n8n/packages/nodes-base/nodes/Databricks/resources/vectorSearch/parameters.ts

import type { INodeProperties } from 'n8n-workflow';

export const vectorSearchParameters: INodeProperties[] = [
	{
		displayName: 'Index Name',
		name: 'indexName',
		type: 'string',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['getIndex', 'queryIndex', 'createIndex'],
			},
		},
		default: '',
		description: 'Name of the vector search index',
	},
	{
		displayName: 'Endpoint Name',
		name: 'endpointName',
		type: 'string',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['listIndexes', 'createIndex'],
			},
		},
		default: '',
		description: 'Name of the vector search endpoint',
	},
	{
		displayName: 'Primary Key',
		name: 'primaryKey',
		type: 'string',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['createIndex'],
			},
		},
		default: '',
		placeholder: 'ID',
		description: 'Primary key column of the index',
	},
	{
		displayName: 'Index Type',
		name: 'indexType',
		type: 'options',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['createIndex'],
			},
		},
		options: [
			{
				name: 'Delta Sync',
				value: 'DELTA_SYNC',
				description: 'Automatically syncs with a source Delta Table',
			},
			{
				name: 'Direct Access',
				value: 'DIRECT_ACCESS',
				description: 'Supports direct read and write of vectors and metadata',
			},
		],
		default: 'DELTA_SYNC',
		description: 'Type of vector search index to create',
	},
	{
		displayName: 'Delta Sync Index Spec',
		name: 'deltaSyncIndexSpec',
		type: 'json',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['createIndex'],
				indexType: ['DELTA_SYNC'],
			},
		},
		default:
			'{\n  "source_table": "catalog.schema.table",\n  "pipeline_type": "TRIGGERED",\n  "embedding_source_columns": [{\n    "name": "text",\n    "embedding_model_endpoint_name": "e5-small-v2"\n  }],\n  "columns_to_sync": ["id", "text"]\n}',
		description: 'Specification for the Delta Sync index',
		typeOptions: {
			rows: 8,
		},
	},
	{
		displayName: 'Direct Access Index Spec',
		name: 'directAccessIndexSpec',
		type: 'json',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['createIndex'],
				indexType: ['DIRECT_ACCESS'],
			},
		},
		default:
			'{\n  "embedding_vector_columns": [{\n    "name": "embedding",\n    "embedding_dimension": 1536\n  }],\n  "schema_json": "{}"\n}',
		description: 'Specification for the Direct Access index',
		typeOptions: {
			rows: 6,
		},
	},
	{
		displayName: 'Query Type',
		name: 'queryType',
		type: 'options',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
			},
		},
		options: [
			{
				name: 'Text Query',
				value: 'text',
				description: 'Query using text (automatically converted to vectors)',
			},
			{
				name: 'Vector Query',
				value: 'vector',
				description: 'Query using pre-computed vector embeddings',
			},
		],
		default: 'text',
		description: 'Type of query to perform',
	},
	{
		displayName: 'Query Text',
		name: 'queryText',
		type: 'string',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
				queryType: ['text'],
			},
		},
		default: '',
		description: 'Text to search for (will be automatically converted to embeddings)',
		placeholder: 'What is machine learning?',
	},
	{
		displayName: 'Query Vector',
		name: 'queryVector',
		type: 'json',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
				queryType: ['vector'],
			},
		},
		default: '[]',
		description: 'Vector embeddings to search for similar vectors (array of numbers)',
		placeholder: '[0.1, 0.2, 0.3, ...]',
	},
	{
		displayName: 'Search Mode',
		name: 'searchMode',
		type: 'options',
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
			},
		},
		options: [
			{
				name: 'Hybrid',
				value: 'HYBRID',
				description: 'Combines semantic (vector) and keyword search for best results',
			},
			{
				name: 'ANN (Approximate Nearest Neighbor)',
				value: 'ANN',
				description: 'Pure vector similarity search',
			},
		],
		default: 'ANN',
		description: 'Search algorithm to use',
	},
	{
		displayName: 'Columns to Return',
		name: 'columns',
		type: 'string',
		required: true,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
			},
		},
		default: '',
		description:
			'Comma-separated list of column names to return in results (e.g., "content,URL,title")',
		placeholder: 'content, url',
	},
	{
		displayName: 'Number of Results',
		name: 'numResults',
		type: 'number',
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
			},
		},
		default: 10,
		description: 'Maximum number of results to return',
	},
	{
		displayName: 'Enable Reranking',
		name: 'enableReranking',
		type: 'boolean',
		default: false,
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
			},
		},
		description: 'Whether to rerank results using a reranker model for improved relevance',
	},
	{
		displayName: 'Reranker Model',
		name: 'rerankerModel',
		type: 'string',
		required: true,
		default: 'databricks_reranker',
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
				enableReranking: [true],
			},
		},
		description: 'Name of the reranker model to use',
	},
	{
		displayName: 'Columns to Rerank',
		name: 'columnsToRerank',
		type: 'string',
		required: true,
		default: '',
		placeholder: 'content',
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
				enableReranking: [true],
			},
		},
		description: 'Comma-separated list of columns to use for reranking (e.g., "content,title")',
	},
	{
		displayName: 'Options',
		name: 'options',
		type: 'collection',
		placeholder: 'Add Option',
		default: {},
		displayOptions: {
			show: {
				resource: ['vectorSearch'],
				operation: ['queryIndex'],
			},
		},
		options: [
			{
				displayName: 'Filter Expression',
				name: 'filterExpression',
				type: 'string',
				default: '',
				description:
					'SQL-like filter expression to apply to the results (e.g., "category = \'docs\' AND published = true")',
				placeholder: 'category = "documentation"',
			},
			{
				displayName: 'Score Threshold',
				name: 'scoreThreshold',
				type: 'number',
				default: 0,
				description: 'Minimum relevance score threshold for results. Must be ≥ 0 and ≤ 1.',
				typeOptions: {
					minValue: 0,
					maxValue: 1,
					numberPrecision: 2,
				},
			},
		],
	},
];