Conversation
Codecov Report
@@ Coverage Diff @@
## dev #147 +/- ##
==========================================
+ Coverage 38.97% 39.05% +0.07%
==========================================
Files 59 59
Lines 5647 5654 +7
==========================================
+ Hits 2201 2208 +7
Misses 3446 3446
Continue to review full report at Codecov.
|
ODiogoSilva
left a comment
ODiogoSilva
left a comment
There was a problem hiding this comment.
We finally have quast! 😃 Thanks @sjackman for the great addition. I've just left a couple of comments, the first being the more relevant. Other than that, it's all good.
| "default": "null", | ||
| "description": "Compare the assembly to this reference genome" | ||
| }, | ||
| "genomeSize": { |
There was a problem hiding this comment.
Is there a reason for the genomeSize to be in "BP"? This parameter already occurs in several other components in "MB" units, so for consistency it would be great for this one to be to. Moreover, when pipelines are built with the --merge-params option, parameters with the same name for different components are merged into a single one and this could cause problems. If you really need the units to be in BP, I suggest changing the name of the parameter (to genomeSizeBp or any other you can think of) to avoid these problems.
There was a problem hiding this comment.
I prefer to use a base SI unit without an implied prefix for the units of parameters. The reason for this is that when I see --genomeSize=3e9 it's clear what's going on. When I see --genomeSize=3000 I don't know if the units are in bp, kbp, Mbp, or Gbp without reading the documentation. Additionally, when specifying the size of small genomes, such as the human mitochondrial genome, I found --genomeSize=0.016569 to be rather awkward and confusing compared to --genomeSize=16569.
There was a problem hiding this comment.
For consistency though and to prevent trouble with --merge-params I'll change the parameter name to genomeSizeBp.
There was a problem hiding this comment.
Thanks, though those are valid point you make. For this PR it's easier and more contained if we change the parameter name, but in the near future it'll be a good idea to standardize the genomeSize parameter like you proposed.
| {% endwith %} | ||
|
|
||
| script: | ||
| "/usr/bin/time -v quast -o . -r $reference -s $assembly -l $sample_id -t $task.cpus >> .command.log 2>&1" |
There was a problem hiding this comment.
There's no trouble in using time to check the run time of the software, but that information (and more) will already be available on the pipeline_stats.txt file generated during the execution of the pipeline 😄
There was a problem hiding this comment.
That's excellent!
This is odd though. /usr/bin/time -v reports
Maximum resident set size (kbytes): 408240
whereas pipeline_stats.txt reports
rss = 58.9 MB
vmem = 476.4 MB
So I'd like to leave /usr/bin/time -v in until I sort out this discrepancy.
There was a problem hiding this comment.
No problem. Though that's interesting. It could be that the rss reported in by nextflow takes into account any possible overhead of a docker/singularity execution, while the time one is purely from quast. Or they simply measure rss differently. Either way I would be interesting to find out why!
There was a problem hiding this comment.
Ah yeah, the Docker thing makes sense for the difference between 408M and 476M. I'm more confused though by /usr/bin/time reporting maxrss=408M whereas pipeline_stats.txt reports rss=58.9M. That's a very big difference!
There was a problem hiding this comment.
Oh, I misread. I though it was 40Mb! That is very weird indeed. It almost seems that time's rss is reporting the virtual memory measure. Does time also provide something like vmem?
There was a problem hiding this comment.
That may be. I doesn't oddly, which is why I usually use zsh -c 'time foobar' rather than /usr/bin/time for this purpose.
There was a problem hiding this comment.
I don't think zsh is in that quast Docker image though.
ODiogoSilva
left a comment
ODiogoSilva
left a comment
There was a problem hiding this comment.
All good to me and welcome to the team! 😄 (also thumbs on the commit squash)
|
Thanks! Flowcraft is pretty cool! 😁 |
4 participants
No description provided.