From 600d4ceb29ee94b5a140d0f72dbbbc38419b9a7a Mon Sep 17 00:00:00 2001
From: ItzCrazyKns <95534749+ItzCrazyKns@users.noreply.github.com>
Date: Thu, 23 Oct 2025 23:06:05 +0530
Subject: [PATCH] feat(hf-transformer): use langchain's inbuilt transformer
 class

---
 docker-compose.yaml                           |  2 +-
 docs/API/SEARCH.md                            |  1 +
 .../Settings/SettingsButtonMobile.tsx         |  2 +-
 src/lib/huggingfaceTransformer.ts             | 76 -------------------
 src/lib/models/providers/transformers.ts      |  3 +-
 5 files changed, 4 insertions(+), 80 deletions(-)
 delete mode 100644 src/lib/huggingfaceTransformer.ts

diff --git a/docker-compose.yaml b/docker-compose.yaml
index d9d811d..50b6785 100644
--- a/docker-compose.yaml
+++ b/docker-compose.yaml
@@ -2,7 +2,7 @@ services:
   perplexica:
     image: itzcrazykns1337/perplexica:latest
     ports:
-      - "3000:3000"
+      - '3000:3000'
     volumes:
       - data:/home/perplexica/data
       - uploads:/home/perplexica/uploads
diff --git a/docs/API/SEARCH.md b/docs/API/SEARCH.md
index bf0db7a..04f11ef 100644
--- a/docs/API/SEARCH.md
+++ b/docs/API/SEARCH.md
@@ -17,6 +17,7 @@ Before making search requests, you'll need to get the available providers and th
 Returns a list of all active providers with their available chat and embedding models.
 
 **Response Example:**
+
 ```json
 {
   "providers": [
diff --git a/src/components/Settings/SettingsButtonMobile.tsx b/src/components/Settings/SettingsButtonMobile.tsx
index 836a4eb..db26f4c 100644
--- a/src/components/Settings/SettingsButtonMobile.tsx
+++ b/src/components/Settings/SettingsButtonMobile.tsx
@@ -9,7 +9,7 @@ const SettingsButtonMobile = () => {
   return (
     <>
       <button className="lg:hidden" onClick={() => setIsOpen(true)}>
-        <Settings size={18}/>
+        <Settings size={18} />
       </button>
       <AnimatePresence>
         {isOpen && <SettingsDialogue isOpen={isOpen} setIsOpen={setIsOpen} />}
diff --git a/src/lib/huggingfaceTransformer.ts b/src/lib/huggingfaceTransformer.ts
deleted file mode 100644
index a8c21bd..0000000
--- a/src/lib/huggingfaceTransformer.ts
+++ /dev/null
@@ -1,76 +0,0 @@
-import { Embeddings, type EmbeddingsParams } from '@langchain/core/embeddings';
-import { chunkArray } from '@langchain/core/utils/chunk_array';
-
-export interface HuggingFaceTransformersEmbeddingsParams
-  extends EmbeddingsParams {
-  modelName: string;
-
-  model: string;
-
-  timeout?: number;
-
-  batchSize?: number;
-
-  stripNewLines?: boolean;
-}
-
-export class HuggingFaceTransformersEmbeddings
-  extends Embeddings
-  implements HuggingFaceTransformersEmbeddingsParams
-{
-  modelName = 'Xenova/all-MiniLM-L6-v2';
-
-  model = 'Xenova/all-MiniLM-L6-v2';
-
-  batchSize = 512;
-
-  stripNewLines = true;
-
-  timeout?: number;
-
-  constructor(fields?: Partial<HuggingFaceTransformersEmbeddingsParams>) {
-    super(fields ?? {});
-
-    this.modelName = fields?.model ?? fields?.modelName ?? this.model;
-    this.model = this.modelName;
-    this.stripNewLines = fields?.stripNewLines ?? this.stripNewLines;
-    this.timeout = fields?.timeout;
-  }
-
-  async embedDocuments(texts: string[]): Promise<number[][]> {
-    const batches = chunkArray(
-      this.stripNewLines ? texts.map((t) => t.replace(/\n/g, ' ')) : texts,
-      this.batchSize,
-    );
-
-    const batchRequests = batches.map((batch) => this.runEmbedding(batch));
-    const batchResponses = await Promise.all(batchRequests);
-    const embeddings: number[][] = [];
-
-    for (let i = 0; i < batchResponses.length; i += 1) {
-      const batchResponse = batchResponses[i];
-      for (let j = 0; j < batchResponse.length; j += 1) {
-        embeddings.push(batchResponse[j]);
-      }
-    }
-
-    return embeddings;
-  }
-
-  async embedQuery(text: string): Promise<number[]> {
-    const data = await this.runEmbedding([
-      this.stripNewLines ? text.replace(/\n/g, ' ') : text,
-    ]);
-    return data[0];
-  }
-
-  private async runEmbedding(texts: string[]) {
-    const { pipeline } = await import('@huggingface/transformers');
-    const pipe = await pipeline('feature-extraction', this.model);
-
-    return this.caller.call(async () => {
-      const output = await pipe(texts, { pooling: 'mean', normalize: true });
-      return output.tolist();
-    });
-  }
-}
diff --git a/src/lib/models/providers/transformers.ts b/src/lib/models/providers/transformers.ts
index 1468155..afd6b9e 100644
--- a/src/lib/models/providers/transformers.ts
+++ b/src/lib/models/providers/transformers.ts
@@ -4,8 +4,7 @@ import BaseModelProvider from './baseProvider';
 import { Embeddings } from '@langchain/core/embeddings';
 import { UIConfigField } from '@/lib/config/types';
 import { getConfiguredModelProviderById } from '@/lib/config/serverRegistry';
-import { HuggingFaceTransformersEmbeddings } from '@/lib/huggingfaceTransformer';
-
+import { HuggingFaceTransformersEmbeddings } from '@langchain/community/embeddings/huggingface_transformers';
 interface TransformersConfig {}
 
 const defaultEmbeddingModels: Model[] = [