Could not find assembly error

[aturling]

I think this might be related to #711 but I'm not sure. When trying to create an annotation, we're getting the error:

Error encountered in client: Error: Could not find assembly "68e6e8f55602846d3a515bf5" to add feature "690130fa529523a5375a2f7b". Data may be out of sync, please refresh the page

In the console:

apollo.js:1 Error: Could not find assembly "68e6e8f55602846d3a515bf5" to add feature "690130fa529523a5375a2f7b"
    at Object.addFeature (apollo.js:1:932370)
    at o.<anonymous> (apollo.js:1:187072)
    at d (apollo.js:1:6419)
    at Generator.<anonymous> (apollo.js:1:7769)
    at Generator.next (apollo.js:1:6848)
submit	@	apollo.js:1

The error persists even after refreshing the page, clearing cookies, logging out and logging back in, etc.

I did notice that if we have some annotations preloaded using the apollo feature import command, we don't get the above error. We tested this with running that import command on a small gff3 file containing a single gene on chromosome 1. After that, we can log in and successfully create annotations on chromosome 1, but we get the error above when trying to create annotations on other chromosomes in the same assembly.

Is it required to start with a set of annotations? We don't usually - on the older version of Apollo, the annotation track is empty on a fresh installation until users start creating them.

[garrettjstevens]

It's not required to start with a set of annotations, Apollo 3 is intended to be used with or without first importing features. This is definitely a bug. How are you creating an annotation when you get this error? I've tried to reproduce this and I do get an error message, but the message is different from yours and I want to make sure I can test the same case as you while working on a bug fix.

[aturling]

An update from this morning: today we were getting the error again even though it worked yesterday. What seemed to resolve it today was zooming ALL the way in as far as possible, creating the annotation, then zooming back out.

Our test installation has three assemblies, each with a fasta file and several gff3 files loaded as evidence tracks. We begin by selecting "Linear genome view" and choosing the assembly and chromosome and clicking the "open" button. Next we open the track selector and click on the checkboxes next to the annotation track and one of the gff3 tracks to display them in the main window, and then zoom in enough to remove the "zoom in to see annotations" message. Then we find a gene in the gff3 track, right click it to bring up the menu, and click "Create Apollo annotation". We leave all of the default checkboxes as they are (all checked) and click the "Create" button. If we get the error, it will display immediately in the console, but it takes a few seconds to pop up on the screen (first it says something like "successfully copied..." then after a couple of seconds it displays the "could not find assembly" error).

Another thing I noticed is that if I log in as Guest - we have ours set to read-only - and try to create an annotation, if it's not working it will display the above error message rather than the expected permission denied error. When it's working correctly, the guest account sees a "forbidden" error message which is expected since that account doesn't have permission to create annotations.

[garrettjstevens]

Thanks for the additional info. It turned out my different error message was related to something else, I've been able to reproduce yours now.

I've got a potential fix figured out, so this should be fixed in the next release.

[garrettjstevens]

This should be fixed now in v0.3.9