Weaviate · Schema

InvertedIndexConfig

Configure the inverted index built into Weaviate. See [Reference: Inverted index](https://docs.weaviate.io/weaviate/config-refs/indexing/inverted-index#inverted-index-parameters) for details.

Vector DatabaseAIMachine LearningSemantic SearchOpen SourceGraphQLKubernetes

Properties

Name Type Description
cleanupIntervalSeconds number Asynchronous index clean up happens every n seconds (default: 60).
bm25 object
stopwords object
indexTimestamps boolean Index each object by its internal timestamps (default: `false`).
indexNullState boolean Index each object with the null state (default: `false`).
indexPropertyLength boolean Index length of properties (default: `false`).
usingBlockMaxWAND boolean Using BlockMax WAND for query execution (default: `false`, will be `true` for new collections created after 1.30).
tokenizerUserDict array User-defined dictionary for tokenization.
stopwordPresets object User-defined named stopword lists. Each key is a preset name that can be referenced by a property's textAnalyzer.stopwordPreset field. The value is an array of stopword strings. Preset names must not
View JSON Schema on GitHub

JSON Schema

weaviate-inverted-index-config-schema.json Raw ↑
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://api-evangelist.github.io/weaviate/json-schema/weaviate-inverted-index-config-schema.json",
  "title": "InvertedIndexConfig",
  "description": "Configure the inverted index built into Weaviate. See [Reference: Inverted index](https://docs.weaviate.io/weaviate/config-refs/indexing/inverted-index#inverted-index-parameters) for details.",
  "type": "object",
  "properties": {
    "cleanupIntervalSeconds": {
      "type": "number",
      "format": "int",
      "description": "Asynchronous index clean up happens every n seconds (default: 60)."
    },
    "bm25": {
      "$ref": "#/components/schemas/BM25Config"
    },
    "stopwords": {
      "$ref": "#/components/schemas/StopwordConfig"
    },
    "indexTimestamps": {
      "type": "boolean",
      "description": "Index each object by its internal timestamps (default: `false`)."
    },
    "indexNullState": {
      "type": "boolean",
      "description": "Index each object with the null state (default: `false`)."
    },
    "indexPropertyLength": {
      "type": "boolean",
      "description": "Index length of properties (default: `false`)."
    },
    "usingBlockMaxWAND": {
      "type": "boolean",
      "description": "Using BlockMax WAND for query execution (default: `false`, will be `true` for new collections created after 1.30)."
    },
    "tokenizerUserDict": {
      "type": "array",
      "description": "User-defined dictionary for tokenization.",
      "items": {
        "$ref": "#/components/schemas/TokenizerUserDictConfig"
      }
    },
    "stopwordPresets": {
      "type": "object",
      "description": "User-defined named stopword lists. Each key is a preset name that can be referenced by a property's textAnalyzer.stopwordPreset field. The value is an array of stopword strings. Preset names must not be empty or whitespace-only; each list must contain at least one word; individual words must not be empty or whitespace-only.",
      "additionalProperties": {
        "type": "array",
        "items": {
          "type": "string"
        }
      }
    }
  }
}