Quast will fail when given assemblies with no contig greater than 500bp, which causes the pipeline to fail. One poor-quality sample could crash a full run, so it would make the overall pipeline more robust if we can prevent the pipeline from crashing in the presence of a single low-quality sample.

Add a testing workflow using GitHub actions

In anticipation of integrating with tools and platforms like Sequera Platform we'd like to evaluate what would be necessary to adopt the nf-core conventions for our existing pipelines. Since this is a fairly simple pipeline, it's a good candidate for conversion to nf-core.

Pull in the latest version of mlst

The pipeline currently creates one output directory per sample and publishes all outputs there. eg:

When combining this pipeline with others, it may be useful to encapsulate the outputs from this pipeline in a sub-directory that is named with the pipeline name and version.

So by default we would create outputs of this structure:

.
├── sample-01
│   ├── sample-01_alleles.csv
│   └── sample-01_sequence_type.csv
├── sample-02
│   ├── sample-02_alleles.csv
│   └── sample-02_sequence_type.csv
└── sample-03
    ├── sample-03_alleles.csv
    └── sample-03_sequence_type.csv

...but when running with a --versioned_outdir flag , we would produce:

.
├── sample-01
│   └── mlst-nf-v0.1-output
│       ├── sample-01_alleles.csv
│       └── sample-01_sequence_type.csv
├── sample-02
│   └── mlst-nf-v0.1-output
│       ├── sample-01_alleles.csv
│       └── sample-01_sequence_type.csv
└── sample-03
    └── mlst-nf-v0.1-output
        ├── sample-01_alleles.csv
        └── sample-01_sequence_type.csv
 

...then a subsequent analysis could produce similar outputs alongside:

.
├── sample-01
│   ├── mlst-nf-v0.1-output
│   │   └── sample-01_mlst.csv
│   └── routine-assembly-v0.2-output
│       ├── sample-01_bakta.gbk
│       └── sample-01_unicycler.fa
├── sample-02
│   ├── mlst-nf-v0.1-output
│   │   └── sample-02_mlst.csv
│   └── routine-assembly-v0.2-output
│       ├── sample-02_bakta.gbk
│       └── sample-02_unicycler.fa
└── sample-03
    ├── mlst-nf-v0.1-output
    │   └── sample-03_mlst.csv
    └── routine-assembly-v0.2-output
        ├── sample-03_bakta.gbk
        └── sample-03_unicycler.fa

There are cases where we run this pipeline on the outputs of another pipeline (generally BCCDC-PHL/routine-assembly. That pipeline may already perform QC on its outputs, so running essentially the same QC on the inputs of this pipeline would be redundant.

Add a --skip_input_qc flag that causes the QUAST analysis on the input assemblies to be skipped.

We currently only generate a separate output directory for each sample. But it would be convenient to collect the sequence types for all samples into a single .csv file as well. The user should be able to specify a prefix for the collected outputs, using a --collected_outputs_prefix flag, whose default value is collected.

The versioned_outdir param hasn't proven to be useful, and it clutters up our publishDir directives.

Remove the versioned_outdir param.

Command error:
  Traceback (most recent call last):
    File "/home/dfornika/.nextflow/assets/BCCDC-PHL/mlst-nf/bin/parse_alleles.py", line 78, in <module>
      main(args)
    File "/home/dfornika/.nextflow/assets/BCCDC-PHL/mlst-nf/bin/parse_alleles.py", line 29, in main
      num_alleles = len(mlst[sample]['alleles'])
  TypeError: object of type 'NoneType' has no len()

json output from mlst was:

{
   "sample-X.fa" : {
      "scheme" : "-",
      "sequence_type" : "-",
      "alleles" : null,
      "filename" : "sample-X.fa"
   }
}

Our tests are currently failing because the provenance files produced by this pipeline don't match our schema

Update the provenance format to match the schema.

The pipeline version as stated by the manifest here:

...is out of sync with our tagged releases as listed here:

https://github.com/BCCDC-PHL/mlst-nf/releases

We should QC incoming data by running quast on input assemblies.