Failures when creating annotation plot

[Corey Smart]

I seem to be having some problems running solve-field.  I have some troubleshooting to do on some of them, but one of the issues is that solve-field falls over trying to create the annotation plot for a successful solve.  Error messages from the apod3.jpg demo file are below:

 

Creating index object overlay plot...

Creating annotation plot...

sip_qfits.c:313:read_polynomial: SIP: warning: key "A_0_0" not found; setting to zero.

sip_qfits.c:313:read_polynomial: SIP: warning: key "A_0_1" not found; setting to zero.

sip_qfits.c:313:read_polynomial: SIP: warning: key "A_1_0" not found; setting to zero.

sip_qfits.c:313:read_polynomial: SIP: warning: key "B_0_0" not found; setting to zero.

...

I'm running astrometry on a Mac with OS X 10.9, installed using HomeBrew.  I ran into a couple of issues with the installation of netpbm where it was insistent on using gcc-4.2 and not being able to install dependencies. The only way I found around that was to use the --devel option to install the developer version of netpbm.  (Seems to be a known issue with netpbm.)  I'm not sure if that may be contributing to the this issue?  (Also had an issue where the formula for astrometry was looking for an old version (4.19) of wcslib and couldn't find it on the server...  I manually downloaded the old version of wcslib and put it in the brew cache as a workaround for that.)

Any thoughts on what I might do to fix this?

Thanks in advance,

Corey

[Dustin Lang]

 Hi,

Note that that's not an error, just a "warning".  It's harmless -- I changed the SIP terms that are read in (to read valid but rarely used linear terms), but forgot to also change the code that reports possible problems.

About netpbm, I don't know anything -- haven't tried reinstalling it recently.

I see that the "camphogg" fork of homebrew-science was pointing to an older version of wcslib (4.19 vs 4.20); that is now fixed.  (The wcslib author doesn't keep old releases around, so once he releases a new version, the link to the old version is broken.)

Does solve-field actually create the annotation plot?  The warnings about shouldn't stop it from working.

--dustin

[Corey Smart]

Thanks for the quick reply!  (And awesome work, by the way!)

The code generates an "annotated" .png file, but there are no annotations.  I've attached the -ngc.png output file, where I was expecting to find the annotations...  The background image is there, but no annotations.  The -indx.png file (also attached) doesn't seem to suffer the same fate.

Corey

[Corey Smart]

Hrmmm...  Found some time to play with things some more this evening and tried with a personal image file.  I'd had troubles with this one not solving before, but down sampling by 4 got me a solution.  (And a match to the solution I received on the Flickr group some time back.)  This time the image is correctly annotated. (Seem to be having some browser issues or I would attach it.)

Not sure what's changed, but seems to be doing what I need with the images that are important to me.  Sorry for the earlier distraction and thanks for your help!

Corey

[Dustin Lang]

Ok, glad to hear that's working.  I just tried submitting the apod3 demo to the web service and it comes back with only a couple of annotations, and probably ones with no radius estimate that might not be plotted by the solve-field version.  My guess is it just has no objects in the database of things to annotate in that region (even though there's obviously some stuff there!).

cheers,
--dustin