Merge pull request #100 from Stranger6667/dd/cache-validator

perf: Cache JSON Schema validators by their schema's JSON representation

Author: Zac-HD

.github/workflows/ci.yml

@@ -65,7 +65,7 @@ jobs:
   release:
     runs-on: ubuntu-latest
     needs: [check, test, test-slow]
-    if: github.repository == 'Zac-HD/hypothesis-jsonschema' &&  github.ref == 'refs/heads/master'
+    if: github.repository == 'python-jsonschema/hypothesis-jsonschema' &&  github.ref == 'refs/heads/master'
     steps:
     - uses: actions/checkout@v2
     - name: Set up Python 3.8

CHANGELOG.md

@@ -1,5 +1,8 @@
 # Changelog
 
+#### 0.22.1 - 2023-02-07
+- Cache JSON Schema validators by their schema's JSON representation
+
 #### 0.22.0 - 2021-12-15
 - never generate trailing newlines for regex patterns ending in `$`
   (allowed by Python, but not by JSON Schema)

deps/check.in

@@ -10,7 +10,6 @@ flake8-docstrings
 flake8-mutable
 # flake8-noqa  # See https://github.com/JBKahn/flake8-print/issues/50
 flake8-print
-flake8-simplify
 flake8-strftime
 mypy
 pep8-naming

deps/check.txt

@@ -4,8 +4,6 @@
 #
 #    pip-compile --output-file=deps/check.txt deps/check.in
 #
-astor==0.8.1
-    # via flake8-simplify
 attrs==21.2.0
     # via flake8-bugbear
 autoflake==1.4
@@ -36,7 +34,6 @@ flake8==4.0.1
     #   flake8-mutable
     #   flake8-polyfill
     #   flake8-print
-    #   flake8-simplify
     #   flake8-strftime
     #   pep8-naming
 flake8-2020==1.6.1
@@ -63,8 +60,6 @@ flake8-polyfill==1.0.2
     #   pep8-naming
 flake8-print==4.0.0
     # via -r deps/check.in
-flake8-simplify==0.14.2
-    # via -r deps/check.in
 flake8-strftime==0.3.2
     # via -r deps/check.in
 gitdb==4.0.9
@@ -77,7 +72,6 @@ importlib-metadata==4.2.0
     #   flake8
     #   flake8-2020
     #   flake8-comprehensions
-    #   flake8-simplify
     #   stevedore
 isort==5.10.1
     # via shed

src/hypothesis_jsonschema/__init__.py

@@ -3,7 +3,7 @@
 The only public API is `from_schema`; check the docstring for details.
 """
 
-__version__ = "0.22.0"
+__version__ = "0.22.1"
 __all__ = ["from_schema"]
 
 from ._from_schema import from_schema

src/hypothesis_jsonschema/_canonicalise.py

@@ -18,6 +18,7 @@
 import math
 import re
 from fractions import Fraction
+from functools import lru_cache
 from typing import Any, Dict, List, Optional, Tuple, Union
 
 import jsonschema
@@ -69,7 +70,33 @@ def next_down(val: float) -> float:
     return out
 
 
+class CacheableSchema:
+    """Cache schema by its JSON representation.
+
+    Canonicalisation is not required as schemas with the same JSON representation
+    will have the same validator.
+    """
+
+    __slots__ = ("schema", "encoded")
+
+    def __init__(self, schema: Schema) -> None:
+        self.schema = schema
+        self.encoded = hash(json.dumps(schema, sort_keys=True))
+
+    def __eq__(self, other: "CacheableSchema") -> bool:  # type: ignore
+        return self.encoded == other.encoded
+
+    def __hash__(self) -> int:
+        return self.encoded
+
+
 def _get_validator_class(schema: Schema) -> JSONSchemaValidator:
+    return __get_validator_class(CacheableSchema(schema))
+
+
+@lru_cache(maxsize=128)
+def __get_validator_class(wrapper: CacheableSchema) -> JSONSchemaValidator:
+    schema = wrapper.schema
     with contextlib.suppress(jsonschema.exceptions.SchemaError):
         validator = jsonschema.validators.validator_for(schema)
         validator.check_schema(schema)