use internal aggregation range class for date aggregation ranges

Author: miguelgrinberg

elasticsearch/dsl/query.py

@@ -1084,7 +1084,7 @@ class Knn(Query):
     :arg similarity: The minimum similarity for a vector to be considered
         a match
     :arg rescore_vector: Apply oversampling and rescoring to quantized
-        vectors *
+        vectors
     :arg boost: Floating point number used to decrease or increase the
         relevance scores of the query. Boost values are relative to the
         default value of 1.0. A boost value between 0 and 1.0 decreases

elasticsearch/dsl/types.py

@@ -324,31 +324,58 @@ class DenseVectorIndexOptions(AttrDict[Any]):
         `int4_flat` index types.
     :arg ef_construction: The number of candidates to track while
         assembling the list of nearest neighbors for each new node.  Only
-        applicable to `hnsw`, `int8_hnsw`, and `int4_hnsw` index types.
-        Defaults to `100` if omitted.
+        applicable to `hnsw`, `int8_hnsw`, `bbq_hnsw`, and `int4_hnsw`
+        index types. Defaults to `100` if omitted.
     :arg m: The number of neighbors each node will be connected to in the
-        HNSW graph.  Only applicable to `hnsw`, `int8_hnsw`, and
-        `int4_hnsw` index types. Defaults to `16` if omitted.
+        HNSW graph.  Only applicable to `hnsw`, `int8_hnsw`, `bbq_hnsw`,
+        and `int4_hnsw` index types. Defaults to `16` if omitted.
+    :arg rescore_vector: The rescore vector options. This is only
+        applicable to `bbq_hnsw`, `int4_hnsw`, `int8_hnsw`, `bbq_flat`,
+        `int4_flat`, and `int8_flat` index types.
     """
 
     type: Union[
-        Literal["flat", "hnsw", "int4_flat", "int4_hnsw", "int8_flat", "int8_hnsw"],
+        Literal[
+            "bbq_flat",
+            "bbq_hnsw",
+            "flat",
+            "hnsw",
+            "int4_flat",
+            "int4_hnsw",
+            "int8_flat",
+            "int8_hnsw",
+        ],
         DefaultType,
     ]
     confidence_interval: Union[float, DefaultType]
     ef_construction: Union[int, DefaultType]
     m: Union[int, DefaultType]
+    rescore_vector: Union[
+        "DenseVectorIndexOptionsRescoreVector", Dict[str, Any], DefaultType
+    ]
 
     def __init__(
         self,
         *,
         type: Union[
-            Literal["flat", "hnsw", "int4_flat", "int4_hnsw", "int8_flat", "int8_hnsw"],
+            Literal[
+                "bbq_flat",
+                "bbq_hnsw",
+                "flat",
+                "hnsw",
+                "int4_flat",
+                "int4_hnsw",
+                "int8_flat",
+                "int8_hnsw",
+            ],
             DefaultType,
         ] = DEFAULT,
         confidence_interval: Union[float, DefaultType] = DEFAULT,
         ef_construction: Union[int, DefaultType] = DEFAULT,
         m: Union[int, DefaultType] = DEFAULT,
+        rescore_vector: Union[
+            "DenseVectorIndexOptionsRescoreVector", Dict[str, Any], DefaultType
+        ] = DEFAULT,
         **kwargs: Any,
     ):
         if type is not DEFAULT:
@@ -359,6 +386,29 @@ def __init__(
             kwargs["ef_construction"] = ef_construction
         if m is not DEFAULT:
             kwargs["m"] = m
+        if rescore_vector is not DEFAULT:
+            kwargs["rescore_vector"] = rescore_vector
+        super().__init__(kwargs)
+
+
+class DenseVectorIndexOptionsRescoreVector(AttrDict[Any]):
+    """
+    :arg oversample: (required) The oversampling factor to use when
+        searching for the nearest neighbor. This is only applicable to the
+        quantized formats: `bbq_*`, `int4_*`, and `int8_*`. When provided,
+        `oversample * k` vectors will be gathered and then their scores
+        will be re-computed with the original vectors.  valid values are
+        between `1.0` and `10.0` (inclusive), or `0` exactly to disable
+        oversampling.
+    """
+
+    oversample: Union[float, DefaultType]
+
+    def __init__(
+        self, *, oversample: Union[float, DefaultType] = DEFAULT, **kwargs: Any
+    ):
+        if oversample is not DEFAULT:
+            kwargs["oversample"] = oversample
         super().__init__(kwargs)
 
 

utils/dsl-generator.py

@@ -345,12 +345,12 @@ def get_python_type(self, schema_type, for_response=False):
             ]["name"].endswith("Analyzer"):
                 # not expanding analyzers at this time, maybe in the future
                 return "str, Dict[str, Any]", None
-            elif (
-                schema_type["name"]["namespace"] == "_types.aggregations"
-                and schema_type["name"]["name"].endswith("AggregationRange")
-                and schema_type["name"]["name"] != "IpRangeAggregationRange"
-            ):
-                return '"wrappers.AggregationRange"', None
+            elif schema_type["name"]["namespace"] == "_types.aggregations":
+                if (
+                    schema_type["name"]["name"].endswith("AggregationRange")
+                    or schema_type["name"]["name"] == "DateRangeExpression"
+                ) and schema_type["name"]["name"] != "IpRangeAggregationRange":
+                    return '"wrappers.AggregationRange"', None
 
             # to handle other interfaces we generate a type of the same name
             # and add the interface to the interfaces.py module