ncihtan / htan-portal

The HTAN Data Portal
https://humantumoratlas.org
9 stars 11 forks source link

Missing thumbnails and stories #332

Open adamjtaylor opened 2 years ago

adamjtaylor commented 2 years ago

Following manual review on 2021-11-03, the following files are missing either thumbnail or story.

Vanderbilt HE missing both thumbnail and story

OHSU CyCIF fix background in thumbnail

OHSU listed mIHC but actually H&E - missing thumbnail and regenerate minerva as h&E

WASHU IMC missing thumbnail

WASHU IMC need minerva stories

WASHU IMC need thumbnails and minerva stories

adamjtaylor commented 2 years ago

@hweej ... last min request... could you look why the washu assets listed above re not appearing in the mapping. they are now present in the assets bucket but when doing the mapping they still show as null

I think this is related to these basenames appearing in two separate folders as well as in an archive folder.

hweej commented 2 years ago

I'll take a look at it!

adamjtaylor commented 2 years ago

Vanderbilt files failed due to out of Java heap memory errors during conversion to OME-TIFF (pyramid required for both minerva rendering and thumbnails) eg.

2021-11-03 13:50:00,114 [pool-1-thread-1] ERROR c.g.bioformats2raw.Converter - Failure processing chunk; resolution=0 plane=0 xx=39936 yy=17408 zz=0 width=1024 height=1024 depth=1
java.lang.OutOfMemoryError: Java heap space

Increasing limit on AWSBatch machines called by Nextflow Tower not sufficient at the moment!

hweej commented 2 years ago

@adamjtaylor , it seems like a new imc_level_2 subdirectory was created when uploading the latest htan-assets.

# Extra imc_level_2
htan-assets/minerva_stories/htan-dcc-washu/imc_level_2/imc_level_2/batch_1_10072021/ 

# I fixed htan assets back to this
htan-assets/minerva_stories/htan-dcc-washu/imc_level_2/batch_1_10072021/