Details
-
Type:
Bug
-
Status: Won't Fix
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: pipe_tasks
-
Labels:None
-
Story Points:1
-
Team:Alert Production
-
Urgent?:No
Description
The calibrateTask.py example in the $PIPE_TASKS_DIR/examples subdir, which exercises both charImg and calibrate, is not working on a current version of the stack, and exits with an error that ends as so:
...
|
File "/tigress/HSC/LSST/stack_20160915/Linux64/meas_astrom/12.1-7-gfd5edcb+2/python/lsst/meas/astrom/multiindex.py", line 53, in getConfigFromEnvironment
|
raise RuntimeError("Unable to find andConfig.py in astrometry_net_data directory %s" % (anDir,))
|
RuntimeError: Unable to find andConfig.py in astrometry_net_data directory /tigress/HSC/LSST/stack_20160915/Linux64/astrometry_net_data/8.0.0.0+17
|
If one sets up an older version of the stack with e.g.
setup pipe_tasks -t w_2016_51
|
And then runs the example straight (i.e. "python calibrateTask.py"), it executes and exits just fine, so clearly something broke in the last few weeks (that seems to involve a_n_d).
(Btw, it needs afwdata also to be separately installed to run correctly.)
Of course, none of our current examples are guaranteed to be working (and that's part of what the docs team will make sure is right in the future), so i don't know what the prioritization should be here, just pointing it out.
Attachments
Issue Links
- relates to
-
DM-9390 Provide a working CalibrateTask & CharacterizeImageTask example
- Won't Fix
-
- Invalid
-
- Done
-
DM-4126 Include examples in CI
- To Do
-
- Implemented
- mentioned in
-
Page Loading...
Activity
I'm hesitant to try to fix examples that aren't being run in CI. Does anyone have an opinion on what I should do with this?
Well.. from my 'novice' angle, this shouldn't be highly prioritized, so it's fine by me (esp as long as i can make it run with an old version), since we plan to rewrite the examples later too, but perhaps others may care more..?
I have no very strong opinions. I certainly don't think it's worth burning more than a few minutes on — if a fix is trivial, great; if it isn't, no big deal. Maybe we should clearly mark the example as broken or even remove it entirely to avoid potentially confusing folks, though?
This code was removed on DM-9520.
I'm guessing that this was broken by
DM-8232, so setting team to AP.Worth adding that this example shouldn't need a_n_d at all: it has its own custom astrometry class which is designed to be self contained, but it's now broken such that it skips that and looks for a_n_d after all.