IndexProps

Properties for defining a Index.

See https://www.alibabacloud.com/help/ros/developer-reference/aliyun-bailian-index

Initializer

import ros_cdk_bailian
ros_cdk_bailian.IndexProps(
  name: typing.Union[str, IResolvable],
  source_type: typing.Union[str, IResolvable],
  structure_type: typing.Union[str, IResolvable],
  workspace_id: typing.Union[str, IResolvable],
  category_ids: typing.Union[IResolvable, typing.List[typing.Union[str, IResolvable]]] = None,
  chunk_mode: typing.Union[str, IResolvable] = None,
  chunk_size: typing.Union[typing.Union[int, float], IResolvable] = None,
  connect_id: typing.Union[str, IResolvable] = None,
  database: typing.Union[str, IResolvable] = None,
  dense_similarity_top_k: typing.Union[typing.Union[int, float], IResolvable] = None,
  description: typing.Union[str, IResolvable] = None,
  document_ids: typing.Union[IResolvable, typing.List[typing.Union[str, IResolvable]]] = None,
  embedding_model_name: typing.Union[str, IResolvable] = None,
  enable_headers: typing.Union[bool, IResolvable] = None,
  enable_rewrite: typing.Union[bool, IResolvable] = None,
  knowledge_scene: typing.Union[str, IResolvable] = None,
  knowledge_type: typing.Union[str, IResolvable] = None,
  meta_extract_columns: typing.Union[IResolvable, typing.List[typing.Union[IResolvable, MetaExtractColumnsProperty]]] = None,
  overlap_size: typing.Union[typing.Union[int, float], IResolvable] = None,
  pipeline_commercial_cu: typing.Union[typing.Union[int, float], IResolvable] = None,
  pipeline_commercial_type: typing.Union[str, IResolvable] = None,
  pipeline_retrieve_rate_limit_strategy: typing.Union[str, IResolvable] = None,
  rerank_instruct: typing.Union[str, IResolvable] = None,
  rerank_min_score: typing.Union[typing.Union[int, float], IResolvable] = None,
  rerank_mode: typing.Union[str, IResolvable] = None,
  rerank_model_name: typing.Union[str, IResolvable] = None,
  separator: typing.Union[str, IResolvable] = None,
  sink_instance_id: typing.Union[str, IResolvable] = None,
  sink_region: typing.Union[str, IResolvable] = None,
  sink_type: typing.Union[str, IResolvable] = None,
  sparse_similarity_top_k: typing.Union[typing.Union[int, float], IResolvable] = None,
  table: typing.Union[str, IResolvable] = None,
  table_ids: typing.Union[IResolvable, typing.List[typing.Union[str, IResolvable]]] = None
)

Properties

name^Required

name: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property name: Knowledge base name.

1-20 characters. Supports Chinese, English, numbers, underscores (_), hyphens (-), periods (.) and colons (:).

source_type^Required

source_type: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property sourceType: Data source type.

DATA_CENTER_CATEGORY: import files from specified categories. DATA_CENTER_FILE: import specified files.

structure_type^Required

structure_type: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property structureType: Knowledge base type.

unstructured: document search or audio/video knowledge base. structured: data query or image Q&A knowledge base.

workspace_id^Required

workspace_id: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property workspaceId: Workspace ID.

The knowledge base will be created in this workspace.

category_ids^Optional

category_ids: typing.Union[IResolvable, typing.List[typing.Union[str, IResolvable]]]

• Type: typing.Union[ros_cdk_core.IResolvable, typing.List[typing.Union[str, ros_cdk_core.IResolvable]]]

Property categoryIds: Category IDs to import when creating the knowledge base.

Required when SourceType is DATA_CENTER_CATEGORY.

chunk_mode^Optional

chunk_mode: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property chunkMode: Chunk strategy.

If not specified, smart chunking is used. When set to regex, Separator must also be provided.

chunk_size^Optional

chunk_size: typing.Union[typing.Union[int, float], IResolvable]

• Type: typing.Union[typing.Union[int, float], ros_cdk_core.IResolvable]

Property chunkSize: Chunk size, the maximum number of characters per text slice.

Range [1-6000]. Defaults to 500. If ChunkSize is less than 100, OverlapSize must also be set.

connect_id^Optional

connect_id: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property connectId: Connection ID.

Replaces the deprecated datasourceCode parameter.

database^Optional

database: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property database: Database name for structured knowledge base.

dense_similarity_top_k^Optional

dense_similarity_top_k: typing.Union[typing.Union[int, float], IResolvable]

• Type: typing.Union[typing.Union[int, float], ros_cdk_core.IResolvable]

Property denseSimilarityTopK: Vector retrieval Top K.

Retrieves the K most similar text slices by vector similarity. Range [0-100]. Defaults to 100. DenseSimilarityTopK + SparseSimilarityTopK must be <= 200. Only supported in UpdateIndex, not in CreateIndex.

description^Optional

description: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property description: Knowledge base description.

0-1000 characters.

document_ids^Optional

document_ids: typing.Union[IResolvable, typing.List[typing.Union[str, IResolvable]]]

• Type: typing.Union[ros_cdk_core.IResolvable, typing.List[typing.Union[str, ros_cdk_core.IResolvable]]]

Property documentIds: Document IDs to import when creating the knowledge base.

Required when SourceType is DATA_CENTER_FILE.

embedding_model_name^Optional

embedding_model_name: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property embeddingModelName: Embedding model used by the knowledge base.

Defaults to text-embedding-v3.

enable_headers^Optional

enable_headers: typing.Union[bool, IResolvable]

• Type: typing.Union[bool, ros_cdk_core.IResolvable]

Property enableHeaders: Whether to enable headers for document parsing.

Defaults to false.

enable_rewrite^Optional

enable_rewrite: typing.Union[bool, IResolvable]

• Type: typing.Union[bool, ros_cdk_core.IResolvable]

Property enableRewrite: Whether to enable multi-turn dialogue rewriting.

Defaults to true.

knowledge_scene^Optional

knowledge_scene: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property knowledgeScene: Knowledge base scene.

Must be provided together with KnowledgeType. The valid values depend on KnowledgeType.

knowledge_type^Optional

knowledge_type: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property knowledgeType: Knowledge base type classification.

Must be provided together with KnowledgeScene. If both are omitted, the system uses default based on StructureType.

meta_extract_columns^Optional

meta_extract_columns: typing.Union[IResolvable, typing.List[typing.Union[IResolvable, MetaExtractColumnsProperty]]]

• Type: typing.Union[ros_cdk_core.IResolvable, typing.List[typing.Union[ros_cdk_core.IResolvable, MetaExtractColumnsProperty]]]

Property metaExtractColumns: Metadata extraction configuration.

overlap_size^Optional

overlap_size: typing.Union[typing.Union[int, float], IResolvable]

• Type: typing.Union[typing.Union[int, float], ros_cdk_core.IResolvable]

Property overlapSize: Overlap size, the number of overlapping characters between adjacent text slices.

Range [0-1024]. Defaults to 100. Must be less than ChunkSize.

pipeline_commercial_cu^Optional

pipeline_commercial_cu: typing.Union[typing.Union[int, float], IResolvable]

• Type: typing.Union[typing.Union[int, float], ros_cdk_core.IResolvable]

Property pipelineCommercialCu: RCU count for enterprise edition.

Only required when PipelineCommercialType is enterprise. Range [1-200].

pipeline_commercial_type^Optional

pipeline_commercial_type: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property pipelineCommercialType: Knowledge base specification type.

standard: Standard edition. enterprise: Enterprise edition. Defaults to standard.

pipeline_retrieve_rate_limit_strategy^Optional

pipeline_retrieve_rate_limit_strategy: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property pipelineRetrieveRateLimitStrategy: Retrieval rate limit strategy for enterprise edition.

Defaults to downgrade.

rerank_instruct^Optional

rerank_instruct: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property rerankInstruct: Custom rerank instruction.

Required when RerankMode is custom.

rerank_min_score^Optional

rerank_min_score: typing.Union[typing.Union[int, float], IResolvable]

• Type: typing.Union[typing.Union[int, float], ros_cdk_core.IResolvable]

Property rerankMinScore: Similarity threshold.

Only text slices with similarity score exceeding this value will be recalled. Range [0.01-1.00]. Defaults to 0.01.

rerank_mode^Optional

rerank_mode: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property rerankMode: Rerank mode.

qa: Q&A scenario (default). similar: similarity scenario. custom: custom rerank instruction, requires RerankInstruct.

rerank_model_name^Optional

rerank_model_name: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property rerankModelName: Rerank model used by the knowledge base.

Defaults to qwen3-rerank when not specified.

separator^Optional

separator: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property separator: Separator for chunking.

Only effective when ChunkMode is regex.

sink_instance_id^Optional

sink_instance_id: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property sinkInstanceId: Vector storage instance ID.

Required when SinkType is ADB.

sink_region^Optional

sink_region: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property sinkRegion: Region of the vector storage instance.

Required when SinkType is ADB.

sink_type^Optional

sink_type: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property sinkType: Vector storage type.

BUILT_IN: managed by Bailian platform. ADB: AnalyticDB for PostgreSQL. Defaults to BUILT_IN. When set to ADB, SinkInstanceId and SinkRegion are required.

sparse_similarity_top_k^Optional

sparse_similarity_top_k: typing.Union[typing.Union[int, float], IResolvable]

• Type: typing.Union[typing.Union[int, float], ros_cdk_core.IResolvable]

Property sparseSimilarityTopK: Keyword retrieval Top K.

Retrieves text slices by exact keyword matching. Range [0-100]. Defaults to 100. DenseSimilarityTopK + SparseSimilarityTopK must be <= 200. Only supported in UpdateIndex, not in CreateIndex.

table^Optional

table: typing.Union[str, IResolvable]

• Type: typing.Union[str, ros_cdk_core.IResolvable]

Property table: Table name for structured knowledge base.

table_ids^Optional

table_ids: typing.Union[IResolvable, typing.List[typing.Union[str, IResolvable]]]

• Type: typing.Union[ros_cdk_core.IResolvable, typing.List[typing.Union[str, ros_cdk_core.IResolvable]]]

Property tableIds: Table IDs to import when creating the knowledge base.