From d574a09529118e7d91468f151d2dd66d416a765c Mon Sep 17 00:00:00 2001
From: Wenwei Lin <75592450+wenwei-lin@users.noreply.github.com>
Date: Sat, 15 Feb 2025 22:50:05 +0800
Subject: [PATCH] fix: support html file in knowledge base (#1703)

---
 src/main/loader/index.ts | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)

diff --git a/src/main/loader/index.ts b/src/main/loader/index.ts
index e8444d2141..e69c3ff224 100644
--- a/src/main/loader/index.ts
+++ b/src/main/loader/index.ts
@@ -2,6 +2,7 @@ import * as fs from 'node:fs'
 
 import { LocalPathLoader, RAGApplication, TextLoader } from '@llm-tools/embedjs'
 import type { AddLoaderReturn } from '@llm-tools/embedjs-interfaces'
+import { WebLoader } from '@llm-tools/embedjs-loader-web'
 import { LoaderReturn } from '@shared/config/types'
 import { FileType, KnowledgeBaseParams } from '@types'
 import Logger from 'electron-log'
@@ -69,8 +70,26 @@ export async function addFileLoader(
     } as LoaderReturn
   }
 
-  // 文本类型
   const fileContent = fs.readFileSync(file.path, 'utf-8')
+  // HTML类型
+  if (['.html', '.htm'].includes(file.ext)) {
+    const loaderReturn = await ragApplication.addLoader(
+      new WebLoader({
+        urlOrContent: fileContent,
+        chunkSize: base.chunkSize,
+        chunkOverlap: base.chunkOverlap
+      }) as any,
+      forceReload
+    )
+    return {
+      entriesAdded: loaderReturn.entriesAdded,
+      uniqueId: loaderReturn.uniqueId,
+      uniqueIds: [loaderReturn.uniqueId],
+      loaderType: loaderReturn.loaderType
+    }
+  }
+
+  // 文本类型
   const loaderReturn = await ragApplication.addLoader(
     new TextLoader({ text: fileContent, chunkSize: base.chunkSize, chunkOverlap: base.chunkOverlap }) as any,
     forceReload