Here's how I create a_n_d index files from a PS1 database query, from https://github.com/HyperSuprime-Cam/hscMisc/blob/master/ps1dbToAnd.py. I suspect it wouldn't be hard to modify this to pull down SDSS or UCAC or similiar, or to have it operate on our own measurements so as to bootstrap a calibration.

Page in progress may be viewed at: https://confluence.lsstcorp.org/display/LSWUG/Astrometry.net+Index+Files

This tutorial on Confluence is ready for review (see https://confluence.lsstcorp.org/display/LSWUG/Building+Astrometry.net+Index+Files). The one caveat is that general users will only be able to build the index files when the astrometry_net package in the Stack is upgraded to v0.46. Still I mention a clue for how to do this on lsst-dev (by pointing to Simon's private copy of v0.38).

Review comments, these are pretty minor:

1. It would be good to point to the original documentation for reference (http://astrometry.net/doc/build-index.html)
2. wget is not universal. curl is more common, so I'd suggest changing to: curl -O http://www.astro.washington.edu/users/ivezic/sdss/catalogs/stripe82calibStars_v2.6.dat.gz
3. Provide downloadable links for the files you embed. It's a little annoying to have to cut and paste to follow along (if that's what you intend).
4. Why is there no 'hide source' button
5. The script you wrote could pretty easily include the capability to detect whether the file is gzipped an unzip it for the user.
6. You suggest that one can create the andConfig.py by hand, so you should probably point to an example and/or a description of the contents.

Is it worth mentioning hpsplit? Its use seems important for large catalogues.

Good point about hpsplit. It's mentioned in the astrometry.net docs, but not called out specifically in the LSST page.

There is some generic documentation on index building; no need to repeat
that part. Really the only LSST-specific part is re-using the same star
trees (via "build-index -1").

http://astrometry.net/doc/build-index.html

cheers,
--dstn

Simon Krughoff<https://jira.lsstcorp.org/secure/ViewProfile.jspa?name=krughoff>

Simon, thanks for the review. Responses below:

1. Links to the astrometry.net website already existed under the subheadings:

• Build a FITS Binary Table (one link)
• Build Indexes (3 links)

Another was added to the new info block (Split the Sky into Pieces) in:

• Build Indexes

Where else would you recommend links?

2. Updated download command to curl.

3. The scripts are now also attached to the page and links to download the files are embedded in the text.

4. Regrettably, a "hide source" feature is not offered in Confluence.

5. Files could be compressed in a variety of formats, and the supporting libraries for the various MIME types do not appear to be uniformly rich. The suggested capability may be added in a future rev of this page.

6. Example of andConfig.py added.

7. (from Paul) hpsplit is now mentioned in an Info block. It is not necessary for the catalog used in this tutorial, and would clutter the narrative to explain it fully. So I've mostly deferred to the Astrometry.net page. I can always embellish this tutorial later if users stumble on this point.

Re: bullet #1

I don't know how I missed those. Thanks for pointing out their existence. I don't think anything others are necessary.