docs: Weaviate 1.39 release

_build_scripts/link-validator.js

@@ -45,6 +45,7 @@ const domainsToIgnore = [
     'medium.com', // TODO[g-despot]: started throwing Forbidden 403 (incl. subdomains, e.g. *.medium.com)
     'https://www.npmjs.com',
     'https://openai.com',
+    'https://platform.deepseek.com', // 403s automated requests (site loads fine in a browser)
     'https://platform.openai.com',
     'https://www.researchgate.net',
     'https://simple/',

_includes/code/csharp/ModelProvidersTest.cs

@@ -154,6 +154,95 @@ await client.Collections.Create(
         await client.Collections.Delete("DemoCollection");
     }
 
+    [Fact(Skip = "Requires MORPH_APIKEY, not configured in CI")]
+    public async Task TestMorphInstantiation()
+    {
+        // START MorphInstantiation
+        // Best practice: store your credentials in environment variables
+        string weaviateUrl = Environment.GetEnvironmentVariable("WEAVIATE_URL");
+        string weaviateApiKey = Environment.GetEnvironmentVariable("WEAVIATE_API_KEY");
+        string morphApiKey = Environment.GetEnvironmentVariable("MORPH_APIKEY");
+
+        // highlight-start
+        // Morph requests are built by Weaviate's OpenAI-compatible client,
+        // so the Morph key is supplied under the OpenAI header name.
+        using var client = await Connect.Cloud(
+            weaviateUrl,
+            weaviateApiKey,
+            headers: new Dictionary<string, string>
+            {
+                ["X-Openai-Api-Key"] = morphApiKey,
+            }
+        );
+
+        var meta = await client.GetMeta();
+        Console.WriteLine(meta.Version);
+        // highlight-end
+        // END MorphInstantiation
+    }
+
+    [Fact(Skip = "Requires MORPH_APIKEY, not configured in CI")]
+    public async Task TestMorphVectorizer()
+    {
+        if (await client.Collections.Exists("DemoCollection"))
+            await client.Collections.Delete("DemoCollection");
+
+        // START BasicVectorizerMorph
+        await client.Collections.Create(
+            new CollectionCreateParams
+            {
+                Name = "DemoCollection",
+                VectorConfig = new VectorConfigList
+                {
+                    Configure.Vector(
+                        "title_vector",
+                        v => v.Text2VecMorph(),
+                        sourceProperties: ["title"]
+                    ),
+                },
+                Properties = [Property.Text("title"), Property.Text("description")],
+            }
+        );
+        // END BasicVectorizerMorph
+
+        var config = await client.Collections.Export("DemoCollection");
+        Assert.True(config.VectorConfig.ContainsKey("title_vector"));
+        Assert.Equal("text2vec-morph", config.VectorConfig["title_vector"].Vectorizer.Identifier);
+
+        await client.Collections.Delete("DemoCollection");
+    }
+
+    [Fact(Skip = "Requires MORPH_APIKEY, not configured in CI")]
+    public async Task TestMorphVectorizerFull()
+    {
+        if (await client.Collections.Exists("DemoCollection"))
+            await client.Collections.Delete("DemoCollection");
+
+        // START FullVectorizerMorph
+        await client.Collections.Create(
+            new CollectionCreateParams
+            {
+                Name = "DemoCollection",
+                VectorConfig = new VectorConfigList
+                {
+                    Configure.Vector(
+                        "title_vector",
+                        v => v.Text2VecMorph(model: "morph-embedding-v3"),
+                        sourceProperties: ["title"]
+                    ),
+                },
+                Properties = [Property.Text("title"), Property.Text("description")],
+            }
+        );
+        // END FullVectorizerMorph
+
+        var config = await client.Collections.Export("DemoCollection");
+        Assert.True(config.VectorConfig.ContainsKey("title_vector"));
+        Assert.Equal("text2vec-morph", config.VectorConfig["title_vector"].Vectorizer.Identifier);
+
+        await client.Collections.Delete("DemoCollection");
+    }
+
     [Fact]
     public async Task TestWeaviateVectorizerModel()
     {

_includes/code/howto/configure.backups.py

@@ -191,6 +191,24 @@
 assert client.collections.exists("Article")
 assert client.collections.exists("Publication")
 
+# START ListBackups
+backups = client.backup.list_backups(
+    backend="filesystem",
+    sort_by_starting_time_asc=True,
+)
+
+for backup in backups:
+    print(backup.backup_id, backup.status, backup.incremental_base_backup_id)
+# END ListBackups
+
+# Test
+backup_ids = [backup.backup_id for backup in backups]
+assert {"base-backup", "incremental-backup-1", "incremental-backup-2"}.issubset(backup_ids)
+bases = {backup.backup_id: backup.incremental_base_backup_id for backup in backups}
+assert bases["base-backup"] is None
+assert bases["incremental-backup-1"] == "base-backup"
+assert bases["incremental-backup-2"] == "incremental-backup-1"
+
 # Clean up
 client.collections.delete(["Article", "Publication"])
 

_includes/code/howto/manage-data.create.py

@@ -182,6 +182,10 @@
 # WithGeoCoordinates
 # ========================================
 
+# Test scaffolding; not rendered in the docs. Start from an empty collection so
+# the exact count asserted below cannot pick up an object left by an earlier run.
+client.collections.delete("Publication")
+
 # START WithGeoCoordinates
 publications = client.collections.use("Publication")
 
@@ -196,23 +200,34 @@
 # END WithGeoCoordinates
 
 # TEST - Confirm insert & delete object
+# Test scaffolding below; not rendered in the docs.
+# A geo property is backed by its own index, and that index is filled in the
+# background when the server runs with ASYNC_INDEXING enabled (the test
+# instance does), so a geo filter does not see the object the moment insert
+# returns. Wait (bounded) for the write to become visible, so the assertion
+# tests the geo range and not index visibility. The assertion stays exact.
+import time
 from weaviate.classes.data import GeoCoordinate
 from weaviate.classes.query import Filter
 
-response = publications.query.fetch_objects(
-    filters=(
-        Filter
-        .by_property("headquartersGeoLocation")
-        .within_geo_range(
-            coordinate=GeoCoordinate(
-                latitude=52.39,
-                longitude=4.84
-            ),
-            distance=1000  # In meters
-        )
+geo_range_filter = (
+    Filter
+    .by_property("headquartersGeoLocation")
+    .within_geo_range(
+        coordinate=GeoCoordinate(
+            latitude=52.39,
+            longitude=4.84
+        ),
+        distance=1000  # In meters
     )
 )
 
+for _ in range(30):
+    response = publications.query.fetch_objects(filters=geo_range_filter)
+    if len(response.objects) >= 1:
+        break
+    time.sleep(1)
+
 assert len(response.objects) == 1
 obj_uuid = response.objects[0].uuid
 publications.data.delete_by_id(obj_uuid)

_includes/code/howto/search.bm25.gql.py

@@ -72,4 +72,29 @@
 
 gqlresponse = client.graphql_raw_query(gql_query)
 
+gql_query = """
+# START BM25OperatorCrossPropertyAnd
+{
+  Get {
+    JeopardyQuestion(
+      limit: 3
+      bm25: {
+        query: "Australian mammal cute"
+        # highlight-start
+        searchOperator: {
+          operator: AndCross,
+        }
+        # highlight-end
+      }
+    ) {
+      question
+      answer
+    }
+  }
+}
+# END BM25OperatorCrossPropertyAnd
+"""
+
+gqlresponse = client.graphql_raw_query(gql_query)
+
 client.close()

_includes/code/howto/search.bm25.py

@@ -107,6 +107,36 @@
 # End test
 
 
+# ============================
+# ===== BM25 w/ AND_CROSS =====
+# ============================
+
+# START BM25OperatorCrossPropertyAnd
+# highlight-start
+from weaviate.classes.query import BM25Operator
+# highlight-end
+
+jeopardy = client.collections.use("JeopardyQuestion")
+response = jeopardy.query.bm25(
+    # highlight-start
+    query="African desert wind",
+    # Each token must be matched by at least one searched property,
+    # but not necessarily all by the same property
+    operator=BM25Operator.and_cross(),
+    # highlight-end
+    limit=3,
+)
+
+for o in response.objects:
+    print(o.properties)
+# END BM25OperatorCrossPropertyAnd
+
+
+# Tests
+assert response.objects[0].collection == "JeopardyQuestion"
+# End test
+
+
 # ================================================
 # ===== BM25 Query with score / explainScore =====
 # ================================================

_includes/code/howto/search.similarity.mmr.py

@@ -126,6 +126,83 @@
 relevant_questions = [o.properties["question"] for o in response_relevant.objects]
 assert diverse_questions != relevant_questions, "Pure diversity and pure relevance should differ"
 
+# START MMRHybridExample
+from weaviate.classes.query import Diversity
+
+collection = client.collections.get("MMRDemo")
+
+# Fuse the keyword and vector results into 20 candidates, then select 5 diverse results
+response = collection.query.hybrid(
+    query="Question",
+    vector=base_vec,
+    limit=20,
+    # highlight-start
+    diversity_selection=Diversity.mmr(
+        limit=5,
+        balance=0.5,
+    ),
+    # highlight-end
+)
+
+for o in response.objects:
+    print(o.properties["question"])
+# END MMRHybridExample
+
+# Test
+assert len(response.objects) == 5
+hybrid_mmr_questions = [o.properties["question"] for o in response.objects]
+
+hybrid_standard = collection.query.hybrid(
+    query="Question",
+    vector=base_vec,
+    limit=5,
+)
+hybrid_standard_questions = [o.properties["question"] for o in hybrid_standard.objects]
+assert set(hybrid_standard_questions) != set(
+    hybrid_mmr_questions
+), "MMR should select different items than standard hybrid search"
+
+# START MMRPagination
+from weaviate.classes.query import Diversity
+
+collection = client.collections.get("MMRDemo")
+
+# The query limit is the size of the window that gets diversified.
+# The diversity limit is the page size.
+query_limit = 10
+page_size = 3
+
+# Advance offset by the query limit, NOT by the number of returned objects
+for offset in range(0, 30, query_limit):
+    page = collection.query.near_vector(
+        near_vector=base_vec,
+        limit=query_limit,
+        # highlight-start
+        offset=offset,
+        diversity_selection=Diversity.mmr(limit=page_size, balance=0.5),
+        # highlight-end
+    )
+
+    for o in page.objects:
+        print(offset, o.properties["question"])
+# END MMRPagination
+
+# Test
+paged_questions = []
+for offset in range(0, 30, query_limit):
+    page = collection.query.near_vector(
+        near_vector=base_vec,
+        limit=query_limit,
+        offset=offset,
+        diversity_selection=Diversity.mmr(limit=page_size, balance=0.5),
+    )
+    paged_questions += [o.properties["question"] for o in page.objects]
+
+assert len(paged_questions) == 9
+assert len(set(paged_questions)) == len(
+    paged_questions
+), "Advancing offset by the query limit must not repeat objects across pages"
+
 # Cleanup
 client.collections.delete("MMRDemo")
 client.close()