Open mmokrejs opened 6 years ago
Hi @mmokrejs,
Thanks again for all the feedback, we'll work on these. I'll mention that (2) is not quite simple as it seems. Specifically, if you write mappings (--writeMappings, -z
), the implicit file is STDOUT. I believe this has been discussed with @tseemann in the past. This is intentional so that one can immediately pipe that output to e.g. samtools to convert the SAM format to a BAM format. I am open to cleaner solutions that (ideally) don't involve having to write the BAM files directly.
Hi, thank you for a prompt answer. Well, now I understand more why you write on STDERR
... Hum, I do not know what to do then.
I think the longer-term solution is for us to directly support BAM writing. This would obviate the need of passing the output through samtools
. I will mark this issue as an enhancement to remind us to look into what will be involved in doing that (while we work on addressing the more obvious issues you raise).
Hi, would you mind cleaning up the STDOUT and STDERR output from
salmon quant
andsalmon index
? CommonlySTDOUT
should be used for normal output messages andSTDERR
for errors or at worst warnings. Salmon writes most of the messages toSTDERR
.Version Info: This is the most recent version of Salmon.
is output toSTDOUT
. I find it useless. If salmon run some network connection to figure out its version it isa)
prone to errors,b)
I would expect a fat warning in the documentation aphone home
feature is builtin,c)
it is likely to fail on more network-restricted installations. Or,d)
, the message is incorrect. I suggest drop the message altogether.salmon quant
writes a lot of normal messages toSTDERR
. Please useSTDOUT
instead. If a program exits with a non-zero exit code it is common to read itsSTDERR
output to learn what was the cause for the error. It is awkward to realize there is lots of unrelated text. Please follow common rules on Unix.The docs at http://salmon.readthedocs.io/en/latest/salmon.html did not mention the
fmd
index is just a plain index frombwa
. Why don't you instruct users to usebwa index
instead? It would be clearer (if that is the type of index you employ).salmon index
behaviorIt would be more interesting to see what filename is used for the index. Showing the "basename" of the index file is not very helpful. I would prefer to see the filename(s) on the beginning and end lines:
I wonder what are these duplicates in a human cdna predicted, as available from
ftp://ftp.ensembl.org/pub/release-92/fasta/homo_sapiens/cdna
:Homo_sapiens.GRCh38.cdna.all/bwaidx
was created.Doh, something unexpected from the logs, isn't it?
Hope this helps.