Release new docs to preview

Author: Milvus-doc-bot

v2.6.x/assets/bitmap.png

@@ -509,6 +509,11 @@
                         "token": "NuokbSgbroyVPQx14fKcm37bnoh",
                         "type": "image",
                         "alt_text": "deduplication-workflow"
+                    },
+                    {
+                        "token": "IbpqwpTFghdA37bU6dDc5eCUnfh",
+                        "type": "board",
+                        "alt_text": "load-workflow"
                     }
                 ]
             }

v2.6.x/assets/load-workflow.png

@@ -1261,6 +1261,12 @@
                 "id": "eviction.md",
                 "order": 2,
                 "children": []
+              },
+              {
+                "label": "Best Practices for Tiered Storage",
+                "id": "best-practices-for-tiered-storage.md",
+                "order": 3,
+                "children": []
               }
             ]
           }

v2.6.x/scripts/config.json

@@ -120,26 +120,17 @@ The following table lists the parameters that can be configured in `params` when
    </tr>
    <tr>
      <td><p><code>build_algo</code></p></td>
-     <td><p>Selects the graph generation algorithm before pruning. Possible values:</p>
-<ul>
-<li><p><code>IVF_PQ</code>: Offers higher quality but slower build time.</p></li>
-<li><p><code>NN_DESCENT</code>: Provides a quicker build with potentially lower recall.</p></li>
-</ul></td>
+     <td><p>Selects the graph generation algorithm before pruning. Possible values:</p><ul><li><p><code>IVF_PQ</code>: Offers higher quality but slower build time.</p></li><li><p><code>NN_DESCENT</code>: Provides a quicker build with potentially lower recall.</p></li></ul></td>
      <td><p><code>IVF_PQ</code></p></td>
    </tr>
    <tr>
      <td><p><code>cache_dataset_on_device</code></p></td>
-     <td><p>Decides whether to cache the original dataset in GPU memory. Possible values:</p>
-<ul>
-<li><p><code>"true"</code>: Caches the original dataset to enhance recall by refining search results.</p></li>
-<li><p><code>"false"</code>: Does not cache the original dataset to save gpu memory.</p></li>
-</ul></td>
+     <td><p>Decides whether to cache the original dataset in GPU memory. Possible values:</p><ul><li><p><code>"true"</code>: Caches the original dataset to enhance recall by refining search results.</p></li><li><p><code>"false"</code>: Does not cache the original dataset to save gpu memory.</p></li></ul></td>
      <td><p><code>"false"</code></p></td>
    </tr>
    <tr>
      <td><p><code>adapt_for_cpu</code></p></td>
-     <td><p>Decides whether to use GPU for index-building and CPU for search.
- Setting this parameter to <code>"true"</code> requires the presence of the <code>ef</code> parameter in the search requests.</p></td>
+     <td><p>Decides whether to use GPU for index-building and CPU for search.</p><p>Setting this parameter to <code>"true"</code> requires the presence of the <code>ef</code> parameter in the search requests.</p></td>
      <td><p><code>"false"</code></p></td>
    </tr>
 </table>
@@ -176,8 +167,7 @@ The following table lists the parameters that can be configured in `search_param
    </tr>
    <tr>
      <td><p><code>ef</code></p></td>
-     <td><p>Specifies the query time/accuracy trade-off. A higher <code>ef</code> value leads to more accurate but slower search.
- This parameter is mandatory if you set <code>adapt_for_cpu</code> to <code>true</code> when you build the index.</p></td>
+     <td><p>Specifies the query time/accuracy trade-off. A higher <code>ef</code> value leads to more accurate but slower search.</p><p>This parameter is mandatory if you set <code>adapt_for_cpu</code> to <code>true</code> when you build the index.</p></td>
      <td><p><code>[top_k, int_max]</code></p></td>
    </tr>
 </table>

v2.6.x/site/en/menuStructure/en.json

@@ -66,9 +66,9 @@ Use the `drop_index()` method to remove an existing index from a collection.
 
 <div class="alert note">
 
-- In **v2.6.3** or earlier, you must release the collection before dropping an index.
+- In **v2.6.3** or earlier, you must release the collection before dropping a scalar index.
 
-- From **v2.6.4** or later, you can drop an index directly once it’s no longer needed—no need to release the collection first.
+- From **v2.6.4** or later, you can drop a scalar index directly once it’s no longer needed—no need to release the collection first.
 
 </div>
 

v2.6.x/site/en/userGuide/indexes/gpu/gpu-cagra.md

@@ -107,9 +107,9 @@ Use the `drop_index()` method to remove an existing index from a collection.
 
 <div class="alert note">
 
-- In **v2.6.3** or earlier, you must release the collection before dropping an index.
+- In **v2.6.3** or earlier, you must release the collection before dropping a scalar index.
 
-- From **v2.6.4** or later, you can drop an index directly once it’s no longer needed—no need to release the collection first.
+- From **v2.6.4** or later, you can drop a scalar index directly once it’s no longer needed—no need to release the collection first.
 
 </div>
 

v2.6.x/site/en/userGuide/indexes/scalar/bitmap.md

@@ -235,9 +235,9 @@ Use the `drop_index()` method to remove an existing index from a collection.
 
 <div class="alert note">
 
-- In **v2.6.3** or earlier, you must release the collection before dropping an index.
+- In **v2.6.3** or earlier, you must release the collection before dropping a scalar index.
 
-- From **v2.6.4** or later, you can drop an index directly once it’s no longer needed—no need to release the collection first.
+- From **v2.6.4** or later, you can drop a scalar index directly once it’s no longer needed—no need to release the collection first.
 
 </div>
 

v2.6.x/site/en/userGuide/indexes/scalar/inverted.md

@@ -113,9 +113,9 @@ Use the `drop_index()` method to remove an existing index from a collection.
 
 <div class="alert note">
 
-- In **v2.6.3** or earlier, you must release the collection before dropping an index.
+- In **v2.6.3** or earlier, you must release the collection before dropping a scalar index.
 
-- From **v2.6.4** or later, you can drop an index directly once it’s no longer needed—no need to release the collection first.
+- From **v2.6.4** or later, you can drop a scalar index directly once it’s no longer needed—no need to release the collection first.
 
 </div>
 

v2.6.x/site/en/userGuide/indexes/scalar/ngram.md

@@ -138,6 +138,8 @@ To use an Array of Structs in Milvus, you need to define an array field when cre
 
 1. Set the field's `max_capacity` attribute to an appropriate value to specify the maximum number of Structs each entity can contain in this field.
 
+1. (**Optional**) You can set `mmap.enabled` for any field within the Struct element to balance the hot and cold data in the Struct.
+
 Here's how you can define a collection schema that includes an Array of Structs:
 
 <div class="multipleCode">
@@ -172,8 +174,8 @@ struct_schema = MilvusClient.create_struct_field_schema()
 struct_schema.add_field("text", DataType.VARCHAR, max_length=65535)
 struct_schema.add_field("chapter", DataType.VARCHAR, max_length=512)
 
-# add multiple vector fields to the struct
-struct_schema.add_field("text_vector", DataType.FLOAT_VECTOR, dim=5)
+# add a vector field to the struct with mmap enabled
+struct_schema.add_field("text_vector", DataType.FLOAT_VECTOR, mmap_enabled=True, dim=5)
 
 # reference the struct schema in an Array field with its 
 # element type set to `DataType.STRUCT`
@@ -445,6 +447,8 @@ Specifically, you can directly use the names of the vector fields within Struct
 
 Milvus provides `EmbeddingList` to help you organize query vectors for searches against an embedding list in an Array of Structs more neatly.
 
+However, `EmbeddingList` can used only in `search()` requests without range search or grouping search parameters, let alone `search_iterator()` requests.
+
 </div>
 
 <div class="multipleCode">