Merge pull request #635 from aws-samples/fix/kendra-document-uri

Kendra RAG のリンクの修正
aws-samples · Aug 30, 2024 · b20f564 · b20f564
2 parents e1cdd77 + 9c89b4f
commit b20f564
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 7 deletions.
diff --git a/packages/web/src/hooks/useRag.ts b/packages/web/src/hooks/useRag.ts
@@ -6,6 +6,7 @@ import { ShownMessage } from 'generative-ai-use-cases-jp';
 import { findModelByModelId } from './useModel';
 import { getPrompter } from '../prompts';
 import { RetrieveResultItem, DocumentAttribute } from '@aws-sdk/client-kendra';
+import { cleanEncode } from '../utils/URLUtils';
 
 // 同一のドキュメントとみなす Key 値
 const uniqueKeyOfItem = (item: RetrieveResultItem): string => {
@@ -154,13 +155,7 @@ const useRag = (id: string) => {
                       ? `(${_excerpt_page_number} ページ)`
                       : ''
                   }](
-                  ${
-                    item.DocumentURI
-                      ? encodeURI(item.DocumentURI)
-                          .replace(/\(/g, '\\(')
-                          .replace(/\)/g, '\\)')
-                      : ''
-                  }${
+                  ${item.DocumentURI ? cleanEncode(item.DocumentURI) : ''}${
                     _excerpt_page_number ? `#page=${_excerpt_page_number}` : ''
                   })`
                 : '';

diff --git a/packages/web/src/utils/URLUtils.ts b/packages/web/src/utils/URLUtils.ts
@@ -0,0 +1,17 @@
+// RFC3986 Encoding
+// https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/encodeURIComponent#encoding_for_rfc3986
+const encodeRFC3986URI = (uri: string) => {
+  return encodeURI(uri).replace(
+    /[!'()*]/g,
+    (c) => `%${c.charCodeAt(0).toString(16).toUpperCase()}`
+  );
+};
+
+export const cleanEncode = (uri: string): string => {
+  // Check if uri is encoded, and encode it if not already encoded
+  if (decodeURI(uri) === uri) {
+    return encodeRFC3986URI(uri);
+  } else {
+    return uri;
+  }
+};