`CrossReferencing()` fails with `object 'my.lib' not found` when loading `compound_db.qs`

[hosseinvk]

Summary

I am using MetaboAnalystR 4.2.0 on a Unix/Linux server with R 4.5.1. Running CrossReferencing(mSet, "name") fails when loading compound_db.qs.

Error

mSet <- CrossReferencing(mSet, "name")

Output:

[1] "Reading data unsuccessful, attempting to re-download file..."
100 1009k  100 1009k    0     0   421k      0  0:00:02  0:00:02 --:--:--  421k
[1] "Loading files from server unsuccessful. Ensure curl is downloaded on your computer."
Error in .get.my.lib("compound_db.qs") : object 'my.lib' not found

Checks performed

The downloaded file exists and has a plausible size:

file.exists("compound_db.qs")
file.info("compound_db.qs")$size

Output:

[1] TRUE
[1] 1033560

But it cannot be read with qs::qread():

packageVersion("qs")
packageVersion("stringfish")
qs::qread("compound_db.qs")

Output:

[1] '0.25.5'
[1] '0.18.0'

Error in qs::qread("compound_db.qs") : QS format not detected

I also tested qs 0.27.3, with the same QS format not detected error.

The file appears to be binary, not an HTML/text error page:

readBin("compound_db.qs", "raw", n = 20)

Output:

[1] 0b 0e 0a c1 01 01 02 01 00 00 00 00 00 00 00 00 55 b7 fb 5a

Environment

R version 4.5.1
Platform: x86_64-pc-linux-gnu
MetaboAnalystR 4.2.0
qs 0.25.5
stringfish 0.18.0

Question

Is the hosted compound_db.qs expected to be readable with qs::qread()? It seems .get.my.lib() retries the download after qread() fails, but then returns my.lib even though it was never created, causing the final object 'my.lib' not found error.

[hosseinvk]

Dear @Zhiqiang-PANG,

Thank you for your recent work on the application and for replacing qs2 with qs. The original installation issue is now resolved.

However, further downstream I am now facing a new error when running CalculateHyperScore().

The following steps run successfully, but the final command fails:

packageVersion("MetaboAnalystR")
# [1] ‘4.3.0’

mSet <- InitDataObjects("conc", "msetora", FALSE, 150) # OK

cmpd.vec<-c("Auricularine", "Xylose", "Hydroquinone" ...)

mSet <- Setup.MapData(mSet, cmpd.vec) # OK
# Warning message:
# In write.csv(qvec, file = "datalist.csv", row.names = FALSE, col.names = FALSE, :
#   attempt to set 'col.names' ignored

mSet <- CrossReferencing(mSet, "name") # OK
# [1] "compound_db.qs"
# [1] "Loaded files from MetaboAnalyst web-server."
# [1] "syn_nms.qs"
# [1] "Loaded files from MetaboAnalyst web-server."
# [1] "2"
# [2] "There are >15 compounds without matches. You can either proceed or if necessary, update these compound IDs and upload again."

mSet <- CreateMappingResultTable(mSet) # OK
# [1] "compound_db.qs"
# [1] "Loaded files from MetaboAnalyst web-server."

mSet <- Setup.HMDBReferenceMetabolome(
  mSet,
  file.path(model_dir, "tensor", "background_FE", "bg_metabolites_CB.txt")
) # OK
# Read 1190 items
# [1] "master_compound_db.qs"
# [1] "Loaded files from MetaboAnalyst web-server."

mSet <- SetMetabolomeFilter(mSet, TRUE) # OK
mSet <- SetCurrentMsetLib(mSet, "smpdb_pathway", 2) # OK

mSet <- CalculateHyperScore(mSet) # ERROR
# [1] "compound_db.qs"
# [1] "Loaded files from MetaboAnalyst web-server."
# Error in mSetObj$analSet : $ operator is invalid for atomic vectors
# In addition: Warning message:
# In value[[3L]](cond) :
#   ExportResultMatArrow failed: cannot connect to localhost:6311

I am sharing this update in case it helps with tracking the current state of the issue.

Many thanks.