@ingram-cloud/sdk 1.8.0 → 1.9.0

package/dist/zod/vector-stores.js

@@ -229,23 +229,37 @@ export const VectorStoreFileBatchOut = z
 })
     .meta({ id: "VectorStoreFileBatchOut" });
 // ── Search ───────────────────────────────────────────────────────────────────
+/** OpenAI's `ranking_options`. `hybrid_search` turns on reciprocal-rank fusion
+ *  of the embedding ranking with a full-text (keyword) ranking, weighted as
+ *  given; without it search is embedding-only. */
+export const VectorStoreRankingOptions = z
+    .object({
+    score_threshold: z.number().min(0).max(1).optional(),
+    hybrid_search: z
+        .object({
+        embedding_weight: z.number().min(0),
+        text_weight: z.number().min(0),
+    })
+        .refine((w) => w.embedding_weight > 0 || w.text_weight > 0, {
+        message: "at least one weight must be positive",
+    })
+        .optional(),
+})
+    .meta({ id: "VectorStoreRankingOptions" });
 export const VectorStoreSearchIn = z
     .object({
     query: z.union([z.string().min(1), z.array(z.string().min(1)).min(1)]),
     max_num_results: z.number().int().min(1).max(50).optional(),
     filters: VectorStoreFilter.nullish(),
-    ranking_options: z
-        .object({
-        score_threshold: z.number().min(0).max(1).optional(),
-    })
-        .optional(),
+    ranking_options: VectorStoreRankingOptions.optional(),
 })
     .meta({ id: "VectorStoreSearchIn" });
 export const VectorStoreSearchResult = z
     .object({
     file_id: z.string(),
     filename: z.string(),
-    /** Cosine similarity in [0, 1]. */
+    /** In [0, 1]: cosine similarity, or — with `hybrid_search` — the fused
+     *  reciprocal-rank score (1 = first in both rankings). */
     score: z.number(),
     attributes: VectorStoreAttributes.nullable(),
     content: z.array(z.object({ type: z.literal("text"), text: z.string() })),

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@ingram-cloud/sdk",
-	"version": "1.8.0",
+	"version": "1.9.0",
 	"description": "Typed wire contract + management-plane client for the Ingram Cloud API: Zod schemas to validate request/response bodies, TypeScript types for the JSON you read back, SSE/webhook event types, and a typed REST client.",
 	"keywords": [
 		"api",

package/ts/zod/vector-stores.ts

@@ -268,16 +268,30 @@ export const VectorStoreFileBatchOut = z
 
 // ── Search ───────────────────────────────────────────────────────────────────
 
+/** OpenAI's `ranking_options`. `hybrid_search` turns on reciprocal-rank fusion
+ *  of the embedding ranking with a full-text (keyword) ranking, weighted as
+ *  given; without it search is embedding-only. */
+export const VectorStoreRankingOptions = z
+	.object({
+		score_threshold: z.number().min(0).max(1).optional(),
+		hybrid_search: z
+			.object({
+				embedding_weight: z.number().min(0),
+				text_weight: z.number().min(0),
+			})
+			.refine((w) => w.embedding_weight > 0 || w.text_weight > 0, {
+				message: "at least one weight must be positive",
+			})
+			.optional(),
+	})
+	.meta({ id: "VectorStoreRankingOptions" });
+
 export const VectorStoreSearchIn = z
 	.object({
 		query: z.union([z.string().min(1), z.array(z.string().min(1)).min(1)]),
 		max_num_results: z.number().int().min(1).max(50).optional(),
 		filters: VectorStoreFilter.nullish(),
-		ranking_options: z
-			.object({
-				score_threshold: z.number().min(0).max(1).optional(),
-			})
-			.optional(),
+		ranking_options: VectorStoreRankingOptions.optional(),
 	})
 	.meta({ id: "VectorStoreSearchIn" });
 
@@ -285,7 +299,8 @@ export const VectorStoreSearchResult = z
 	.object({
 		file_id: z.string(),
 		filename: z.string(),
-		/** Cosine similarity in [0, 1]. */
+		/** In [0, 1]: cosine similarity, or — with `hybrid_search` — the fused
+		 *  reciprocal-rank score (1 = first in both rankings). */
 		score: z.number(),
 		attributes: VectorStoreAttributes.nullable(),
 		content: z.array(z.object({ type: z.literal("text"), text: z.string() })),