name: nextclade_datasetget
description: Get dataset for SARS-CoV-2 genome clade assignment, mutation calling, and sequence quality checks (C++ implementation)
keywords:
  - nextclade
  - variant
  - consensus
tools:
  - nextclade:
      description: SARS-CoV-2 genome clade assignment, mutation calling, and sequence quality checks
      homepage: https://github.com/nextstrain/nextclade
      documentation: https://github.com/nextstrain/nextclade
      tool_dev_url: https://github.com/nextstrain/nextclade
      doi: ""
      licence: ["MIT"]

input:
  - dataset:
      type: string
      description: Name of dataset to retrieve. A list of available datasets can be obtained using the nextclade dataset list command.
      pattern: ".+"
  - reference:
      type: string
      description: Accession id to download dataset based on a particular reference sequence. A list of available datasets can be obtained using the nextclade dataset list command.
      pattern: ".+"
  - tag:
      type: string
      description: Version tag of the dataset to download. A list of available datasets can be obtained using the nextclade dataset list command.
      pattern: ".+"

output:
  - versions:
      type: file
      description: File containing software versions
      pattern: "versions.yml"
  - prefix:
      type: path
      description: A directory containing the dataset files needed for nextclade run
      pattern: "prefix"

authors:
  - "@antunderwood"
  - "@drpatelh"