nf-core / sarek

Analysis pipeline to detect germline or somatic variants (pre-processing, variant calling and annotation) from WGS / targeted sequencing
https://nf-co.re/sarek
MIT License
393 stars 402 forks source link

[BUG] * --vep_cache_version: expected type: String, found: Integer (99) #518

Closed leipzig closed 2 years ago

leipzig commented 2 years ago

Check Documentation

I have checked the following places for your error:

Description of the bug

* --vep_cache_version: expected type: String, found: Integer (99)

Steps to reproduce

Steps to reproduce the behaviour:

./nextflow run sarek-2.7.1  \
--input tiny-manta-https.gcs.tsv     \
--step 'mapping'     \
--outdir './results'             \
--nucleotides_per_second 1000.0                                              \
--aligner 'bwa-mem'  \
--markdup_java_options '-Xms4000m -Xmx7g'                        \
--cf_coeff 0.05          \
--cf_ploidy 2.0                                                                              \
--genome 'smallGRCh37'                                                                   \
--snpeff_db 'WBcel235.86'    \
--species 'caenorhabditis_elegans'   \
--igenomes_base 's3://ngi-igenomes/igenomes'     \
--genomes_base 'raw.githubusercontent.com/nf-core/test-datasets/sarek/reference'     \
--igenomes_ignore        \
--publish_dir_mode 'copy'        \
--validate_params            \
--max_multiqc_email_size '25.MB'             \
--tracedir './results/pipeline_info'             \
--cpus 8     \
--single_cpu_mem '7 GB'  \
--max_cpus 16    \
--max_memory '128.GB'    \
--max_time '240.h'   \
--custom_config_version 'master'     \
--vep_cache_version '99'         \
--custom_config_base 'https://raw.githubusercontent.com/nf-core/configs/master'  \
-w gs://somebucket

Expected behaviour

not treat vep_cache_version as numeric when it is quoted

Log files

System

Nextflow Installation

Container engine

Additional context

there is no type of quote that will solve this to my knowledge

tdanhorn commented 2 years ago

Don't know what OS/shell you are using, but Unix shells strip the quotes, so nextflow (or the pipeline) never gets to see them, i.e it makes no difference (in this case) if you put them there. If you really want quotes to be seen by the pipeline, \'99\' or "'99'" should do the trick. On the other hand, if something should be read as a string and the code autodetects a number and complains, because it thinks it is smarter than you, I'd call that a bug.

maxulysse commented 2 years ago

@leipzig sorry for late reply. As @tdanhorn said, it is indeed a bug which is fixed on the current dev.

maxulysse commented 2 years ago

@leipzig can you try that: --vep_cache_version \"99\"

FriederikeHanssen commented 2 years ago

Hm I think there is still something :fish: going on with the vep_cache_version number:


ERROR: Validation of pipeline parameters failed!
* --vep_cache_version: expected type: Number, found: String (104)

when running:

nextflow run nf-core/sarek -r dev -profile cfc \
--input samplesheet.csv \
--outdir ./results \
--genome GATK.GRCh38 \
--aligner bwa-mem2 \
--bwa false \
--save_reference \
-resume

Dev version: 5d0a0b9755

Edit: tried also --vep_cache_version \"104\", same thing Edit2: --vep_cache_version 104 works with no quotes. So something in the igenomes.config not quite right yet

maxulysse commented 2 years ago

So the issue with igenomes has been fixed. Afraid there's nothing we can do at the moment params from the command line, except using --schema_ignore_params and said params to avoid validation from the schema.

Closing the issue for now.