name: ncbigenomedownload
description: A tool to quickly download assemblies from NCBI's Assembly database
keywords:
  - fasta
  - download
  - assembly
tools:
  - ncbigenomedownload:
      description: Download genome files from the NCBI FTP server.
      homepage: https://github.com/kblin/ncbi-genome-download
      documentation: https://github.com/kblin/ncbi-genome-download
      tool_dev_url: https://github.com/kblin/ncbi-genome-download
      doi: ""
      licence: ['Apache Software License']

input:
  - meta:
      type: map
      description: |
        Groovy Map containing sample information
        e.g. [ id:'test', single_end:false ]
  - accessions:
      type: file
      description: List of accessions (one per line) to download
      pattern: "*.txt"

output:
  - meta:
      type: map
      description: |
        Groovy Map containing sample information
        e.g. [ id:'test', single_end:false ]
  - versions:
      type: file
      description: File containing software versions
      pattern: "versions.yml"
  - gbk:
      type: file
      description: GenBank format of the genomic sequence(s) in the assembly
      pattern: "*_genomic.gbff.gz"
  - fna:
      type: file
      description: FASTA format of the genomic sequence(s) in the assembly.
      pattern: "*_genomic.fna.gz"
  - rm:
      type: file
      description: RepeatMasker output for eukaryotes.
      pattern: "*_rm.out.gz"
  - features:
      type: file
      description: Tab-delimited text file reporting locations and attributes for a subset of annotated features
      pattern: "*_feature_table.txt.gz"
  - gff:
      type: file
      description: Annotation of the genomic sequence(s) in GFF3 format
      pattern: "*_genomic.gff.gz"
  - faa:
      type: file
      description: FASTA format of the accessioned protein products annotated on the genome assembly.
      pattern: "*_protein.faa.gz"
  - gpff:
      type: file
      description: GenPept format of the accessioned protein products annotated on the genome assembly.
      pattern: "*_protein.gpff.gz"
  - wgs_gbk:
      type: file
      description: GenBank flat file format of the WGS master for the assembly
      pattern: "*_wgsmaster.gbff.gz"
  - cds:
      type: file
      description: FASTA format of the nucleotide sequences corresponding to all CDS features annotated on the assembly
      pattern: "*_cds_from_genomic.fna.gz"
  - rna:
      type: file
      description: FASTA format of accessioned RNA products annotated on the genome assembly
      pattern: "*_rna.fna.gz"
  - rna_fna:
      type: file
      description: FASTA format of the nucleotide sequences corresponding to all RNA features annotated on the assembly
      pattern: "*_rna_from_genomic.fna.gz"
  - report:
      type: file
      description: Tab-delimited text file reporting the name, role and sequence accession.version for objects in the assembly
      pattern: "*_assembly_report.txt"
  - stats:
      type: file
      description: Tab-delimited text file reporting statistics for the assembly
      pattern: "*_assembly_stats.txt"

authors:
  - "@rpetit3"