Hey everyone,

We finished up the doc sprint today.  I think we made some big improvements to the doc base in the last two days with help from: Sam Skillman, Geoffrey So, Nathan Goldbaum, John Zuhone, Jeff Oishi, Britton Smith, Matt Turk, and me.  Thanks everyone!  But we're not done with everything yet.  By my tally, everyone still has a task to complete, and there are still some open tickets on the bitbucket docs issues list.  In this email, I'll try to summarize what we did, what remains to be done, and what the timeline forward is.

First of all, the latest docs exist in my PR here:

There is a build of this available for viewing here:  

--Restructured the doc layout to be more hierarchical, so that now all of the docs fall into the following categories: install/, help/, bootcamp/, cookbook/, development/, examine/, visualize/, analyze/, and reference/.
--Restructured the top-level page to reflect this, resulting in a much simpler and cleaner starting point.
--Added support for embedded notebooks, embedded executed scripts, and embedded notebook snippets in the docs which get executed and built dynamically with the docs.
--Updated what codes are supported in 2.x (removed support for gadget, ramses and art, since these don't work well in 2.x)
--Updated the installation instructions
--Moved the bootcamp from its own repository into the docs as the main introduction to the code using embedded notebooks.
--Updated the "How to get help" to be "how to solve problems" with several suggestions that one can do on their own before turning to the mailing list.
--Split the cookbook into Example Scripts and Example Notebooks (longer demonstrations of more complex functionality).
--Updated the "How to Make Plots" and "Plot Modification" parts of the visualization docs making extensive use of the new embedded scripts and embedded notebook snippets.  Looks great!
--Added a section on how to make SZ synthetic observations on one's datasets using the new embedded notebook
--Getting rid of the old workshop lessons
--Removing the bitbucket wiki, since it contained outdated and conflicting redundant information from the docs
--Numerous updates and typo fixes

(see open documentation issues here: https://bitbucket.org/yt_analysis/yt/issues?component=documentation&status=open&status=new )
--Explain use of CuttingRegion/FieldCuts -- Britton
--Write up ProfilePlot docs -- Britton
--Make PhasePlot work in yt and document it -- Matt and Britton
--Introduce how to use a notebook -- Nathan
--Update VR docs -- Sam
--Make a VR notebook with a great example -- Sam
--Demonstrate how to bring in arbitrary grid datasets into the code -- John
--Purge, fix, move dangling links in the code from the restructure -- Cameron
--Describe how to export to sketchfab -- Matt
--Add section in developer docs on "How to dig into the source" --Cameron
--Describe how to get the sample datasets and set up test_dir in config.
--Document how to do VR with two fields (cookbook)
--Update docs for command-line arguments
--Document cosmology units

I think we need to finish these things before the release of 2.6.  I think if people continue to work on these issues with the same momentum we've had the last couple of days, we'll get this finished in the next week.  

Overall, I think the docs are looking a lot better at this point.  Let us know if you have any additional comments/suggestions.  Thanks again to those of you who put in a lot of hard work the last couple of days!


On Mon, Oct 28, 2013 at 6:16 PM, Cameron Hummels <chummels@gmail.com> wrote:
Hey everyone,

This is a status update after day one of the doc sprint.  We had a good turnout with Matt Turk, Sam Skillman, John Zuhone, Jeff Oishi, Nathan Goldbaum, and myself present.  We were able to get a lot done, some synchronously, some asynchronously, over the course of the day.

Work has begun on restructuring the docs to make them more easily navigable.  The front page has been cleaned up to have a limited number of top-level options for users.  The installation page has been updated.  The cookbook repository has been merged into the docs inline so that new users can view (or execute) that as their introduction to the functionality of yt.  Minor changes have occurred on the help pages and the developer pages.  Significant effort has gone into making the docs more succinct and having fewer overlap points (where some concept is explained more than once).  The workshop materials were removed and many of the former top-level directories (e.g. advanced) have been split up and redistributed into the other code directories.  A new sample dataset was added to the yt-project.org/data site by Nathan of one of his high-res isolated galaxy sims--it's almost 3 GB, but it's a beautiful dataset for demonstrating some cool functionality with real data.

I've issued a PR to the yt-doc repo with many of these changes here: 

You can view this version of the docs looks here (minus the api docs and the bootcamp renderings):

We seemed to settle on grouping yt functionality into 3 tiers: visualizing data, analyzing data, and examining data (accessing it at a low level).  Sam Skillman made a cool illustration of this here: https://docs.google.com/drawings/d/15aWmO71uHaxjx0qK5bMGxHir9wPI27Cgk-qG03Y00nc/edit .  The plan is to have the docs reflect this hierarchy and to have it show up in the introductory statements on the front page.

--Nathan - Creating a set of notebooks to demonstrate how to use the plot window and plot modifications effectively.  This will update the entire plotting interface portion of the docs.  Also, create an introduction to how to use notebooks from scratch.
--Jeff & Sam - Writing a notebook demonstrating how to use it for doing real science, not just merely as a toy.
--Cameron - Going over the top-level docs with a fine-toothed comb looking to update/clarify.  Restructuring the hierarchy of docs.
--John - Updating the Athena frontend; Documenting SZ analysis and particle trajectories in the analysis modules section.
--Matt - Restructuring the hierarchy of docs.  Purging out-of-date information.  Picking a new theme for the docs.

But there is still a lot left to do.  For those of you who want to participate but don't yet know how you can assist, you can do a few things.  
--Look over the docs and see if you find anything that needs fixing, pruning, updating, etc.  Add an issue in the yt_analysis/yt repo and tag it with documentation, so we know to fix it.  Or create an issue and fix it yourself and PR it.  Really, the docs just need some attention to clean them up.
--Look at the existing documentation issues on bitbucket.  Address one of them.

I think we settled on a solution to the notebook vs script debate.  Continue to use scripts for the simpler tasks in the cookbook and use notebooks for more complex examples.  Right now I've split the cookbook into Example Scripts and Example Notebooks for this task.  Tomorrow will be spent adding more coding examples to both of these subdivisions.

Finish restructuring the docs.  Add lots more recipes (both notebook and scripts) to the cookbook.  Address more of the issues from the bb list.

If you want to join us, come by IRC or google+ at 8amPST tomorrow, and we'll put you to good use.


On Sat, Oct 26, 2013 at 9:40 AM, Britton Smith <brittonsmith@gmail.com> wrote:
Hi Cameron,

Thanks for organizing this!  This will be very useful I think.  I will be traveling for most of Monday during the time of the sprint but should be able to participate on Tuesday.  It would be useful if you could provide a short summary of the events of the first day for those of us who can only join in on the second day.  Thanks again for doing this.  See you all there.


On Thu, Oct 24, 2013 at 2:16 AM, Cameron Hummels <chummels@gmail.com> wrote:
Hey everyone,

Next Monday and Tuesday we will be having a documentation sprint to try to beef up the yt docs (and docstrings) before the final release of the 2.x branch when development (and documentation) will switch over to the 3.0 branch.  All developers are welcome to attend/participate for as much time as they can spare.  

I know documentation isn't always glamorous, but I think this will be pretty fun.  Furthermore, I think this will be very beneficial to the community, in that it will be fewer frustrated users, fewer questions on IRC and the mailing list, and better understanding for everyone of all the features and functionality of the codebase!

We will be meeting up as a Google + hangout over the course of those two days, with periodic "check-in" meeting times to make sure everyone is on the same page (see below for schedule).  Yes, I know it is early for Pacific Coasters on a Monday @ 8am, but you can meet for a few minutes from your bed without your camera turned off if need be.  We all know that people have meetings and need to eat and such, so you are not required to be in the G+ hangout constantly.  That said, people are encouraged to remain in the G+ hangout throughout both days for fast turnover time if they have questions/discussion or want to work on something new.  Documentation tasks will be identified so that attendees can volunteer for them at the various check-in meetings.  This way, individuals can work on these tasks semi-autonomously.  The various things I think we should aim to accomplish include:

--Add documentation for features that are currently undocumented in the yt codebase.
--Add documentation for new features from yt 2.5 to present (in docstrings, cookbooks, and narrative docs as applicable).
--Fulfill the tasks identified in the BitBucket Documentation Issues List: 
--Remove outdated information from the docs.
--Pare down multiple locations of how to do individual tasks, so as to make maintainability easier.
--Fix any typos or mistakes in the content.

--8am PST/11am EST: Initial meetup in G+ to discuss the goals of the sprint and to layout the specific individual tasks which need to be accomplished.  I'll include a template for a docstrings example, a template for a cookbook example, and an example of a good narrative docs section.  Developers will choose what task they want to work on.  Meeting should take 30-60 minutes.
--11am PST/2pm EST: Status check, and reshuffling of projects as needed.  15 minutes.
--2pm PST/5pm EST: Status check and conclusion for the day. 15 minutes.

Same schedule as Monday.

Please write back to this email saying whether or not you're going to be attending any/all of the sprint.  That way, I can invite you personally to the google docs hangout to join up.  I may try to make this hangout "on-air" so that it is recorded, and so if we go above the 10-person limit, those unable to fit in the hangout can watch it streaming live.

Before Monday, please think for a few minutes about things that you may have noticed in the past that were lacking from the documentation, but at the time you found them you didn't have time to fix.  If you identify anything, please mark it down as a documentation task in the bitbucket issue tracker so that we know to work on it next week: https://bitbucket.org/yt_analysis/yt/issues?component=documentation&status=open&status=new . We will be using this as our main means of tracking what gets accomplished during the sprint.  I encourage you to look at the docs as well to look for ways they can be improved.

Thanks everyone, and I look forward to meeting with you next week!


Cameron Hummels
Postdoctoral Researcher
Steward Observatory
University of Arizona

yt-dev mailing list

yt-dev mailing list

Cameron Hummels
Postdoctoral Researcher
Steward Observatory
University of Arizona

Cameron Hummels
Postdoctoral Researcher
Steward Observatory
University of Arizona