vector: take the list item field from the schema when batching vectors

Lance stores the fixed-size list item as nullable even when the write
schema declared it non-null, so PutVectors built batches that no longer
matched: expected FixedSizeList(4 x Float32) but found FixedSizeList(4 x
non-null Float32).

seaweed-vector/Cargo.lock

@@ -6486,6 +6486,7 @@ dependencies = [
  "clap",
  "futures",
  "lance",
+ "lance-arrow",
  "lance-linalg",
  "prost",
  "prost-types",

seaweed-vector/crates/server/Cargo.toml

@@ -16,6 +16,7 @@ path = "src/main.rs"
 weed-vector-core = { path = "../core" }
 weed-lance-worker = { path = "../../../seaweed-worker/crates/lance" }
 lance = { version = "10", default-features = false, features = ["aws"] }
+lance-arrow = "10"
 lance-linalg = "10"
 arrow-array = "58"
 arrow-cast = "58"

seaweed-vector/crates/server/src/service.rs

@@ -110,8 +110,18 @@ fn batch_of(vectors: &[pb::Vector], dimension: usize, schema: Arc<Schema>) -> Re
         .flat_map(|v| v.data.iter().copied())
         .collect::<Vec<_>>();
     let values: ArrayRef = Arc::new(Float32Array::from(flat));
+    // Lance stores the list item as nullable even when the write schema
+    // declared it non-null, so the child field must come from the batch's
+    // schema rather than a fixed declaration.
+    let DataType::FixedSizeList(item_field, _) = schema
+        .field_with_name(VECTOR_COLUMN)
+        .map_err(anyhow::Error::from)?
+        .data_type()
+    else {
+        bail!("vector column is not a fixed-size list");
+    };
     let vectors_array = FixedSizeListArray::try_new(
-        Field::new("item", DataType::Float32, false).into(),
+        item_field.clone(),
         dimension as i32,
         values,
         None,

weed/s3api/s3vectors/handlers.go

@@ -788,7 +788,7 @@ func filterExpr(filter map[string]interface{}) (string, error) {
 }
 
 func fieldClause(field string, cond interface{}) (string, error) {
-	col := fmt.Sprintf("json_get_str(`metadata`, '%s')", strings.ReplaceAll(field, "'", "''"))
+	col := fmt.Sprintf("json_get_string(`metadata`, '%s')", strings.ReplaceAll(field, "'", "''"))
 	switch c := cond.(type) {
 	case map[string]interface{}:
 		parts := []string{}