perf(schema): cache compiled JSONPath expressions in document validation

redisvl/schema/validation.py

@@ -6,6 +6,7 @@
 """
 
 import json
+from functools import lru_cache
 from typing import Any
 
 from jsonpath_ng import parse as jsonpath_parse
@@ -210,6 +211,28 @@ def _validate_vector(cls, value):
         return type(model_name, (BaseModel,), class_dict)
 
 
+@lru_cache(maxsize=512)
+def _compile_json_path(path: str) -> Any:
+    """
+    Parse a JSONPath expression, reusing the result for repeated paths.
+
+    Schema field paths are fixed, and parsing costs ~1ms against microseconds to
+    evaluate, so validating per field re-did the expensive half every time.
+    Sharing one expression is safe: it holds no per-evaluation state.
+
+    Args:
+        path: JSONPath expression, with or without the leading ``$``
+
+    Returns:
+        The parsed jsonpath-ng expression
+    """
+    # If path doesn't start with $, add it as per JSONPath spec
+    if not path.startswith("$"):
+        path = f"$.{path}"
+
+    return jsonpath_parse(path)
+
+
 def extract_from_json_path(obj: dict[str, Any], path: str) -> Any:
     """
     Extract a value from a nested JSON object using a JSONPath expression.
@@ -226,13 +249,7 @@ def extract_from_json_path(obj: dict[str, Any], path: str) -> Any:
         and supports the full JSONPath specification including filters, wildcards,
         and array indexing.
     """
-    # If path doesn't start with $, add it as per JSONPath spec
-    if not path.startswith("$"):
-        path = f"$.{path}"
-
-    # Parse and find the JSONPath expression
-    jsonpath_expr = jsonpath_parse(path)
-    matches = jsonpath_expr.find(obj)
+    matches = _compile_json_path(path).find(obj)
 
     # Return the first match value, or None if no matches
     if matches:

tests/unit/test_validation.py

@@ -14,11 +14,13 @@
 import pytest
 
 from redisvl.schema import IndexSchema
+from redisvl.schema import validation as validation_module
 from redisvl.schema.fields import FieldTypes, VectorDataType
 from redisvl.schema.schema import StorageType
 from redisvl.schema.type_utils import TypeInferrer
 from redisvl.schema.validation import (
     SchemaModelGenerator,
+    _compile_json_path,
     extract_from_json_path,
     validate_object,
 )
@@ -301,6 +303,51 @@ def test_extract_from_json_path(self, valid_json_data, path, expected_value):
         """Test extracting values using JSON paths."""
         assert extract_from_json_path(valid_json_data, path) == expected_value
 
+    @pytest.fixture
+    def parse_calls(self, monkeypatch):
+        """Record every path handed to jsonpath-ng's parser."""
+        calls: list[str] = []
+        real_parse = validation_module.jsonpath_parse
+
+        def counting_parse(path):
+            calls.append(path)
+            return real_parse(path)
+
+        _compile_json_path.cache_clear()
+        monkeypatch.setattr(validation_module, "jsonpath_parse", counting_parse)
+        return calls
+
+    def test_repeated_path_compiles_once(self, valid_json_data, parse_calls):
+        """A repeated path compiles once, not once per call."""
+        for _ in range(25):
+            assert (
+                extract_from_json_path(valid_json_data, "$.metadata.user") == "user123"
+            )
+
+        assert parse_calls == ["$.metadata.user"]
+
+    def test_distinct_paths_each_compile(self, valid_json_data, parse_calls):
+        """Caching must not collapse distinct paths onto one expression."""
+        assert extract_from_json_path(valid_json_data, "$.metadata.user") == "user123"
+        assert extract_from_json_path(valid_json_data, "$.metadata.rating") == 4.5
+        assert len(parse_calls) == 2
+
+    def test_shared_expression_holds_no_state(self):
+        """Reusing one expression must not leak values between evaluations."""
+        _compile_json_path.cache_clear()
+
+        for expected in ("alice", "bob", "carol"):
+            obj = {"metadata": {"user": expected}}
+            assert extract_from_json_path(obj, "$.metadata.user") == expected
+
+        assert extract_from_json_path({"metadata": {}}, "$.metadata.user") is None
+
+    def test_leading_dollar_is_optional(self, valid_json_data):
+        """Both spellings of a path resolve to the same value."""
+        assert extract_from_json_path(
+            valid_json_data, "metadata.user"
+        ) == extract_from_json_path(valid_json_data, "$.metadata.user")
+
 
 # # -------------------- CATEGORY 2: PARAMETRIZED VALIDATOR TESTS --------------------