You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi Luis,
I currently get an error I cannot trace back and our IT doesn't know what's the issue either. I am running a job on one of our huge nodes 2TB, so memory should not be an issue and get the following error (just removed some personal information (path)):
Exiting after fatal error:
ESC[31mAn unhandled error occurred (this should not happen)!
If you can reproduce this issue, please run your script
with the --trace flag and report a bug (including the script and the trace) at
https://github.com/ngless-toolkit/ngless/issues
The error message was: `fd:12: hPutBuf: resource vanished (Broken pipe)`)```
here are the last lines from the trace log:
[main] CMD:/bactopia-20201013/anaconda3/envs/ngless/bin/bwa mem -t 1 -K 100000000 -p -a /NGLESSmodules/Modules/gmgc.ngm/1.0/cached/gmgc:no-rare.fna.splits_250000m.0-bwa-0.7.17.fna -
[main] Real time: 79285.296 sec; CPU: 79095.875 sec
[Wed 22-06-2022 12:52:27] Line 16: Success
[Wed 22-06-2022 12:52:27] Line 16: Mapped readset stats (/NGLESSmodules/Modules/gmgc.ngm/1.0/cached/gmgc:no-rare.fna.splits_250000m.0.fna):
[Wed 22-06-2022 12:52:27] Line 16: Total reads: 21950778
[Wed 22-06-2022 12:52:27] Line 16: Total reads aligned: 13303553 [60.61%]
[Wed 22-06-2022 12:52:27] Line 16: Total reads Unique map: 7767800 [35.39%]
[Wed 22-06-2022 12:52:27] Line 16: Total reads Non-Unique map: 5535753 [25.22%]
[Wed 22-06-2022 12:52:27] Line 17: Running garbage collection.
[Wed 22-06-2022 12:52:27] Line 17: Interpreting [interpretIO]: gmgc_mapped_post = select(Lookup 'gmgc_mapped' as NGLMappedReadSet)using {Block {blockVariable = Variable "mr", blockBody = Sequence [mr = (Lookup 'mr' as NGLMappedRead).MethodName {unwrapMethodName = "filter"}( Nothing; min_match_size=45; min_identity_pc=95; action={drop} ),if [UOpNot((Lookup 'mr' as NGLMappedRead).MethodName {unwrapMethodName = "flag"}( Just {mapped} ))] then {Sequence [discard]} else {Sequence []}]}}
[Wed 22-06-2022 12:52:27] Line 17: Interpreting [assignment]: select(Lookup 'gmgc_mapped' as NGLMappedReadSet)using {Block {blockVariable = Variable "mr", blockBody = Sequence [mr = (Lookup 'mr' as NGLMappedRead).MethodName {unwrapMethodName = "filter"}( Nothing; min_match_size=45; min_identity_pc=95; action={drop} ),if [UOpNot((Lookup 'mr' as NGLMappedRead).MethodName {unwrapMethodName = "flag"}( Just {mapped} ))] then {Sequence [discard]} else {Sequence []}]}}
[Wed 22-06-2022 12:52:27] Line 17: Executing blocked select on file /temp/mapped_gmgc:no-rare.sam6057-0.zstd
[Wed 22-06-2022 12:52:27] Line 17: Created & opened temporary file //temp/block_selected_mapped_gmgc:no-rare.sam6057-1.zstd
[Wed 22-06-2022 13:00:19] Line 21: Running garbage collection.
[Wed 22-06-2022 13:00:19] Line 21: Interpreting [interpretIO]: temp$4 = mapstats(Lookup 'gmgc_mapped_post' as NGLMappedReadSet)
[Wed 22-06-2022 13:00:19] Line 21: Interpreting [assignment]: mapstats(Lookup 'gmgc_mapped_post' as NGLMappedReadSet)
[Wed 22-06-2022 13:00:19] Line 21: Computing mapstats on File /temp/block_selected_mapped_gmgc:no-rare.sam6057-1.zstd
[Wed 22-06-2022 13:03:03] Line 21: Created & opened temporary file /temp/sam_stats_block_selected_mapped_gmgc:no-rare6057-2.stats
[Wed 22-06-2022 13:03:03] Line 21: Running garbage collection.
[Wed 22-06-2022 13:03:03] Line 21: Interpreting [interpretIO]: __check_ofile(BinaryOp("gmgc" -BOpPathAppend- BinaryOp(Lookup 'RESULTS' as NGLString -BOpAdd- "gmgc_norare.stats.txt")); original_lno=21)
[Wed 22-06-2022 13:03:03] Line 21: Interpreting [executing module function: '__check_ofile']: NGOString "gmgc/D16gmgc_norare.stats.txt"
[Wed 22-06-2022 13:03:03] Line 21: Running garbage collection.
[Wed 22-06-2022 13:03:03] Line 21: Interpreting [interpretIO]: write(Lookup 'temp$4' as NGLCounts; __can_move=True; __hash="7b8566ff57fbce4d04ada723875d32b8"; ofile=BinaryOp("gmgc" -BOpPathAppend- BinaryOp(Lookup 'RESULTS' as NGLString -BOpAdd- "gmgc_norare.stats.txt")))
[Wed 22-06-2022 13:03:03] Line 21: Interpreting [write]: NGOCounts File /temp/sam_stats_block_selected_mapped_gmgc:no-rare6057-2.stats
[Wed 22-06-2022 13:03:03] Line 21: Writing counts to: gmgc/D16gmgc_norare.stats.txt
[Wed 22-06-2022 13:03:03] Line 22: Running garbage collection.
[Wed 22-06-2022 13:03:03] Line 22: Interpreting [interpretIO]: __check_ofile(BinaryOp("gmgc" -BOpPathAppend- BinaryOp(Lookup 'RESULTS' as NGLString -BOpAdd- ".gmgc_norare.bam")); original_lno=22)
[Wed 22-06-2022 13:03:03] Line 22: Interpreting [executing module function: '__check_ofile']: NGOString "gmgc/D16.gmgc_norare.bam"
[Wed 22-06-2022 13:03:03] Line 22: Running garbage collection.
[Wed 22-06-2022 13:03:03] Line 22: Interpreting [interpretIO]: write(Lookup 'gmgc_mapped_post' as NGLMappedReadSet; __can_move=True; __hash="dcb070e78c288900aed230a2e77b322c"; ofile=BinaryOp("gmgc" -BOpPathAppend- BinaryOp(Lookup 'RESULTS' as NGLString -BOpAdd- ".gmgc_norare.bam")))
[Wed 22-06-2022 13:03:03] Line 22: Interpreting [write]: NGOMappedReadSet {nglgroupName = "preprocessed/D16filtered_HG_HS_qc.pair.1.fq.gz", nglSamFile = File /temp/block_selected_mapped_gmgc:no-rare.sam6057-1.zstd, nglReference = Just "gmgc:no-rare"}
[Wed 22-06-2022 13:03:03] Line 22: Created & opened temporary file /temp/converted_block_selected_mapped_gmgc:no-rare6057-3.bam
[Wed 22-06-2022 13:03:03] Line 22: SAM->BAM Conversion start ('/block_selected_mapped_gmgc:no-rare.sam6057-1.zstd' -> '/temp/converted_block_selected_mapped_gmgc:no-rare6057-3.bam')
ngless "1.4"
import "gmgc" version "1.0"
import "parallel" version "1.0"
import "samtools" version "1.0"
samples = readlines('samplelist.txt')
current = lock1(samples)
input = paired ("preprocessed"</>current + "filtered_HG_HS_qc.pair.1.fq.gz","preprocessed"</>current + "filtered_HG_HS_qc.pair.2.fq.gz")
#different result outputdirs not necessary anymore, since all results will be "collected" in one output with the collect function
RESULTS = current
##GMGC
##"How counts are adjusted in the presence of multiple annotations is defined by the multiple argument. Generally, for obtaining gene abundances, distribution of multiple mappers is the best (using multiple={dist1}), while for functional annotations, you want to count them all (using multiple={all1}). This implies that the functional annotations will sum to a higher value than the number of reads. This may seem strange at first, but it is the intended behaviour."
gmgc_mapped = map (input, reference='gmgc:no-rare',mode_all=True,block_size_megabases=250000)
gmgc_mapped_post = select(gmgc_mapped) using |mr|:
mr = mr.filter(min_match_size=45, min_identity_pc=95, action={drop})
if not mr.flag({mapped}):
discard
write(mapstats(gmgc_mapped_post),ofile="gmgc"</>RESULTS+'gmgc_norare.stats.txt')
write(gmgc_mapped_post,ofile="gmgc"</>RESULTS+'.gmgc_norare.bam')
Thanks in advance!
Ulrike
The text was updated successfully, but these errors were encountered:
Hi Luis,
I currently get an error I cannot trace back and our IT doesn't know what's the issue either. I am running a job on one of our huge nodes 2TB, so memory should not be an issue and get the following error (just removed some personal information (path)):
Thanks in advance!
Ulrike
The text was updated successfully, but these errors were encountered: