[yt-dev] Issue #701: Documentation improvements (yt_analysis/yt)

gsiisg issues-reply at bitbucket.org
Tue Oct 29 15:38:43 PDT 2013


New issue 701: Documentation improvements
https://bitbucket.org/yt_analysis/yt/issue/701/documentation-improvements

gsiisg:

Some might be nitpicking or personal preference, but some are just plain typos.  I was not able to read through the entire doc in one sitting, so I did not get to the later sections below Halo Profiling.  Hopefully these are not all already identified errors that others caught.

From
G.S.

yt Bootcamp: A Worked Introduction:

 -> Data Inspection: 
"This section can be skipped!" If someone added comments on what each command line demonstrates, this section would not have to be skipped.  Seems like useful things to know like how to access the data within grids, printing out min/max values of each field, but not sure about the lines and their purpose:
print (g2.Parent.child_mask == 0).sum() * 8
print g2.ActiveDimensions.prod()

-> Simple Visualization
zoom I think is self explanatory, but I think a short description like
p.pan_rel((0.1, 0.0))
"this pans the camera relative to the previous location (x,y) amount", will be helpful.

-> Data Objects and Time Series:
No units associated with three line graphs plotted.

-> Derived Fields and Profiles:
to show how LaTeX friendly the units are, we can change the units of
"Trex/s" to "Trex s$^{-1}$"
and actually print out the units in the subsequent graph that use the field.  Again missing units in the graphs.

The yt Cookbook
-> A Few Complex Plots
Under "Projecting Off-Axis with a Colorbar" there's the following note:

"Please note that this same write_projection function will work with a volume rendering to generate a colorbar in the same fashion."

This ought to have an example, I think showing a volume rendering with a colorbar will be extremely useful.  Maybe it is in the latest version but I do not recall from the dev, or 2.5 version of the docs.

Visualizing Data
-> Plot Modification Mechanisms:
- Thanks to whom ever that put in the available callbacks, this is a huge improvement from previous version of the doc.

-> Using the Manual Plotting Interface:
- Should there be a example of how one may save an .PNG/.JPG and use it in an .EPS file as per the EPS writer tutorial from the workshop material here, or is that functionality being phased out?  At least in ApJ they require all graphs to be in EPS format, but maybe since they're no longer so strict on the file size, outputting it in native eps would be fine.  However, EPS looks pixelated and blurry when you zoom in to look at individual pixels, PNG looks sharp.  Or stick this in the "Generating Processed Data" section, which might be a better place.

-> Volume Rendering:
"As of yt 2.4, this code is threaded using OpenMP. Many of the commands (including snapshot) will accept a num_threads option."
- An example of the commands using this would be helpful, and a hybrid OpenMP+MPI one as well.

"See transfer_functions."
- link missing to transfer_functions

"Unfortunately, due to issues spherical-projection issues,"
- typo too many "issues"

- HEALPix graphic should have LaTeX label, currently it shows "g/cm^2", should switch to "g cm$^{-2}$" or at least "g/cm$^2$"

- Consistency in using "HEALPix" or "HEALpix"

- "using the imshow command:"--> using the imshow() command:

Analyzing Data
-> Using and Manipulating Objects and Fields
- typo "only regions that posses certain properties are of interest." - should be "possesses"

-> Examining and Manipulating Particles
- typo "For simulations only including dark matter particles" -> "including only"

-> Parallel Computation With YT
- typo "if you wanted it to run in parallel on 16 cores (you can always the number of cores you want to run on)"--> you can always select the number of cores

--> Analysis Modules
---->Halo Finding
- parallel HOP is published so no need to link the ArXiv pre-print at "http://adsabs.harvard.edu/abs/2010ApJS..191...43S"


How to Develop yt 
"Numpy is to be imported as na not np. While this may change in the future, for now this is the correct idiom."
- I thought this was changed long ago to np now






More information about the yt-dev mailing list