{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://raw.githubusercontent.com/FAIR-bioHeaders/FHR-Specification/main/fhr.json",
  "title": "FHR",
  "description": "FAIR Header Reference genome Schema for Genome Assemblies",
  "type": "object",
  "properties": {
    "schema":{
      "type": "string",
      "description": "centralized schema file"
    },
    "schemaVersion":{
      "type": "number",
      "description": "Version of FHR"
    },
    "genome":{
      "type": "string",
      "description": "Name of the Genome"
    },
    "genomeSynonym": {
      "type": "array",
      "description": "Other common names of the genome",
      "items": {
        "type":"string"
      }
    },
    "taxon":{
      "type": "object",
      "description": "Species name and URL of the species information at identifiers.org",
      "properties": {
        "name": {
          "type": "string"
        },
        "uri": {
          "type": "string",
          "format": "uri",
          "pattern": "https://identifiers.org/taxonomy:[0-9]+"
        }
      }
    },
    "version":{
      "type": "string",
      "description": "Version number of Genome eg. 1.2.0"
    },
    "metadataAuthor":{
      "type": "array",
      "items": {
        "type": "object",
        "description": "Author of the FHR Instance (Person or Org)",
        "properties": {
          "name": {
            "type": "string"
          },
          "uri": { "$ref": "#/definitions/orcidUri" }
        }
      }
    },
    "assemblyAuthor":{
      "type": "array",
        "items": {
        "type": "object",
        "description": "Assembler of the Genome (Person or Org)",
        "properties": {
          "name": {
            "type": "string"
          },
          "uri": { "$ref": "#/definitions/orcidUri" }
        }
      }
    },
    "dateCreated":{
      "type": "string",
      "format": "date",
      "description": "Date the genome assembly was created"
    },
    "voucherSpecimen":{
      "type": "string",
      "description": "Description of the physical sample"
    },
    "accessionID":{
      "type": "object",
      "description": "accessionID genome assembly was created",
      "properties": {
        "name": {
          "type": "string"
        },
        "url": {
          "type": "string",
          "format": "uri"
        }
      }
    },
    "instrument": {
      "type": "array",
      "description": "Physical tools and instruments used in the creation of the genome assembly",
      "items": {
        "type": "string"
      }
    },
    "scholarlyArticle": {
      "type": "string",
      "pattern": "^10.",
      "description": "Scholarly article genome was published e.g. 10.5281/zenodo.6762550 "
    },
    "documentation": {
      "type": "string",
      "description": "Documentation about the genome"
    },
    "identifier": {
      "type": "array",
      "description": "Identifies of the genome",
      "items": {
        "type": "string",
        "pattern": "[a-z0-9]*:.*"
        }
    },
    "relatedLink": {
      "type": "array",
      "description": "Related URLS to the genome",
      "items": {
        "type": "string",
        "format": "uri"
      }
    },
    "funding": {
      "type": "string",
      "description": "Grant Line Item"
    },
    "reuseConditions": {
      "type": "string",
      "description": "license and reuse conditions for the use of the Genome"
    },
    "masking": {
      "type": "string",
      "pattern": "(not-masked|hard-masked|soft-masked|repeat-masked|unknown)",
    "description": "masking applied to the genome assembly"
    },
    "vitalStats": {
      "type": "object",
      "description": "general statistics about the genome assembly",
      "properties": {
        "L50": { 
            "type": "integer",
            "description": "L50 of the genome assembly"
        },
        "N50": { 
            "type": "integer",
             "description": "N50 of the genome assembly"
        },
        "L90": { 
            "type": "integer" ,
            "description": "L90 of the genome assembly"
        },
        "totalBasePairs": { 
            "type": "integer" ,
            "description": "total base pairs of the genome assembly"
        },
        "numberContigs": { 
            "type": "integer" ,
            "description": "number of contigs of the genome assembly"
        },
        "numberScaffolds": { 
            "type": "integer" ,
            "description": "number of scaffolds of the genome assembly"
        },
        "readTechnology": { 
            "type": "string",
            "description": "read technology of the genome assembly (short, long, hifi, etc...)"
        }
      }
    },
    "checksum": { "$ref": "#/definitions/sha2" }
  },
  "required": [
    "schema",
    "schemaVersion",
    "genome",
    "taxon",
    "version",
    "metadataAuthor",
    "assemblyAuthor",
    "dateCreated",
    "masking",
    "checksum"
  ],
  "definitions": {
    "orcidUri": { "format": "uri", 
                  "pattern": "^https://orcid.org/[0-9][0-9][0-9][0-9]-[0-9][0-9][0-9][0-9]-[0-9][0-9][0-9][0-9]-[0-9][0-9][0-9][0-9X]" },
    "sha2": {
      "type": "string",
      "minLength": 44,
      "maxLength": 44,
      "pattern": "^[A-Za-z0-9/+=]+$",
      "description": "sha2-512/256 checksum value for hashing"
    }
  },
  "additionalProperties": false
}