mm10 is broken
Closed this issue · 10 comments
it's giving this error:
::Loading cache:: /mnt/q/shefflab/LOLAweb/databases//LOLACore/mm10/encodeTFBSmm10/encodeTFBSmm10.RData
Timing stopped at: 4.916 0.044 4.984
Warning: Error in combine_CompressedList_objects: the objects to combine must be CompressedList objects (or NULLs)
Stack trace (innermost first):
96: combine_CompressedList_objects
95: .local
94: c
93: c
92: readRegionGRL
91: loadRegionDB
90: system.time [/srv/shiny-server/LOLAweb/apps/LOLAweb/app.R#578]
89: eventReactiveHandler [/srv/shiny-server/LOLAweb/apps/LOLAweb/app.R#509]
69: rawdat_nocache
68: withCallingHandlers [/srv/shiny-server/LOLAweb/apps/LOLAweb/app.R#385]
67: observeEventHandler [/srv/shiny-server/LOLAweb/apps/LOLAweb/app.R#382]
1: runApp
Warning: Error in combine_CompressedList_objects: the objects to combine must be CompressedList objects (or NULLs)
Stack trace (innermost first):
61: <Anonymous>
60: stop
59: rawdat_nocache
58: observerFunc [/srv/shiny-server/LOLAweb/apps/LOLAweb/app.R#753]
1: runApp
Something wrong with that cache, I think...
docker run --rm -it --volume /mnt/q:/mnt/q databio/lolaweb bash
R
LOLA::loadRegionDB("/mnt/q/shefflab/LOLAweb/databases//LOLACore/mm10/")"
I am trying regenerating the cache now.
that didn't work. it may be a LOLA version thing. can we try updating the dev container to the latest R/bioconductor/LOLA version?
turns out it's not a LOLA version problem, but could be a GenomicRanges version problem. I still think we should update...
ok yeah i think the shinyBase container image we are pulling should include the bioconductor (not github) version of LOLA:
https://github.com/databio/shinyBase/blob/master/Dockerfile#L29
that gets pulled in via the LOLAweb dockerfile ... but i can't remember if that base image gets rebuilt automatically? every time we push a change to databio/shinyBase ? manually?
@nmagee is going to get sick of explaining this to me :)
it only gets rebuilt if we push a change to databio/shinyBase or trigger it manually.
Ok, I rebuilt it locally and pushed changes -- can we trigger a re-deploy of the containers @nmagee to see if this solves the problem?
Yes, that worked! So, turns out I was right, it was probably some kind GRanges version issue. it's working now on mm10.
@nmagee thanks for redeploy -- at some point can you write down some brief instructions or a link somewhere so that in the future I don't have to bother you to do that? I forgot we were using travis for those builds and that was the way to rebuild it...