fix: typesense doc sync (#11807)

Co-authored-by: Giorgio Boa <35845425+gioboa@users.noreply.github.com>

.github/workflows/docsearch.yml

@@ -12,7 +12,35 @@ jobs:
 
     steps:
       - uses: actions/checkout@v5
+      - name: Delete unaliased collections
+        env:
+          TYPESENSE_API_KEY: ${{ secrets.TYPESENSE_API_KEY }}
+          TYPESENSE_HOST: ${{ secrets.TYPESENSE_HOST }}
+          TYPESENSE_PROTOCOL: https
+          TYPESENSE_PORT: 443
+        run: |
+          ALIAS_COLLECTION=$(curl -s -H "X-TYPESENSE-API-KEY: $TYPESENSE_API_KEY" \
+            "$TYPESENSE_PROTOCOL://$TYPESENSE_HOST:$TYPESENSE_PORT/aliases/typeorm-docs" \
+            | jq -r '.collection_name')
 
+          if [ "$ALIAS_COLLECTION" = "null" ] || [ -z "$ALIAS_COLLECTION" ]; then
+            echo "Alias does not exist; skipping collection cleanup."
+            exit 0
+          fi
+
+          echo "Alias currently points to: $ALIAS_COLLECTION"
+
+          COLLECTIONS=$(curl -s -H "X-TYPESENSE-API-KEY: $TYPESENSE_API_KEY" \
+            "$TYPESENSE_PROTOCOL://$TYPESENSE_HOST:$TYPESENSE_PORT/collections" \
+            | jq -r '.[].name')
+
+          for col in $COLLECTIONS; do
+            if [ "$col" != "$ALIAS_COLLECTION" ]; then
+              echo "Deleting unaliased collection: $col"
+              curl -s -X DELETE -H "X-TYPESENSE-API-KEY: $TYPESENSE_API_KEY" \
+                "$TYPESENSE_PROTOCOL://$TYPESENSE_HOST:$TYPESENSE_PORT/collections/$col"
+            fi
+          done
       - run: |
           docker run \
             -e TYPESENSE_API_KEY=${{ secrets.TYPESENSE_API_KEY }} \

docs/docsearch-scraper-config.json

@@ -27,12 +27,7 @@
   "strip_chars": " .,;:#",
   "custom_settings": {
     "separatorsToIndex": "_",
-    "attributesForFaceting": [
-      "language",
-      "version",
-      "type",
-      "docusaurus_tag"
-    ],
+    "attributesForFaceting": [],
     "attributesToRetrieve": [
       "hierarchy",
       "content",
@@ -46,4 +41,4 @@
     "833762294"
   ],
   "nb_hits": 0
-}
+}