AppSearchSchema.EmbeddingPropertyConfig
@RequiresFeature(enforcement = "androidx.appsearch.app.Features#isFeatureSupported", name = Features.SCHEMA_EMBEDDING_PROPERTY_CONFIG)
class AppSearchSchema.EmbeddingPropertyConfig : AppSearchSchema.PropertyConfig
| kotlin.Any | ||
| ↳ | androidx.appsearch.app.AppSearchSchema.PropertyConfig | |
| ↳ | androidx.appsearch.app.AppSearchSchema.EmbeddingPropertyConfig | |
Configuration for a property of type EmbeddingVector in a Document.
Summary
Nested types |
|---|
|
Builder for |
Constants |
|
|---|---|
const Int |
Content in this property will not be indexed. |
const Int |
Embedding vectors in this property will be indexed. |
const Int |
Contents in this property will be quantized to 8 bits. |
const Int |
Contents in this property will not be quantized. |
Public functions |
|
|---|---|
Int |
Returns how the property is indexed. |
Int |
Returns how the embedding contents of this property should be quantized. |
Inherited Constants |
|---|
Inherited functions |
|---|
Constants
INDEXING_TYPE_NONE
const val INDEXING_TYPE_NONE = 0: Int
Content in this property will not be indexed.
INDEXING_TYPE_SIMILARITY
const val INDEXING_TYPE_SIMILARITY = 1: Int
Embedding vectors in this property will be indexed.
The index offers 100% accuracy, but has linear time complexity based on the number of embedding vectors within the index.
QUANTIZATION_TYPE_8_BIT
@ExperimentalAppSearchApi
const val QUANTIZATION_TYPE_8_BIT = 1: Int
Contents in this property will be quantized to 8 bits.
QUANTIZATION_TYPE_NONE
@ExperimentalAppSearchApi
const val QUANTIZATION_TYPE_NONE = 0: Int
Contents in this property will not be quantized.
Public functions
getIndexingType
fun getIndexingType(): Int
Returns how the property is indexed.
getQuantizationType
@ExperimentalAppSearchApi
fun getQuantizationType(): Int
Returns how the embedding contents of this property should be quantized.
If the property isn't indexed, returns QUANTIZATION_TYPE_NONE.