@retrivora-ai/rag-engine 0.3.3 → 0.3.4

package/dist/{chunk-SNPFZBDK.mjs → chunk-GE4AVGA5.mjs}

@@ -1009,8 +1009,18 @@ var GeminiProvider = class {
   constructor(llmConfig, embeddingConfig) {
     if (!llmConfig.apiKey) throw new Error("[GeminiProvider] llmConfig.apiKey is required");
     this.client = new GoogleGenAI({ apiKey: llmConfig.apiKey });
-    this.llmConfig = llmConfig;
-    this.embeddingConfig = embeddingConfig;
+    this.llmConfig = __spreadProps(__spreadValues({}, llmConfig), {
+      model: this.sanitizeModel(llmConfig.model)
+    });
+    if (embeddingConfig) {
+      this.embeddingConfig = __spreadProps(__spreadValues({}, embeddingConfig), {
+        model: this.sanitizeModel(embeddingConfig.model)
+      });
+    }
+  }
+  sanitizeModel(model) {
+    if (!model) return model;
+    return model.split(":")[0];
   }
   async chat(messages, context, options) {
     var _a, _b, _c, _d, _e, _f;
@@ -1039,8 +1049,10 @@ ${context}`;
     return (_f = response.text) != null ? _f : "";
   }
   async embed(text, options) {
-    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j;
-    const model = (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004";
+    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j, _k;
+    const model = this.sanitizeModel(
+      (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004"
+    );
     const apiKey = (_e = (_d = this.embeddingConfig) == null ? void 0 : _d.apiKey) != null ? _e : this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey ? new GoogleGenAI({ apiKey }) : this.client;
     let content = text;
@@ -1057,14 +1069,20 @@ ${context}`;
     }
     const response = await client.models.embedContent({
       model,
-      contents: content
+      contents: content,
+      config: ((_h = this.embeddingConfig) == null ? void 0 : _h.dimensions) ? { outputDimensionality: this.embeddingConfig.dimensions } : void 0
+    }).catch((err) => {
+      console.error(`[GeminiProvider] Embedding failed for model "${model}":`, err.message);
+      throw err;
     });
-    return (_j = (_i = (_h = response.embeddings) == null ? void 0 : _h[0]) == null ? void 0 : _i.values) != null ? _j : [];
+    return (_k = (_j = (_i = response.embeddings) == null ? void 0 : _i[0]) == null ? void 0 : _j.values) != null ? _k : [];
   }
   async batchEmbed(texts, options) {
-    var _a, _b, _c, _d, _e, _f, _g, _h, _i;
+    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j;
     const vectors = [];
-    const model = (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004";
+    const model = this.sanitizeModel(
+      (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004"
+    );
     const apiKey = (_e = (_d = this.embeddingConfig) == null ? void 0 : _d.apiKey) != null ? _e : this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey ? new GoogleGenAI({ apiKey }) : this.client;
     let contents = texts;
@@ -1077,12 +1095,16 @@ ${context}`;
     }
     const response = await client.models.embedContent({
       model,
-      contents
+      contents,
+      config: ((_h = this.embeddingConfig) == null ? void 0 : _h.dimensions) ? { outputDimensionality: this.embeddingConfig.dimensions } : void 0
+    }).catch((err) => {
+      console.error(`[GeminiProvider] Batch embedding failed for model "${model}":`, err.message);
+      throw err;
     });
-    return (_i = (_h = response.embeddings) == null ? void 0 : _h.map((e) => {
+    return (_j = (_i = response.embeddings) == null ? void 0 : _i.map((e) => {
       var _a2;
       return (_a2 = e.values) != null ? _a2 : [];
-    })) != null ? _i : [];
+    })) != null ? _j : [];
   }
   async ping() {
     try {

package/dist/handlers/index.js

@@ -2177,8 +2177,18 @@ var GeminiProvider = class {
   constructor(llmConfig, embeddingConfig) {
     if (!llmConfig.apiKey) throw new Error("[GeminiProvider] llmConfig.apiKey is required");
     this.client = new import_genai.GoogleGenAI({ apiKey: llmConfig.apiKey });
-    this.llmConfig = llmConfig;
-    this.embeddingConfig = embeddingConfig;
+    this.llmConfig = __spreadProps(__spreadValues({}, llmConfig), {
+      model: this.sanitizeModel(llmConfig.model)
+    });
+    if (embeddingConfig) {
+      this.embeddingConfig = __spreadProps(__spreadValues({}, embeddingConfig), {
+        model: this.sanitizeModel(embeddingConfig.model)
+      });
+    }
+  }
+  sanitizeModel(model) {
+    if (!model) return model;
+    return model.split(":")[0];
   }
   async chat(messages, context, options) {
     var _a, _b, _c, _d, _e, _f;
@@ -2207,8 +2217,10 @@ ${context}`;
     return (_f = response.text) != null ? _f : "";
   }
   async embed(text, options) {
-    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j;
-    const model = (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004";
+    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j, _k;
+    const model = this.sanitizeModel(
+      (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004"
+    );
     const apiKey = (_e = (_d = this.embeddingConfig) == null ? void 0 : _d.apiKey) != null ? _e : this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey ? new import_genai.GoogleGenAI({ apiKey }) : this.client;
     let content = text;
@@ -2225,14 +2237,20 @@ ${context}`;
     }
     const response = await client.models.embedContent({
       model,
-      contents: content
+      contents: content,
+      config: ((_h = this.embeddingConfig) == null ? void 0 : _h.dimensions) ? { outputDimensionality: this.embeddingConfig.dimensions } : void 0
+    }).catch((err) => {
+      console.error(`[GeminiProvider] Embedding failed for model "${model}":`, err.message);
+      throw err;
     });
-    return (_j = (_i = (_h = response.embeddings) == null ? void 0 : _h[0]) == null ? void 0 : _i.values) != null ? _j : [];
+    return (_k = (_j = (_i = response.embeddings) == null ? void 0 : _i[0]) == null ? void 0 : _j.values) != null ? _k : [];
   }
   async batchEmbed(texts, options) {
-    var _a, _b, _c, _d, _e, _f, _g, _h, _i;
+    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j;
     const vectors = [];
-    const model = (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004";
+    const model = this.sanitizeModel(
+      (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004"
+    );
     const apiKey = (_e = (_d = this.embeddingConfig) == null ? void 0 : _d.apiKey) != null ? _e : this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey ? new import_genai.GoogleGenAI({ apiKey }) : this.client;
     let contents = texts;
@@ -2245,12 +2263,16 @@ ${context}`;
     }
     const response = await client.models.embedContent({
       model,
-      contents
+      contents,
+      config: ((_h = this.embeddingConfig) == null ? void 0 : _h.dimensions) ? { outputDimensionality: this.embeddingConfig.dimensions } : void 0
+    }).catch((err) => {
+      console.error(`[GeminiProvider] Batch embedding failed for model "${model}":`, err.message);
+      throw err;
     });
-    return (_i = (_h = response.embeddings) == null ? void 0 : _h.map((e) => {
+    return (_j = (_i = response.embeddings) == null ? void 0 : _i.map((e) => {
       var _a2;
       return (_a2 = e.values) != null ? _a2 : [];
-    })) != null ? _i : [];
+    })) != null ? _j : [];
   }
   async ping() {
     try {

package/dist/handlers/index.mjs

@@ -3,7 +3,7 @@ import {
   createHealthHandler,
   createIngestHandler,
   createUploadHandler
-} from "../chunk-SNPFZBDK.mjs";
+} from "../chunk-GE4AVGA5.mjs";
 import "../chunk-EDLTMSNY.mjs";
 import "../chunk-FWCSY2DS.mjs";
 export {

package/dist/server.js

@@ -2221,8 +2221,18 @@ var GeminiProvider = class {
   constructor(llmConfig, embeddingConfig) {
     if (!llmConfig.apiKey) throw new Error("[GeminiProvider] llmConfig.apiKey is required");
     this.client = new import_genai.GoogleGenAI({ apiKey: llmConfig.apiKey });
-    this.llmConfig = llmConfig;
-    this.embeddingConfig = embeddingConfig;
+    this.llmConfig = __spreadProps(__spreadValues({}, llmConfig), {
+      model: this.sanitizeModel(llmConfig.model)
+    });
+    if (embeddingConfig) {
+      this.embeddingConfig = __spreadProps(__spreadValues({}, embeddingConfig), {
+        model: this.sanitizeModel(embeddingConfig.model)
+      });
+    }
+  }
+  sanitizeModel(model) {
+    if (!model) return model;
+    return model.split(":")[0];
   }
   async chat(messages, context, options) {
     var _a, _b, _c, _d, _e, _f;
@@ -2251,8 +2261,10 @@ ${context}`;
     return (_f = response.text) != null ? _f : "";
   }
   async embed(text, options) {
-    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j;
-    const model = (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004";
+    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j, _k;
+    const model = this.sanitizeModel(
+      (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004"
+    );
     const apiKey = (_e = (_d = this.embeddingConfig) == null ? void 0 : _d.apiKey) != null ? _e : this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey ? new import_genai.GoogleGenAI({ apiKey }) : this.client;
     let content = text;
@@ -2269,14 +2281,20 @@ ${context}`;
     }
     const response = await client.models.embedContent({
       model,
-      contents: content
+      contents: content,
+      config: ((_h = this.embeddingConfig) == null ? void 0 : _h.dimensions) ? { outputDimensionality: this.embeddingConfig.dimensions } : void 0
+    }).catch((err) => {
+      console.error(`[GeminiProvider] Embedding failed for model "${model}":`, err.message);
+      throw err;
     });
-    return (_j = (_i = (_h = response.embeddings) == null ? void 0 : _h[0]) == null ? void 0 : _i.values) != null ? _j : [];
+    return (_k = (_j = (_i = response.embeddings) == null ? void 0 : _i[0]) == null ? void 0 : _j.values) != null ? _k : [];
   }
   async batchEmbed(texts, options) {
-    var _a, _b, _c, _d, _e, _f, _g, _h, _i;
+    var _a, _b, _c, _d, _e, _f, _g, _h, _i, _j;
     const vectors = [];
-    const model = (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004";
+    const model = this.sanitizeModel(
+      (_c = (_b = options == null ? void 0 : options.model) != null ? _b : (_a = this.embeddingConfig) == null ? void 0 : _a.model) != null ? _c : "text-embedding-004"
+    );
     const apiKey = (_e = (_d = this.embeddingConfig) == null ? void 0 : _d.apiKey) != null ? _e : this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey ? new import_genai.GoogleGenAI({ apiKey }) : this.client;
     let contents = texts;
@@ -2289,12 +2307,16 @@ ${context}`;
     }
     const response = await client.models.embedContent({
       model,
-      contents
+      contents,
+      config: ((_h = this.embeddingConfig) == null ? void 0 : _h.dimensions) ? { outputDimensionality: this.embeddingConfig.dimensions } : void 0
+    }).catch((err) => {
+      console.error(`[GeminiProvider] Batch embedding failed for model "${model}":`, err.message);
+      throw err;
     });
-    return (_i = (_h = response.embeddings) == null ? void 0 : _h.map((e) => {
+    return (_j = (_i = response.embeddings) == null ? void 0 : _i.map((e) => {
       var _a2;
       return (_a2 = e.values) != null ? _a2 : [];
-    })) != null ? _i : [];
+    })) != null ? _j : [];
   }
   async ping() {
     try {

package/dist/server.mjs

@@ -34,7 +34,7 @@ import {
   createIngestHandler,
   createUploadHandler,
   getRagConfig
-} from "./chunk-SNPFZBDK.mjs";
+} from "./chunk-GE4AVGA5.mjs";
 import "./chunk-EDLTMSNY.mjs";
 import {
   PineconeProvider

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@retrivora-ai/rag-engine",
-  "version": "0.3.3",
+  "version": "0.3.4",
   "description": "Retrivora AI is a plug-and-play AI engine for RAG chat experiences — generic vector DB + LLM provider, embeddable or standalone.",
   "author": "Abhinav Alkuchi",
   "license": "MIT",

package/src/llm/providers/GeminiProvider.ts

@@ -26,10 +26,28 @@ export class GeminiProvider implements ILLMProvider {
   constructor(llmConfig: LLMConfig, embeddingConfig?: EmbeddingConfig) {
     if (!llmConfig.apiKey) throw new Error('[GeminiProvider] llmConfig.apiKey is required');
     this.client = new GoogleGenAI({ apiKey: llmConfig.apiKey });
-    this.llmConfig = llmConfig;
-    this.embeddingConfig = embeddingConfig;
+    
+    // Sanitize model names (e.g. remove ':latest' suffix used by Ollama)
+    this.llmConfig = {
+      ...llmConfig,
+      model: this.sanitizeModel(llmConfig.model)
+    };
+    
+    if (embeddingConfig) {
+      this.embeddingConfig = {
+        ...embeddingConfig,
+        model: this.sanitizeModel(embeddingConfig.model)
+      };
+    }
   }
 
+  private sanitizeModel(model: string): string {
+    if (!model) return model;
+    // Strip :latest suffix common in Ollama but invalid in Google SDK
+    return model.split(':')[0];
+  }
+
+
   async chat(messages: ChatMessage[], context: string, options?: ChatOptions): Promise<string> {
     const systemPrompt =
       this.llmConfig.systemPrompt ??
@@ -62,10 +80,11 @@ export class GeminiProvider implements ILLMProvider {
   }
 
   async embed(text: string, options?: EmbedOptions): Promise<number[]> {
-    const model =
+    const model = this.sanitizeModel(
       options?.model ??
       this.embeddingConfig?.model ??
-      'text-embedding-004';
+      'text-embedding-004'
+    );
 
     const apiKey = this.embeddingConfig?.apiKey ?? this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey
@@ -91,6 +110,10 @@ export class GeminiProvider implements ILLMProvider {
     const response = await client.models.embedContent({
       model,
       contents: content,
+      config: this.embeddingConfig?.dimensions ? { outputDimensionality: this.embeddingConfig.dimensions } : undefined
+    }).catch(err => {
+      console.error(`[GeminiProvider] Embedding failed for model "${model}":`, err.message);
+      throw err;
     });
 
     return response.embeddings?.[0]?.values ?? [];
@@ -101,10 +124,11 @@ export class GeminiProvider implements ILLMProvider {
     // Sequential fallback for Gemini (API doesn't have a direct batch endpoint in the same way, though you can pass multiple contents.
     // For simplicity and to match the other providers' structure, we'll iterate or pass array.
     // The new SDK supports an array of contents.
-    const model =
+    const model = this.sanitizeModel(
       options?.model ??
       this.embeddingConfig?.model ??
-      'text-embedding-004';
+      'text-embedding-004'
+    );
 
     const apiKey = this.embeddingConfig?.apiKey ?? this.llmConfig.apiKey;
     const client = apiKey !== this.llmConfig.apiKey
@@ -126,6 +150,10 @@ export class GeminiProvider implements ILLMProvider {
     const response = await client.models.embedContent({
       model,
       contents,
+      config: this.embeddingConfig?.dimensions ? { outputDimensionality: this.embeddingConfig.dimensions } : undefined
+    }).catch(err => {
+      console.error(`[GeminiProvider] Batch embedding failed for model "${model}":`, err.message);
+      throw err;
     });
 
     return response.embeddings?.map(e => e.values ?? []) ?? [];