Tutorials for release 18.9

weaverba137 commented 5 years ago

This PR is in preparation for getting the tutorials working with release 18.9.

DO NOT MERGE YET!

Below are comments on individual tutorials.

dc17a-truth

Code versions used for this tutorial:
python      : 3.6.6
numpy       : 1.13.3
scipy       : 1.1.0
astropy     : 2.0.7
yaml        : 3.12
matplotlib  : 2.1.2
desiutil    : 1.9.13.dev693
desispec    : 0.23.1.dev2840
specter     : 0.9.0.dev536
speclite    : 0.8

Works, but with the caveat that this is dc17a.

GFA_targets

Looks like the problems described in the note at the top got fixed, because everything worked.
- What should we replace this with? Is the selection still being updated?
Uses DR4 sweeps.
Does get_all_gfa_targets() really need to be that slow, for a tutorial?
get_all_gfa_targets() does import numpy as np, but doesn't actually use it.

Intro_to_DESI_spectra

Testing with /global/cscratch1/sd/sjbailey/desi/dev/end2end.
Lots of RA_TARGET --> TARGET_RA.

Hit failure at

stds = np.where(specobj.fibermap["DESI_TARGET"] & desi_mask["STD_FSTAR"])[0]

What is the new bit in desitarget.targetmask.desi_mask?

redshift-database

Works with truth table fix on /global/cscratch1/sd/sjbailey/desi/dev/end2end.
- Double check on reference run 18.9 when available.
Questions:
- The Moon example now only has one exposure with the Moon up. Is this expected?
- Nights that have data has also changed.

simulating-desi-spectra

make_templates() includes a fourth output, objmeta.

rrdesi fails with:


Running on a batch node:
rrdesi /global/cscratch1/sd/bweaver/desi/simspec/qso-observed-spectra.fits --zbest /global/cscratch1/sd/bweaver/desi/simspec/qso-zbest.fits

srun: job 15438852 queued and waiting for resources srun: job 15438852 has been allocated resources Running with 32 processes Loading targets... Proc 0: Traceback (most recent call last): Proc 0: File "/global/common/software/desi/cori/desiconda/20180709-1.2.6-spec/code/redrock/master/py/redrock/external/desi.py", line 588, in rrdesi comm=comm, cache_Rcsr=True) Proc 0: File "/global/common/software/desi/cori/desiconda/20180709-1.2.6-spec/code/redrock/master/py/redrock/external/desi.py", line 286, in init bname = frow["BRICKNAME"] Proc 0: File "/global/common/software/desi/cori/desiconda/20180709-1.2.6-spec/conda/lib/python3.6/site-packages/astropy/table/row.py", line 46, in getitem return self._table.columns[item][self._index] Proc 0: File "/global/common/software/desi/cori/desiconda/20180709-1.2.6-spec/conda/lib/python3.6/site-packages/astropy/table/table.py", line 109, in getitem return OrderedDict.getitem(self, item) Proc 0: KeyError: 'BRICKNAME'



## survey-simulations

* Uses datachallenge/surveysim2017
* Seems to work fine.

## redrock

* Do we need to test the redrock tutorials?

dkirkby commented 5 years ago

Who is the audience for the GFA_targets notebook? It looks more like (useful) internal documentation than a tutorial on a user-facing API.

moustakas commented 5 years ago

Who is the audience for the GFA_targets notebook? It looks more like (useful) internal documentation than a tutorial on a user-facing API.

I totally agree that we should relegate the GFA_targets notebook. I'm testing it with the 18.7 software release and having a boatload of issues. In addition, I think the fact that it uses DR4 suggests it's due for a major rewrite / update (e.g., in that it doesn't use Gaia).

weaverba137 commented 5 years ago

Would anyone object if I close this PR? I think we'll just need to start from scratch once the next software release comes out (e.g. 18.11).

weaverba137 commented 5 years ago

Having heard no objections, I'm closing without merging.

desihub / tutorials