ds" (for "dataset") everywhere in the code base. Fortunately, there's a backcompat layer, but from here on out, long live "ds"! -Matt

2 1

Big pf => ds PR
by Matthew Turk 18 Jul '14

18 Jul '14

Hi all, Nathan's big pf=>ds pull request has been outstanding for some time now, and it's ready to go. https://bitbucket.org/yt_analysis/yt/pull-request/954/replacing-pf-with-ds Unless someone objects, I would like to bring this in before the doc sprint today. -Matt

3 5

Issue #859: HaloCatalog tries to use non exisiting field (yt_analysis/yt)
by Kacper Kowalik 18 Jul '14

18 Jul '14

New issue 859: HaloCatalog tries to use non exisiting field https://bitbucket.org/yt_analysis/yt/issue/859/halocatalog-tries-to-use-non… Kacper Kowalik: Following script: ``` #!python from yt.mods import load from yt.analysis_modules.halo_analysis.api import HaloCatalog import tempfile import os # Create temporary directory for storing files tmpdir = tempfile.mkdtemp() # Load the data set with the full simulation information data_pf = load('Enzo_64/RD0006/RedshiftOutput0006') # Load the rockstar data files halos_pf = load('rockstar_halos/halos_0.0.bin') # Instantiate a catalog using those two paramter files hc = HaloCatalog(data_pf=data_pf, halos_pf=halos_pf, output_dir=os.path.join(tmpdir, 'halo_catalog')) ``` fails with current tip (d44a659ea62c) yielding ``` #!text Traceback (most recent call last): File "ala.py", line 17, in <module> output_dir=os.path.join(tmpdir, 'halo_catalog')) File "yt/analysis_modules/halo_analysis/halo_catalog.py", line 139, in __init__ self.add_default_quantities() File "yt/analysis_modules/halo_analysis/halo_catalog.py", line 448, in add_default_quantities self.add_quantity("virial_radius", field_type=field_type,prepend=True) File "yt/analysis_modules/halo_analysis/halo_catalog.py", line 218, in add_quantity raise RuntimeError("HaloCatalog quantity must be a registered function or a field of a known type.") RuntimeError: HaloCatalog quantity must be a registered function or a field of a known type. ``` Regression was introduced in 39697406e012 *viral_radius* was renamed to *radius* in `yt/frontends/halo_catalogs/rockstar/definitions.py`, however `yt/analysis_modules/halo_analysis/halo_catalog.py:add_default_quantities` still tries to use it.

1 0