Merge pull request #195 from tschaub/schema-caching

Cache parsed schema

Author: jonhealy1

CHANGELOG.md

@@ -4,6 +4,11 @@ All notable changes to this project will be documented in this file.
 
 The format is (loosely) based on [Keep a Changelog](http://keepachangelog.com/) and this project adheres to [Semantic Versioning](http://semver.org/).
 
+## [v2.4.3] - 2022-03-10
+### Changed
+
+ - Add schema caching
+
 ## [v2.4.2] - 2022-03-02
 ### Changed
 

setup.py

@@ -2,7 +2,7 @@
 
 from setuptools import setup
 
-__version__ = "2.4.2"
+__version__ = "2.4.3"
 
 with open("README.md", "r") as fh:
     long_description = fh.read()
@@ -28,8 +28,8 @@
     keywords="STAC validation raster",
     long_description=long_description,
     long_description_content_type="text/markdown",
-    url="https://github.com/sparkgeo/stac-validator",
-    download_url="https://github.com/sparkgeo/stac-validator/archive/v2.3.0.tar.gz",
+    url="https://github.com/stac-utils/stac-validator",
+    download_url="https://github.com/stac-utils/stac-validator/archive/v2.4.3.tar.gz",
     install_requires=[
         "requests>=2.19.1",
         "jsonschema>=3.2.0",

stac_validator/utilities.py

@@ -1,3 +1,4 @@
+import functools
 import json
 from urllib.parse import urlparse
 from urllib.request import urlopen
@@ -56,6 +57,11 @@ def fetch_and_parse_file(input_path) -> dict:
     return data
 
 
+@functools.lru_cache(maxsize=48)
+def fetch_and_parse_schema(input_path) -> dict:
+    return fetch_and_parse_file(input_path)
+
+
 # validate new versions at schemas.stacspec.org
 def set_schema_addr(version, stac_type: str):
     if version in NEW_VERSIONS:

stac_validator/validate.py

@@ -11,6 +11,7 @@
 
 from .utilities import (
     fetch_and_parse_file,
+    fetch_and_parse_schema,
     get_stac_type,
     link_request,
     set_schema_addr,
@@ -150,15 +151,15 @@ def extensions_validator(self, stac_type: str) -> dict:
     def custom_validator(self):
         # in case the path to custom json schema is local
         # it may contain relative references
-        schema = fetch_and_parse_file(self.custom)
+        schema = fetch_and_parse_schema(self.custom)
         if os.path.exists(self.custom):
             custom_abspath = os.path.abspath(self.custom)
             custom_dir = os.path.dirname(custom_abspath).replace("\\", "/")
             custom_uri = f"file:///{custom_dir}/"
             resolver = RefResolver(custom_uri, self.custom)
             jsonschema.validate(self.stac_content, schema, resolver=resolver)
         else:
-            schema = fetch_and_parse_file(self.custom)
+            schema = fetch_and_parse_schema(self.custom)
             jsonschema.validate(self.stac_content, schema)
 
     def core_validator(self, stac_type: str):
@@ -233,7 +234,7 @@ def recursive_validator(self, stac_type: str):
                     self.custom = set_schema_addr(self.version, stac_type.lower())
                     message = self.create_message(stac_type, "recursive")
                     if self.version == "0.7.0":
-                        schema = fetch_and_parse_file(self.custom)
+                        schema = fetch_and_parse_schema(self.custom)
                         # this next line prevents this: unknown url type: 'geojson.json' ??
                         schema["allOf"] = [{}]
                         jsonschema.validate(self.stac_content, schema)