schema.json

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://earthindex.ai/spec/emb.schema.json",
  "title": "Parquet Embedding Metadata",
  "description": "Parquet Embedding Metadata",
  "type": "object",
  "allOf": [
    {
      "$ref": "https://schemas.stacspec.org/v1.0.0/item-spec/json-schema/provider.json"
    },
    {
      "$ref": "https://schemas.stacspec.org/v1.0.0/item-spec/json-schema/datetime.json"
    },
    {
      "$ref": "https://schemas.stacspec.org/v1.0.0/item-spec/json-schema/licensing.json"
    }
  ],
  "properties": {
    "version": {
      "description": "Spec version",
      "type": "string",
      "const": "0.0.1"
    },
    "model": {
      "description": "Information about the model that created these embeddings",
      "type": "object",
      "properties": {
        "id": {
          "description": "Id of the model",
          "type": "string"
        },
        "family": {
          "description": "",
          "type": "string"
        },
        "name": {
          "description": "",
          "type": "string"
        },
        "description": {
          "description": "",
          "type": "string"
        },
        "config": {
          "description": "",
          "type": "string"
        },
        "source": {
          "description": "",
          "type": "string"
        }
      },
      "required": [
        "id",
        "source"
      ]
    },
    "embedding": {
      "description": "Embedding metadata",
      "type": "object",
      "properites": [
        {
          "dim": {
            "description": "the length of the embeddings vector",
            "type": "number"
          },
          "quantization": {
            "description": "a description of quantization applied, if any",
            "type": "string"
          }
        }
      ]
    },
    "source_datasets": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "id": {
            "description": "",
            "type": "string"
          },
          "source": {
            "description": "",
            "type": "string"
          },
          "name": {
            "description": "",
            "type": "string"
          },
          "description": {
            "description": "",
            "type": "string"
          }
        },
        "required": [
          "id",
          "source"
        ]
      }
    }
  },
  "required": [
    "version",
    "model",
    "providers",
    "datetime",
    "start_datetime",
    "end_datetime",
    "datasets",
    "embedding"
  ]
}