Skip to main content Crate chroma_types Copy item path Source pub use operators ::*;chroma_proto error operator operators Well-known operator IDs and names that are pre-populated in the database. operators_generated optional_u128 plan regex AddAttachedFunctionInputRequest AddAttachedFunctionInputResponse AddCollectionRecordsPayload Payload for adding records to a collection. AddCollectionRecordsRequest AddCollectionRecordsResponse AdvanceAttachedFunctionResponse AttachFunctionRequest AttachFunctionResponse AttachedFunction AttachedFunctionApiResponse API response struct for attached function with function_name instead of function_id AttachedFunctionInfo AttachedFunctionUpdateInfo AttachedFunctionUuid AttachedFunctionUuid is a wrapper around Uuid to provide a type for attached function identifiers. BoolInvertedIndexConfig BoolInvertedIndexType BoolValueType Boolean value type index configurations ChecklistResponse ChromaResourceName Chunk Collection CollectionAndSegments CollectionConfiguration CollectionFlushInfo CollectionUuid CompositeExpression CountCollectionsRequest CountRequest CreateCollectionPayload CreateCollectionRequest CreateDatabaseRequest CreateDatabaseResponse CreateTenantRequest CreateTenantResponse DataChunkIteraror DataRecord Database DatabaseName DatabaseUuid DatabaseUuid is a wrapper around Uuid to provide a type for the database id. DeleteCollectionRecordsPayload Payload for deleting records from a collection. DeleteCollectionRecordsRequest DeleteCollectionRecordsResponse DeleteCollectionRequest DeleteCollectionResponse DeleteDatabaseRequest DeleteDatabaseResponse DetachFunctionRequest DetachFunctionResponse Directory In-memory directory covering all posting blocks for a dimension. DirectoryBlock A metadata block summarizing the posting blocks for a single dimension. DocumentExpression EmbeddingFunctionNewConfiguration FloatInvertedIndexConfig FloatInvertedIndexType FloatListValueType Float list value type index configurations (for vectors) FloatValueType Float value type index configurations FlushCompactionAndAttachedFunctionResponse FlushCompactionResponse ForkCollectionRequest ForkLogsResponse FtsIndexConfig FtsIndexType GetAttachedFunctionResponse GetCollectionByCrnRequest GetCollectionByIdRequest GetCollectionRequest GetDatabaseRequest GetRequest GetRequestPayload Records can be retrieved by their IDs or by a metadata filter. At least one of ids or where
must be provided. Use include to specify which fields to return in the response. GetResponse All arrays have the same length, with each index representing a single record.
Only fields specified in the request’s include parameter are populated. GetTenantRequest GetTenantResponse GetVectorsResult HealthCheckResponse HnswConfiguration HnswIndexConfig Configuration for HNSW vector index algorithm parameters IncludeList IncludeParsingError IndexStatusResponse IntInvertedIndexConfig IntInvertedIndexType IntValueType Integer value type index configurations InternalCollectionConfiguration InternalHnswConfiguration InternalSpannConfiguration InternalUpdateCollectionConfiguration JobId JobId is a wrapper around Uuid to provide a unified type for job identifiers.
Jobs can be either collection compaction jobs or task execution jobs. ListCollectionsRequest ListDatabasesRequest LogRecord MetadataDelta MetadataEmbeddingRecord MetadataExpression MultiCloudMultiRegionConfiguration Configuration for multi-cloud, multi-region deployments. OperationRecord PostingBlockHeader Header read from the first 16 bytes of a serialized block. ProviderRegion A cloud provider and geographic region. QuantizedCluster QuantizedClusterOwned QueryRequest QueryRequestPayload QueryResponse RawWhereFields RegionName A strongly-typed region name. ResetResponse Schema Schema representation for collection index configurations SearchRequest SearchRequestPayload SearchResponse Segment SegmentFlushInfo SegmentShard SegmentUuid SegmentUuid is a wrapper around Uuid to provide a type for the segment id. SpannConfiguration SpannIndexConfig Configuration for SPANN vector index algorithm parameters SpannPostingList SparsePostingBlock A compressed block of posting list entries for sparse vector search. SparseVector Represents a sparse vector using parallel arrays for indices and values. SparseVectorIndexConfig SparseVectorIndexType SparseVectorLengthMismatch Length mismatch between indices, values, and tokens in a sparse vector. SparseVectorValueType Sparse vector value type index configurations StringInvertedIndexConfig StringInvertedIndexType StringValueType String value type index configurations Tenant Topology A named replication topology spanning multiple provider regions. TopologyName A strongly-typed topology name. UpdateCollectionConfiguration UpdateCollectionPayload UpdateCollectionRecordsPayload Payload for updating existing records in a collection. UpdateCollectionRecordsRequest UpdateCollectionRecordsResponse UpdateCollectionRequest UpdateCollectionResponse UpdateHnswConfiguration UpdateSpannConfiguration UpdateTenantRequest UpdateTenantResponse UpsertCollectionRecordsPayload Payload for upserting records in a collection. UpsertCollectionRecordsRequest UpsertCollectionRecordsResponse ValidationError Errors that can occur when validating a MultiCloudMultiRegionConfiguration ValueTypes Strongly-typed value type configurations
Contains optional configurations for each supported value type VectorEmbeddingRecord VectorIndexConfig VectorIndexType VectorQueryResult AddCollectionRecordsError AdvanceAttachedFunctionError AttachFunctionError AttachedFunctionConversionError Base64DecodeError BatchGetCollectionSoftDeleteStatusError BatchGetCollectionVersionFilePathsError BooleanOperator ClientResolutionError Cmek Customer-managed encryption key for storage encryption. CollectionConfigurationToInternalConfigurationError CollectionConversionError CollectionFlushInfoConversionError CollectionMetadataUpdate CollectionToProtoError ContainsOperator ConversionError CountCollectionsError CountForksError CreateCollectionError CreateDatabaseError CreateTenantError DeleteCollectionError DeleteCollectionRecordsError DeleteDatabaseError DetachFunctionError DirtyMarker Markers for tracking collection compaction state changes. DistributedSpannParametersFromSegmentError DocumentOperator EmbeddingFunctionConfiguration EmbeddingsPayload ExecutorError FilterValidationError FinishAttachedFunctionError FinishCreateAttachedFunctionError FinishDatabaseDeletionError FlushCompactionResponseConversionError ForkCollectionError FtsAlgorithm Full-text search index algorithm. GetAttachedFunctionError GetCollectionByCrnError GetCollectionByIdError GetCollectionError GetCollectionSizeError GetCollectionWithSegmentsError GetCollectionsError GetDatabaseError GetMinCompletionOffsetError GetSegmentsError GetTenantError HeartbeatError HnswParametersFromSegmentError Include Use this enum to specify which fields should be returned when retrieving records. IndexConfig Union type for all index configurations (used by builder pattern) IndexStatusError KnnIndex ListAttachedFunctionsError ListCollectionVersionsError ListDatabasesError MaterializedLogOperation MetadataComparison MetadataSetValue MetadataValue MetadataValueConversionError MetadataValueType NameError Errors that can occur when creating a RegionNameTopologyName Operation OperationConversionError PrimitiveOperator Quantization Quantization implementation for SPANN vector index. QueryError RecordConversionError ResetError ScalarEncoding ScalarEncodingConversionError SchemaBuilderError SchemaError SchemaMismatchError Error returned by Segment::matches_schema SegmentConversionError SegmentFlushInfoConversionError SegmentScope SegmentScopeConversionError SegmentShardError SegmentType SetOperator SignedRoaringBitmap This enum helps to delay the evaluation of set minus in metadata filtering: Space SparseFilePathKind Which sparse-index blockfile family a file_path map key refers to. SparseIndexAlgorithm Sparse vector index algorithm. SparsePostingBlockError UpdateCollectionConfigurationToInternalConfigurationError UpdateCollectionConfigurationToInternalUpdateConfigurationError UpdateCollectionError UpdateCollectionRecordsError UpdateEmbeddingsPayload UpdateMetadataValue UpdateMetadataValueConversionError UpdateTenantError UpdateVectorIndexConfiguration UpsertCollectionRecordsError VectorConversionError VectorIndexConfiguration Where This Where enum serves as an unified representation for the where and where_document clauses.
Although this is not unified in the API level due to legacy design choices, in the future we will be
unifying them together, and the structure of the unified AST should be identical to the one here.
Currently both where and where_document clauses will be translated into Where, and if both are
present we simply create a conjunction of both clauses as the actual filter. This is consistent with
the semantics we used to have when the where and where_document clauses are treated seperately. WhereConversionError WhereError WhereValidationError WrappedSerdeJsonError BOOL_INVERTED_INDEX_NAME BOOL_METADATA BOOL_VALUE_NAME CHROMA_DOCUMENT_KEY CHROMA_GROUP_CHUNK_SIBLINGS_KEY Collection-metadata flag (a MetadataValue::Bool(true)) that opts a
collection into chunk-sibling grouping during log partitioning: records
whose ids share a base before a trailing -{idx} are kept in one
partition so they materialize in WAL order. Foundation source
collections set this so the attached function observes a trailing
end-of-job marker after all sibling chunks. Read by the worker’s
PartitionOperator; set by the foundation /init endpoint. CHROMA_KEY CHROMA_URI_KEY DIRECTORY_PREFIX Prefix tag prepended to a dimension-id key for directory block parts. DOCUMENT_KEY EMBEDDING_KEY F32_METADATA FLOAT_INVERTED_INDEX_NAME FLOAT_LIST_VALUE_NAME FLOAT_VALUE_NAME FTS_INDEX_NAME FULL_TEXT_PLS FULL_TEXT_TOKEN HNSW_PATH INT_INVERTED_INDEX_NAME INT_VALUE_NAME MAX_BLOCK_ENTRIES Maximum number of posting entries per block. Chosen so that the
decompressed block fits comfortably in L1 cache:
4096 * (4 bytes offset + 4 bytes f32 value) = 32 KB. MAX_HEAD_ID_BF_PATH MAX_OFFSET_ID OFFSET_ID_TO_DATA OFFSET_ID_TO_USER_ID POSTING_LIST_PATH QUANTIZED_SPANN_CLUSTER QUANTIZED_SPANN_EMBEDDING_METADATA QUANTIZED_SPANN_QUANTIZED_CENTROID QUANTIZED_SPANN_RAW_CENTROID QUANTIZED_SPANN_SCALAR_METADATA SPARSE_FILE_PATH_DELIMITER Delimiter separating the sparse-index file_path family (one of the
SPARSE_* constants) from the metadata key it indexes, e.g.
sparse_posting::my_field. The file_path map key is internal to Chroma
(never used as a storage path component), so this delimiter only needs to
be unambiguous when parsing the family back out. SPARSE_MAX SPARSE_OFFSET_VALUE SPARSE_POSTING SPARSE_VECTOR_INDEX_NAME SPARSE_VECTOR_VALUE_NAME STRING_INVERTED_INDEX_NAME STRING_METADATA STRING_VALUE_NAME U32_METADATA USER_ID_BLOOM_FILTER USER_ID_TO_OFFSET_ID VECTOR_INDEX_NAME VERSION_MAP_PATH are_metadatas_close_to_equal are_update_metadatas_close_to_equal Check if two metadata are close to equal. Ignores small differences in float values. convert_f16_to_f32 Convert a slice of little-endian f16 bytes into f32 values.
Dispatches to SIMD on supported architectures, scalar fallback otherwise. convert_f16_to_f32_scalar Scalar f16→f32 conversion via the half crate. decode_base64_embedding decode_base64_embeddings decode_base64_update_embeddings decode_embeddings default_batch_size default_center_drift_threshold default_construction_ef default_construction_ef_spann default_default_knn_index default_initial_lambda default_m default_m_spann default_merge_threshold default_nreplica_count default_num_centers_to_merge_to default_num_samples_kmeans default_num_threads default_reassign_neighbor_count default_resize_factor default_search_ef default_search_ef_spann default_search_nprobe default_search_rng_epsilon default_search_rng_factor default_space default_split_threshold default_sync_threshold default_write_nprobe default_write_rng_epsilon default_write_rng_factor dirty_log_path_from_hostname get_metadata_value_as is_embedding_function_default is_hnsw_config_default Check if HNSW config is default is_space_default Check if space is default (None means default, or if present, should be default space) logical_size_of_metadata maybe_decode_update_embeddings parse_sparse_file_path_key Classify a file_path map key as a sparse-index entry, returning its
family and the metadata key it indexes. Returns None for non-sparse
keys. The metadata key is None for the legacy/global anonymous layout
(bare SPARSE_* constants written before per-key indexing existed) and
Some(key) for the per-key layout (SPARSE_*::key). parse_where parse_where_document sparse_max_key Per-key file_path map key for the WAND block-max blockfile. sparse_offset_value_key Per-key file_path map key for the WAND offset-value blockfile. sparse_posting_key Per-key file_path map key for the MaxScore posting blockfile. test_segment validate_name CountCollectionsResponse CountResponse CreateCollectionResponse DeletedMetadata ForkCollectionResponse GetCollectionByCrnResponse GetCollectionByIdResponse GetCollectionResponse GetDatabaseResponse ListCollectionsResponse ListDatabasesResponse Metadata UpdateMetadata