manifest.schema.json

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://github.com/SysBioChalmers/raven-python/manifest.schema.json",
  "title": "RAVEN data/binary manifest",
  "description": "Language-agnostic registry of downloadable raven-python / RAVEN data artefacts and external binary bundles. Consumed by the Python resolvers (raven_python.data / raven_python.binaries) and by MATLAB RAVEN. Every file carries a SHA256 so consumers verify integrity after download.",
  "type": "object",
  "required": ["manifest_version"],
  "additionalProperties": false,
  "properties": {
    "manifest_version": {
      "type": "integer",
      "const": 1,
      "description": "Format version of this manifest document."
    },
    "generated": {
      "type": "string",
      "description": "ISO-8601 date the manifest was generated (informational)."
    },
    "data": {
      "type": "object",
      "description": "Data-artefact datasets, keyed by dataset id (e.g. 'kegg'). Maps onto raven_python.data._DATA_REGISTRY.",
      "additionalProperties": { "$ref": "#/$defs/dataset" }
    },
    "binaries": {
      "type": "object",
      "description": "External command-line tool bundles, keyed by bundle id (e.g. 'blast', 'diamond', 'hmmer'). Maps onto raven_python.binaries._REGISTRY.",
      "additionalProperties": { "$ref": "#/$defs/bundle" }
    }
  },
  "$defs": {
    "file": {
      "type": "object",
      "required": ["url", "sha256"],
      "additionalProperties": false,
      "properties": {
        "url": { "type": "string", "format": "uri", "description": "Direct download URL (GitHub release asset, Zenodo file, etc.)." },
        "sha256": { "type": "string", "pattern": "^[0-9a-f]{64}$", "description": "Lowercase hex SHA256 of the file." },
        "bytes": { "type": "integer", "minimum": 0, "description": "File size in bytes (informational; for progress bars / sanity checks)." }
      }
    },
    "dataset": {
      "type": "object",
      "required": ["version", "files"],
      "additionalProperties": false,
      "properties": {
        "version": { "type": "string", "description": "Dataset version tag, e.g. 'kegg116'. Used in the cache path." },
        "description": { "type": "string" },
        "license": { "type": "string", "description": "SPDX id or free text. NOTE: KEGG-derived artefacts are subject to KEGG's terms — confirm redistribution rights before publishing." },
        "doi": { "type": "string", "description": "Zenodo (or other) DOI for this dataset version, if archived." },
        "source": { "type": "string", "format": "uri", "description": "Human-facing page for the release/record (GitHub release or Zenodo landing page)." },
        "files": {
          "type": "object",
          "minProperties": 1,
          "description": "Artefact files keyed by filename.",
          "additionalProperties": { "$ref": "#/$defs/file" }
        }
      }
    },
    "bundle": {
      "type": "object",
      "required": ["version", "provides", "platforms"],
      "additionalProperties": false,
      "properties": {
        "version": { "type": "string", "description": "Upstream tool version, e.g. '2.16.0'." },
        "provides": {
          "type": "array",
          "items": { "type": "string" },
          "minItems": 1,
          "description": "Executable names this bundle provides, e.g. ['blastp', 'makeblastdb']."
        },
        "description": { "type": "string" },
        "license": { "type": "string", "description": "Upstream tool license (e.g. DIAMOND is GPL-3.0-only — ship its license text alongside the ZIP)." },
        "platforms": {
          "type": "object",
          "minProperties": 1,
          "description": "One entry per platform, keyed '<os>-<arch>' (e.g. 'linux-x86_64', 'macos-arm64', 'windows-x86_64'). Matches raven_python.binaries._platform_key().",
          "additionalProperties": { "$ref": "#/$defs/file" }
        }
      }
    }
  }
}