no disk caching (#20)

* better caching

* no disk caching at all

* update manifest

* remove gitignore

* fix network requests

* remove unused imports

* remove potential key error

* rename url

* change ns check

* undo path modification

.github/workflows/release.yml

@@ -21,12 +21,13 @@ jobs:
       - name: Install Dependencies
         run: |
           python -m pip install --upgrade pip
-          pip install -U setuptools setuptools_scm wheel
+          pip install -U setuptools setuptools_scm wheel check-manifest
           pip install -e .[autogen]
       - name: Build Distribution
         run: |
           TAG="${GITHUB_REF/refs\/tags\/v/}"
           echo "::set-env name=tag::$TAG"
+          check-manifest
           python setup.py sdist bdist_wheel
       - name: Create Release
         id: create_release

.gitignore

@@ -108,5 +108,4 @@ venv.bak/
 .DS_Store
 _test_data
 src/ome_types/model/
-src/ome_types/*.pkl
-src/ome_types/_version.py
+src/ome_types/_version.py

MANIFEST.in

@@ -1 +1,3 @@
-include LICENSE README.md setup.py pyproject.toml
+include LICENSE README.md setup.py pyproject.toml
+recursive-include src *.xsd
+recursive-exclude testing *

pyproject.toml

@@ -1,3 +1,10 @@
 [build-system]
 requires = ["setuptools>=42", "wheel", "setuptools_scm[toml]>=3.4", "black", "isort<5.0", "xmlschema", "pydantic"]
 build-backend = "setuptools.build_meta"
+
+[tool.check-manifest]
+ignore = [
+    "*src/ome_types/model/*",
+    "mypy.ini",
+    "src/ome_types/_version.py",  # added by setuptools_scm during build
+]

setup.cfg

@@ -31,6 +31,9 @@ install_requires =
     pydantic[email]>=1.0
     xmlschema>=1.0.16
 
+[options.package_data]
+* = *.xsd
+
 [options.extras_require]
 autogen = isort<5.0; black
 

src/ome_types/schema.py

@@ -1,16 +1,16 @@
-import pickle
 import re
-from os.path import dirname, exists, join
+from functools import lru_cache
+from pathlib import Path
+from typing import Any, Dict, Optional, Union
 from xml.etree import ElementTree
-from typing import Any, Dict, Optional
 
 import xmlschema
 from xmlschema.converters import XMLSchemaConverter
 
 from .model import _field_plurals
 
-
-NS_OME = "{http://www.openmicroscopy.org/Schemas/OME/2016-06}"
+URI_OME = "http://www.openmicroscopy.org/Schemas/OME/2016-06"
+NS_OME = "{" + URI_OME + "}"
 
 __cache__: Dict[str, xmlschema.XMLSchema] = {}
 
@@ -23,32 +23,44 @@ def camel_to_snake(name: str) -> str:
     return result
 
 
-def get_schema(xml: str) -> xmlschema.XMLSchema:
-    url = xmlschema.fetch_schema(xml)
-    version = (
-        re.split("(.com|.org)", url)[-1]
-        .replace("/", "_")
-        .lstrip("_")
-        .replace(".xsd", "")
-    )
-    if version not in __cache__:
-        local = join(dirname(__file__), f"{version}.pkl")
-        if exists(local):
-            with open(local, "rb") as f:
-                __cache__[version] = pickle.load(f)
-        else:
-            schema = xmlschema.XMLSchema(url)
-
-            # FIXME Hack to work around xmlschema poor support for keyrefs to
-            # substitution groups
-            ls_sgs = schema.maps.substitution_groups[f"{NS_OME}LightSourceGroup"]
-            ls_id_maps = schema.maps.identities[f"{NS_OME}LightSourceIDKey"]
-            ls_id_maps.elements = {e: None for e in ls_sgs}
-
-            __cache__[version] = schema
-            with open(local, "wb") as f:
-                pickle.dump(__cache__[version], f)
-    return __cache__[version]
+@lru_cache(maxsize=8)
+def _build_schema(namespace: str) -> xmlschema.XMLSchema:
+    """Return Schema object for a url.
+
+    For the special case of retrieving the 2016-06 OME Schema, use local file.
+    """
+    if namespace == URI_OME:
+        schema = xmlschema.XMLSchema(str(Path(__file__).parent / "ome-2016-06.xsd"))
+        # FIXME Hack to work around xmlschema poor support for keyrefs to
+        # substitution groups
+        ls_sgs = schema.maps.substitution_groups[f"{NS_OME}LightSourceGroup"]
+        ls_id_maps = schema.maps.identities[f"{NS_OME}LightSourceIDKey"]
+        ls_id_maps.elements = {e: None for e in ls_sgs}
+    else:
+        schema = xmlschema.XMLSchema(namespace)
+    return schema
+
+
+def get_schema(source: Union[xmlschema.XMLResource, str]) -> xmlschema.XMLSchema:
+    """Fetch an XMLSchema object given XML source.
+
+    Parameters
+    ----------
+    source : XMLResource or str
+        can be an :class:`xmlschema.XMLResource` instance, a file-like object, a path
+        to a file or an URI of a resource or an Element instance or an ElementTree
+        instance or a string containing the XML data.
+
+    Returns
+    -------
+    xmlschema.XMLSchema
+        An XMLSchema object for the source
+    """
+    if not isinstance(source, xmlschema.XMLResource):
+        resource = xmlschema.XMLResource(source)
+    else:
+        resource = source
+    return _build_schema(resource.namespace)
 
 
 def validate(xml: str, schema: Optional[xmlschema.XMLSchema] = None) -> None:

testing/test_autogen.py

@@ -1,14 +1,13 @@
 import importlib
-import os
 import sys
-from glob import glob
 from pathlib import Path
 
 import pytest
 from xmlschema.validators.exceptions import XMLSchemaValidationError
 
 TESTING_DIR = Path(__file__).parent
-SRC_MODEL = TESTING_DIR.parent / "src" / "ome_types" / "model"
+SRC = TESTING_DIR.parent / "src" / "ome_types"
+SRC_MODEL = SRC / "model"
 
 
 @pytest.fixture(scope="session")
@@ -18,15 +17,18 @@ def model(tmp_path_factory, request):
             raise RuntimeError(
                 f"Please generate local {SRC_MODEL} before using --nogen"
             )
-        sys.path.insert(0, str(SRC_MODEL.parent))
-        return importlib.import_module(SRC_MODEL.name)
-    from ome_autogen import convert_schema
-
-    target_dir = tmp_path_factory.mktemp("test_model")
-    xsd = TESTING_DIR / "ome-2016-06.xsd"
-    convert_schema(url=xsd, target_dir=target_dir)
-    sys.path.insert(0, str(target_dir.parent))
-    return importlib.import_module(target_dir.name)
+        sys.path.insert(0, str(SRC))
+        model = importlib.import_module(SRC_MODEL.name)
+    else:
+        from ome_autogen import convert_schema
+
+        target_dir = tmp_path_factory.mktemp("test_model")
+        xsd = SRC / "ome-2016-06.xsd"
+        convert_schema(url=xsd, target_dir=target_dir)
+        sys.path.insert(0, str(target_dir.parent))
+        model = importlib.import_module(target_dir.name)
+    sys.path.pop(0)
+    return model
 
 
 SHOULD_FAIL = {