jsonschema-path 0.4.2__tar.gz → 0.4.3__tar.gz

{jsonschema_path-0.4.2 → jsonschema_path-0.4.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: jsonschema-path
-Version: 0.4.2
+Version: 0.4.3
 Summary: JSONSchema Spec with object-oriented paths
 License: Apache-2.0
 License-File: LICENSE
@@ -120,6 +120,21 @@ Usage
    {'type': 'string', 'default': '1.0'}
 
 
+Resolved cache
+##############
+
+The resolved-path cache is intended for repeated path lookups and may significantly improve
+``read_value``/membership hot paths. Cache entries are invalidated when the
+resolver registry evolves during reference resolution.
+
+This cache is optional and disabled by default
+(``resolved_cache_maxsize=0``). You can enable it when creating paths or
+accessors, for example:
+
+.. code-block:: python
+
+   >>> path = SchemaPath.from_dict(d, resolved_cache_maxsize=64)
+
 Benchmarks
 ##########
 
@@ -140,12 +155,13 @@ For a quick smoke run:
    poetry run python -m tests.benchmarks.bench_parse --output reports/bench-parse.quick.json --quick
    poetry run python -m tests.benchmarks.bench_lookup --output reports/bench-lookup.quick.json --quick
 
-You can also control repeats/warmup via env vars:
+You can also control repeats/warmup and resolved cache maxsize via env vars:
 
 .. code-block:: console
 
    export JSONSCHEMA_PATH_BENCH_REPEATS=5
    export JSONSCHEMA_PATH_BENCH_WARMUP=1
+   export JSONSCHEMA_PATH_BENCH_RESOLVED_CACHE_MAXSIZE=64
 
 Compare two results:
 

{jsonschema_path-0.4.2 → jsonschema_path-0.4.3}/README.rst

@@ -90,6 +90,21 @@ Usage
    {'type': 'string', 'default': '1.0'}
 
 
+Resolved cache
+##############
+
+The resolved-path cache is intended for repeated path lookups and may significantly improve
+``read_value``/membership hot paths. Cache entries are invalidated when the
+resolver registry evolves during reference resolution.
+
+This cache is optional and disabled by default
+(``resolved_cache_maxsize=0``). You can enable it when creating paths or
+accessors, for example:
+
+.. code-block:: python
+
+   >>> path = SchemaPath.from_dict(d, resolved_cache_maxsize=64)
+
 Benchmarks
 ##########
 
@@ -110,12 +125,13 @@ For a quick smoke run:
    poetry run python -m tests.benchmarks.bench_parse --output reports/bench-parse.quick.json --quick
    poetry run python -m tests.benchmarks.bench_lookup --output reports/bench-lookup.quick.json --quick
 
-You can also control repeats/warmup via env vars:
+You can also control repeats/warmup and resolved cache maxsize via env vars:
 
 .. code-block:: console
 
    export JSONSCHEMA_PATH_BENCH_REPEATS=5
    export JSONSCHEMA_PATH_BENCH_WARMUP=1
+   export JSONSCHEMA_PATH_BENCH_RESOLVED_CACHE_MAXSIZE=64
 
 Compare two results:
 

{jsonschema_path-0.4.2 → jsonschema_path-0.4.3}/jsonschema_path/__init__.py

@@ -4,7 +4,7 @@ from jsonschema_path.paths import SchemaPath
 
 __author__ = "Artur Maciag"
 __email__ = "maciag.artur@gmail.com"
-__version__ = "0.4.2"
+__version__ = "0.4.3"
 __url__ = "https://github.com/p1c2u/jsonschema-path"
 __license__ = "Apache-2.0"
 

{jsonschema_path-0.4.2 → jsonschema_path-0.4.3}/jsonschema_path/accessors.py

@@ -1,5 +1,6 @@
 """JSONSchema spec accessors module."""
 
+from collections import OrderedDict
 from collections.abc import Hashable
 from collections.abc import Iterator
 from collections.abc import Sequence
@@ -27,9 +28,23 @@ from jsonschema_path.utils import is_ref
 class SchemaAccessor(LookupAccessor):
     _resolver_refs: dict[int, Resolver[Schema] | None] = {}
 
-    def __init__(self, schema: Schema, resolver: Resolver[Schema]):
+    def __init__(
+        self,
+        schema: Schema,
+        resolver: Resolver[Schema],
+        resolved_cache_maxsize: int = 0,
+    ):
+        if resolved_cache_maxsize < 0:
+            raise ValueError("resolved_cache_maxsize must be >= 0")
+
         super().__init__(cast(LookupNode, schema))
         self.resolver = resolver
+        self._resolved_cache_maxsize = resolved_cache_maxsize
+        self._resolved_cache: OrderedDict[
+            tuple[tuple[LookupKey, ...], int],
+            Resolved[LookupNode],
+        ] = OrderedDict()
+        self._resolver_version = 0
 
         self._resolver_refs[id(schema)] = resolver
 
@@ -40,6 +55,7 @@ class SchemaAccessor(LookupAccessor):
         specification: Specification[Schema] = DRAFT202012,
         base_uri: str = "",
         handlers: ResolverHandlers | None = None,
+        resolved_cache_maxsize: int = 0,
     ) -> "SchemaAccessor":
         if handlers is None:
             handlers = default_handlers
@@ -50,7 +66,11 @@ class SchemaAccessor(LookupAccessor):
         )
         registry = registry.with_resource(base_uri, base_resource)
         resolver = registry.resolver(base_uri=base_uri)
-        return cls(schema, resolver)
+        return cls(
+            schema,
+            resolver,
+            resolved_cache_maxsize=resolved_cache_maxsize,
+        )
 
     def __getitem__(self, parts: Sequence[LookupKey]) -> LookupNode:
         resolved = self.get_resolved(parts)
@@ -143,13 +163,46 @@ class SchemaAccessor(LookupAccessor):
             pass
 
     def get_resolved(self, parts: Sequence[LookupKey]) -> Resolved[LookupNode]:
+        cache_key = self._resolved_cache_key(parts)
+        if cache_key is not None:
+            cached_resolved = self._resolved_cache.get(cache_key)
+            if cached_resolved is not None:
+                self._resolved_cache.move_to_end(cache_key)
+                return cached_resolved
+
         resolved = self._get_resolved(self.node, parts, resolver=self.resolver)
-        self.resolver = self.resolver._evolve(
-            self.resolver._base_uri,
-            registry=resolved.resolver._registry,
-        )
+        if resolved.resolver._registry is not self.resolver._registry:
+            self.resolver = self.resolver._evolve(
+                self.resolver._base_uri,
+                registry=resolved.resolver._registry,
+            )
+            self._resolver_version += 1
+            self._resolved_cache.clear()
+
+        cache_key = self._resolved_cache_key(parts)
+        if cache_key is not None:
+            self._resolved_cache[cache_key] = resolved
+            self._resolved_cache.move_to_end(cache_key)
+            if len(self._resolved_cache) > self._resolved_cache_maxsize:
+                self._resolved_cache.popitem(last=False)
+
         return resolved
 
+    def _resolved_cache_key(
+        self,
+        parts: Sequence[LookupKey],
+    ) -> tuple[tuple[LookupKey, ...], int] | None:
+        if self._resolved_cache_maxsize <= 0:
+            return None
+
+        parts_tuple = tuple(parts)
+        try:
+            hash(parts_tuple)
+        except TypeError:
+            return None
+
+        return (parts_tuple, self._resolver_version)
+
     @classmethod
     def _get_resolved(
         cls,

{jsonschema_path-0.4.2 → jsonschema_path-0.4.3}/jsonschema_path/paths.py

@@ -105,6 +105,7 @@ class SchemaPath(AccessorPath[SchemaNode, SchemaKey, SchemaValue]):
         specification: Specification[Schema] = DRAFT202012,
         base_uri: str = "",
         handlers: ResolverHandlers = default_handlers,
+        resolved_cache_maxsize: int = 0,
         spec_url: str | None = None,
         ref_resolver_handlers: ResolverHandlers | None = None,
     ) -> TSchemaPath:
@@ -127,6 +128,7 @@ class SchemaPath(AccessorPath[SchemaNode, SchemaKey, SchemaValue]):
             specification=specification,
             base_uri=base_uri,
             handlers=handlers,
+            resolved_cache_maxsize=resolved_cache_maxsize,
         )
 
         return cls(accessor, *args, separator=separator)
@@ -135,19 +137,29 @@ class SchemaPath(AccessorPath[SchemaNode, SchemaKey, SchemaValue]):
     def from_path(
         cls: type[TSchemaPath],
         path: Path,
+        resolved_cache_maxsize: int = 0,
     ) -> TSchemaPath:
         reader = PathReader(path)
         data, base_uri = reader.read()
-        return cls.from_dict(data, base_uri=base_uri)
+        return cls.from_dict(
+            data,
+            base_uri=base_uri,
+            resolved_cache_maxsize=resolved_cache_maxsize,
+        )
 
     @classmethod
     def from_file_path(
         cls: type[TSchemaPath],
         file_path: str,
+        resolved_cache_maxsize: int = 0,
     ) -> TSchemaPath:
         reader = FilePathReader(file_path)
         data, base_uri = reader.read()
-        return cls.from_dict(data, base_uri=base_uri)
+        return cls.from_dict(
+            data,
+            base_uri=base_uri,
+            resolved_cache_maxsize=resolved_cache_maxsize,
+        )
 
     @classmethod
     def from_file(
@@ -155,10 +167,16 @@ class SchemaPath(AccessorPath[SchemaNode, SchemaKey, SchemaValue]):
         fileobj: SupportsRead,
         base_uri: str = "",
         spec_url: str | None = None,
+        resolved_cache_maxsize: int = 0,
     ) -> TSchemaPath:
         reader = FileReader(fileobj)
         data, _ = reader.read()
-        return cls.from_dict(data, base_uri=base_uri, spec_url=spec_url)
+        return cls.from_dict(
+            data,
+            base_uri=base_uri,
+            spec_url=spec_url,
+            resolved_cache_maxsize=resolved_cache_maxsize,
+        )
 
     def str_keys(self) -> Sequence[str]:
         keys = list(self.keys())

{jsonschema_path-0.4.2 → jsonschema_path-0.4.3}/pyproject.toml

@@ -19,7 +19,7 @@ ignore_missing_imports = true
 
 [tool.poetry]
 name = "jsonschema-path"
-version = "0.4.2"
+version = "0.4.3"
 description = "JSONSchema Spec with object-oriented paths"
 authors = ["Artur Maciag <maciag.artur@gmail.com>"]
 license = "Apache-2.0"
@@ -93,7 +93,7 @@ message_template = "Version {new_version}"
 tag_template = "{new_version}"
 
 [tool.tbump.version]
-current = "0.4.2"
+current = "0.4.3"
 regex = '''
   (?P<major>\d+)
   \.