Update cambridge.md #573
Update cambridge.md #573
Conversation
The instructions include the latest changes () in config file: https://github.com/nf-core/configs/blob/master/conf/cambridge.config
|
@nf-core-bot fix linting |
docs/cambridge.md
Outdated
| tar xvfz jdk-20_linux-x64_bin.tar.gz | ||
| # add these lines to .bashrc | ||
| export JAVA_HOME=/home/ef479/jdk-20.0.1 | ||
| export PATH=/home/ef479/jdk-20.0.1/bin:$PATH |
There was a problem hiding this comment.
This also seems dangerous, what if someone needs to use a different tool that isn't v20 compatible? Another reason why I suggest writing Conde instructions above (and install mxf into a separate env)
There was a problem hiding this comment.
I have updated this.
docs/cambridge.md
Outdated
|
|
||
| # Launch the nf-core pipeline for a test database | ||
| # with the Cambridge profile | ||
| nextflow run nf-core/sarek -profile test,cambridge.config --partition "cclake" --project "NAME-SL3-CPU" --cacheDir "/home/<username>/rds/hpc-work/path/to/cache/dir" --outdir nf-sarek-test |
There was a problem hiding this comment.
CacheDir isn't a parameter in any nf-core pipeline afaik, this could break input validatio - particularly if someone uses the config in a non nf-core pipeline that does stricter testing!!!.
You sound instead tell users to use the NXF_SINGULARITY_CACHEDIR bash env variable
There was a problem hiding this comment.
Thanks for this. I changed the documentation here fe1d2ec.
alternative way of getting nextflow using conda
Conda install is suggested first as a preferred method. Singularity cache is now a bash variable.
This removes the params config that could have broken input validation.
removing cachedir in params
- singularity is now an option to load as module, because it should be loaded on login by default. - the cache directory contains more explanation as to why it can be created
|
@nf-core-bot fix linting (Also I've not forgotten @EmelineFavreau ! Getting to this this afternoon!) |
There was a problem hiding this comment.
I'm still unsure about the non-pipeline Params, but I'm assuming you're testing this with no issue for the time being so I approve this for now. However I would consider coming up with a different method in the future as it may conflict with some pipelines (particularly non-nf-core ones)
| @@ -5,7 +5,6 @@ params { | |||
| config_profile_url = "https://docs.hpc.cam.ac.uk/hpc" | |||
| partition = null | |||
| project = null | |||
There was a problem hiding this comment.
This and partition should also be removed as well theoretically
| ``` | ||
| # Launch the nf-core pipeline for a test database | ||
| # with the Cambridge profile | ||
| nextflow run nf-core/sarek -profile test,cambridge.config --partition "cclake" --project "NAME-SL3-CPU" --outdir nf-sarek-test |
There was a problem hiding this comment.
Is partition and project in sarek?
There was a problem hiding this comment.
It might be ok to keep, but you may end up with validation errors as they aren't pipeline specific parameters. It would make make more sense to have them set as environmental variables and insert them via that way
The instructions include the latest changes in config file: https://github.com/nf-core/configs/blob/master/conf/cambridge.config
name: Updated Config documentation for Cambridge HPC
about: An existing cluster config
This allows the user to understand how to use the options partition, project and cacheDir.