@knowledge-stack/ksapi 1.106.2 → 1.107.0

package/README.md

@@ -1,4 +1,4 @@
-# @knowledge-stack/ksapi@1.106.2
+# @knowledge-stack/ksapi@1.107.0
 
 A TypeScript SDK client for the localhost API.
 
@@ -577,7 +577,7 @@ and is automatically generated by the
 [OpenAPI Generator](https://openapi-generator.tech) project:
 
 - API version: `0.1.0`
-- Package version: `1.106.2`
+- Package version: `1.107.0`
 - Generator version: `7.21.0`
 - Build package: `org.openapitools.codegen.languages.TypeScriptFetchClientCodegen`
 

package/dist/esm/models/ChunkMetadata.d.ts

@@ -30,11 +30,17 @@ export interface ChunkMetadata {
      */
     s3Urls?: Array<string>;
     /**
-     * LLM-generated summary of the chunk content. Used for TABLE and HTML chunks to enrich embedding text.
+     * LLM-generated summary of the chunk content. Used for TABLE and HTML chunks to enrich embedding text, and for JSON/YAML chunks (with summarize_for_embedding) as the sole dense embedding text.
      * @type {string}
      * @memberof ChunkMetadata
      */
     summary?: string | null;
+    /**
+     * When True, this chunk's dense embedding is built from its LLM-generated summary (see summary) instead of its raw content. Set for parsed JSON/YAML single chunks so noisy structured text does not dominate the vector; the raw content is still kept for display and sparse (keyword) retrieval. Enrichment generates the summary when this is set and summary is empty.
+     * @type {boolean}
+     * @memberof ChunkMetadata
+     */
+    summarizeForEmbedding?: boolean;
     /**
      * S3 URI to extracted PDF text used for LLM grounding during enrichment
      * @type {string}

package/dist/esm/models/ChunkMetadata.js

@@ -31,6 +31,7 @@ export function ChunkMetadataFromJSONTyped(json, ignoreDiscriminator) {
         'polygons': json['polygons'] == null ? undefined : (json['polygons'].map(PolygonReferenceFromJSON)),
         's3Urls': json['s3_urls'] == null ? undefined : json['s3_urls'],
         'summary': json['summary'] == null ? undefined : json['summary'],
+        'summarizeForEmbedding': json['summarize_for_embedding'] == null ? undefined : json['summarize_for_embedding'],
         'extractedTextS3Uri': json['extracted_text_s3_uri'] == null ? undefined : json['extracted_text_s3_uri'],
         'secondaryTaxonomy': json['secondary_taxonomy'] == null ? undefined : ImageTaxonomyFromJSON(json['secondary_taxonomy']),
         'sheetName': json['sheet_name'] == null ? undefined : json['sheet_name'],
@@ -55,6 +56,7 @@ export function ChunkMetadataToJSONTyped(value, ignoreDiscriminator = false) {
         'polygons': value['polygons'] == null ? undefined : (value['polygons'].map(PolygonReferenceToJSON)),
         's3_urls': value['s3Urls'],
         'summary': value['summary'],
+        'summarize_for_embedding': value['summarizeForEmbedding'],
         'extracted_text_s3_uri': value['extractedTextS3Uri'],
         'secondary_taxonomy': ImageTaxonomyToJSON(value['secondaryTaxonomy']),
         'sheet_name': value['sheetName'],

package/dist/esm/models/DocumentType.d.ts

@@ -21,6 +21,8 @@ export declare const DocumentType: {
     readonly Xlsx: "XLSX";
     readonly Csv: "CSV";
     readonly Pptx: "PPTX";
+    readonly Json: "JSON";
+    readonly Yaml: "YAML";
     readonly Unknown: "UNKNOWN";
 };
 export type DocumentType = typeof DocumentType[keyof typeof DocumentType];

package/dist/esm/models/DocumentType.js

@@ -23,6 +23,8 @@ export const DocumentType = {
     Xlsx: 'XLSX',
     Csv: 'CSV',
     Pptx: 'PPTX',
+    Json: 'JSON',
+    Yaml: 'YAML',
     Unknown: 'UNKNOWN'
 };
 export function instanceOfDocumentType(value) {

package/dist/models/ChunkMetadata.d.ts

@@ -30,11 +30,17 @@ export interface ChunkMetadata {
      */
     s3Urls?: Array<string>;
     /**
-     * LLM-generated summary of the chunk content. Used for TABLE and HTML chunks to enrich embedding text.
+     * LLM-generated summary of the chunk content. Used for TABLE and HTML chunks to enrich embedding text, and for JSON/YAML chunks (with summarize_for_embedding) as the sole dense embedding text.
      * @type {string}
      * @memberof ChunkMetadata
      */
     summary?: string | null;
+    /**
+     * When True, this chunk's dense embedding is built from its LLM-generated summary (see summary) instead of its raw content. Set for parsed JSON/YAML single chunks so noisy structured text does not dominate the vector; the raw content is still kept for display and sparse (keyword) retrieval. Enrichment generates the summary when this is set and summary is empty.
+     * @type {boolean}
+     * @memberof ChunkMetadata
+     */
+    summarizeForEmbedding?: boolean;
     /**
      * S3 URI to extracted PDF text used for LLM grounding during enrichment
      * @type {string}

package/dist/models/ChunkMetadata.js

@@ -39,6 +39,7 @@ function ChunkMetadataFromJSONTyped(json, ignoreDiscriminator) {
         'polygons': json['polygons'] == null ? undefined : (json['polygons'].map(PolygonReference_1.PolygonReferenceFromJSON)),
         's3Urls': json['s3_urls'] == null ? undefined : json['s3_urls'],
         'summary': json['summary'] == null ? undefined : json['summary'],
+        'summarizeForEmbedding': json['summarize_for_embedding'] == null ? undefined : json['summarize_for_embedding'],
         'extractedTextS3Uri': json['extracted_text_s3_uri'] == null ? undefined : json['extracted_text_s3_uri'],
         'secondaryTaxonomy': json['secondary_taxonomy'] == null ? undefined : (0, ImageTaxonomy_1.ImageTaxonomyFromJSON)(json['secondary_taxonomy']),
         'sheetName': json['sheet_name'] == null ? undefined : json['sheet_name'],
@@ -63,6 +64,7 @@ function ChunkMetadataToJSONTyped(value, ignoreDiscriminator = false) {
         'polygons': value['polygons'] == null ? undefined : (value['polygons'].map(PolygonReference_1.PolygonReferenceToJSON)),
         's3_urls': value['s3Urls'],
         'summary': value['summary'],
+        'summarize_for_embedding': value['summarizeForEmbedding'],
         'extracted_text_s3_uri': value['extractedTextS3Uri'],
         'secondary_taxonomy': (0, ImageTaxonomy_1.ImageTaxonomyToJSON)(value['secondaryTaxonomy']),
         'sheet_name': value['sheetName'],

package/dist/models/DocumentType.d.ts

@@ -21,6 +21,8 @@ export declare const DocumentType: {
     readonly Xlsx: "XLSX";
     readonly Csv: "CSV";
     readonly Pptx: "PPTX";
+    readonly Json: "JSON";
+    readonly Yaml: "YAML";
     readonly Unknown: "UNKNOWN";
 };
 export type DocumentType = typeof DocumentType[keyof typeof DocumentType];

package/dist/models/DocumentType.js

@@ -31,6 +31,8 @@ exports.DocumentType = {
     Xlsx: 'XLSX',
     Csv: 'CSV',
     Pptx: 'PPTX',
+    Json: 'JSON',
+    Yaml: 'YAML',
     Unknown: 'UNKNOWN'
 };
 function instanceOfDocumentType(value) {

package/docs/ChunkMetadata.md

@@ -10,6 +10,7 @@ Name | Type
 `polygons` | [Array<PolygonReference>](PolygonReference.md)
 `s3Urls` | Array<string>
 `summary` | string
+`summarizeForEmbedding` | boolean
 `extractedTextS3Uri` | string
 `secondaryTaxonomy` | [ImageTaxonomy](ImageTaxonomy.md)
 `sheetName` | string
@@ -32,6 +33,7 @@ const example = {
   "polygons": null,
   "s3Urls": null,
   "summary": null,
+  "summarizeForEmbedding": null,
   "extractedTextS3Uri": null,
   "secondaryTaxonomy": null,
   "sheetName": null,

package/docs/DocumentsApi.md

@@ -589,7 +589,7 @@ example().catch(console.error);
 | **sortOrder** | `PathOrder` | Sort order for results (default: LOGICAL) | [Optional] [Defaults to `undefined`] [Enum: LOGICAL, NAME, UPDATED_AT, CREATED_AT] |
 | **sortDir** | `SortDirection` | Sort direction; overrides the column\'s natural default | [Optional] [Defaults to `undefined`] [Enum: ASC, DESC] |
 | **ownerId** | `string` | Filter to documents owned by this user | [Optional] [Defaults to `undefined`] |
-| **documentType** | `DocumentType` | Filter to documents of this type | [Optional] [Defaults to `undefined`] [Enum: PDF, DOCX, PLAINTEXT, IMAGE, XLSX, CSV, PPTX, UNKNOWN] |
+| **documentType** | `DocumentType` | Filter to documents of this type | [Optional] [Defaults to `undefined`] [Enum: PDF, DOCX, PLAINTEXT, IMAGE, XLSX, CSV, PPTX, JSON, YAML, UNKNOWN] |
 | **withTags** | `boolean` | Include tags in the response (default: false) | [Optional] [Defaults to `false`] |
 | **limit** | `number` | Number of items per page | [Optional] [Defaults to `20`] |
 | **offset** | `number` | Number of items to skip | [Optional] [Defaults to `0`] |

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@knowledge-stack/ksapi",
-  "version": "1.106.2",
+  "version": "1.107.0",
   "description": "OpenAPI client for @knowledge-stack/ksapi",
   "author": "OpenAPI-Generator",
   "repository": {

package/src/models/ChunkMetadata.ts

@@ -47,11 +47,17 @@ export interface ChunkMetadata {
      */
     s3Urls?: Array<string>;
     /**
-     * LLM-generated summary of the chunk content. Used for TABLE and HTML chunks to enrich embedding text.
+     * LLM-generated summary of the chunk content. Used for TABLE and HTML chunks to enrich embedding text, and for JSON/YAML chunks (with summarize_for_embedding) as the sole dense embedding text.
      * @type {string}
      * @memberof ChunkMetadata
      */
     summary?: string | null;
+    /**
+     * When True, this chunk's dense embedding is built from its LLM-generated summary (see summary) instead of its raw content. Set for parsed JSON/YAML single chunks so noisy structured text does not dominate the vector; the raw content is still kept for display and sparse (keyword) retrieval. Enrichment generates the summary when this is set and summary is empty.
+     * @type {boolean}
+     * @memberof ChunkMetadata
+     */
+    summarizeForEmbedding?: boolean;
     /**
      * S3 URI to extracted PDF text used for LLM grounding during enrichment
      * @type {string}
@@ -159,6 +165,7 @@ export function ChunkMetadataFromJSONTyped(json: any, ignoreDiscriminator: boole
         'polygons': json['polygons'] == null ? undefined : ((json['polygons'] as Array<any>).map(PolygonReferenceFromJSON)),
         's3Urls': json['s3_urls'] == null ? undefined : json['s3_urls'],
         'summary': json['summary'] == null ? undefined : json['summary'],
+        'summarizeForEmbedding': json['summarize_for_embedding'] == null ? undefined : json['summarize_for_embedding'],
         'extractedTextS3Uri': json['extracted_text_s3_uri'] == null ? undefined : json['extracted_text_s3_uri'],
         'secondaryTaxonomy': json['secondary_taxonomy'] == null ? undefined : ImageTaxonomyFromJSON(json['secondary_taxonomy']),
         'sheetName': json['sheet_name'] == null ? undefined : json['sheet_name'],
@@ -187,6 +194,7 @@ export function ChunkMetadataToJSONTyped(value?: ChunkMetadata | null, ignoreDis
         'polygons': value['polygons'] == null ? undefined : ((value['polygons'] as Array<any>).map(PolygonReferenceToJSON)),
         's3_urls': value['s3Urls'],
         'summary': value['summary'],
+        'summarize_for_embedding': value['summarizeForEmbedding'],
         'extracted_text_s3_uri': value['extractedTextS3Uri'],
         'secondary_taxonomy': ImageTaxonomyToJSON(value['secondaryTaxonomy']),
         'sheet_name': value['sheetName'],

package/src/models/DocumentType.ts

@@ -25,6 +25,8 @@ export const DocumentType = {
     Xlsx: 'XLSX',
     Csv: 'CSV',
     Pptx: 'PPTX',
+    Json: 'JSON',
+    Yaml: 'YAML',
     Unknown: 'UNKNOWN'
 } as const;
 export type DocumentType = typeof DocumentType[keyof typeof DocumentType];