{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://bendyline.com/gezk/0.5/schemas/embedding-profile.schema.json",
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "minLength": 1
    },
    "model": {
      "type": "object",
      "properties": {
        "repo": {
          "type": "string",
          "minLength": 1
        },
        "revision": {
          "type": "string",
          "minLength": 1
        },
        "onnxFile": {
          "type": "string",
          "minLength": 1
        },
        "onnxDigest": {
          "type": "string",
          "pattern": "^sha256:[0-9a-f]{64}$"
        }
      },
      "required": [
        "repo",
        "revision"
      ]
    },
    "tokenizer": {
      "type": "object",
      "properties": {
        "kind": {
          "type": "string",
          "minLength": 1
        },
        "file": {
          "type": "string",
          "minLength": 1
        },
        "digest": {
          "type": "string",
          "pattern": "^sha256:[0-9a-f]{64}$"
        }
      },
      "required": [
        "kind"
      ]
    },
    "pooling": {
      "type": "string",
      "enum": [
        "mean",
        "cls",
        "last"
      ]
    },
    "normalized": {
      "type": "boolean"
    },
    "dimensions": {
      "type": "integer",
      "exclusiveMinimum": 0,
      "maximum": 9007199254740991
    },
    "maxTokens": {
      "type": "integer",
      "exclusiveMinimum": 0,
      "maximum": 9007199254740991
    },
    "queryInstruction": {
      "type": "string"
    },
    "passageInstruction": {
      "type": "string"
    },
    "vectorEncoding": {
      "type": "string",
      "enum": [
        "bit+int8"
      ]
    },
    "distance": {
      "type": "object",
      "properties": {
        "stage1": {
          "type": "string",
          "const": "hamming"
        },
        "stage2": {
          "type": "string",
          "const": "cosine"
        }
      },
      "required": [
        "stage1",
        "stage2"
      ]
    },
    "quantization": {
      "type": "object",
      "properties": {
        "int8": {
          "type": "object",
          "properties": {
            "method": {
              "type": "string",
              "const": "symmetric-linear"
            },
            "scale": {
              "type": "number",
              "const": 127
            }
          },
          "required": [
            "method",
            "scale"
          ]
        },
        "binary": {
          "type": "object",
          "properties": {
            "method": {
              "type": "string",
              "const": "sign"
            },
            "threshold": {
              "type": "number",
              "const": 0
            },
            "packing": {
              "type": "string",
              "const": "lsb-first"
            }
          },
          "required": [
            "method",
            "threshold",
            "packing"
          ]
        }
      },
      "required": [
        "int8",
        "binary"
      ]
    }
  },
  "required": [
    "id",
    "model",
    "tokenizer",
    "pooling",
    "normalized",
    "dimensions",
    "maxTokens",
    "queryInstruction",
    "passageInstruction",
    "vectorEncoding",
    "distance",
    "quantization"
  ]
}