global: add class based registry to extend idutils schemes

idutils/config.py

@@ -0,0 +1,80 @@
+# -*- coding: utf-8 -*-
+#
+# This file is part of IDUtils
+# Copyright (C) 2024 CERN.
+#
+# IDUtils is free software; you can redistribute it and/or modify
+# it under the terms of the Revised BSD License; see LICENSE file for
+# more details.
+#
+# In applying this license, CERN does not waive the privileges and immunities
+# granted to it by virtue of its status as an Intergovernmental Organization
+# or submit itself to any jurisdiction.
+
+"""Functions for detecting the persistent identifier."""
+
+from . import validators
+
+IDUTILS_PID_SCHEMES_CONFIG = [
+    ("doi", validators.is_doi),
+    ("ark", validators.is_ark),
+    ("handle", validators.is_handle),
+    ("purl", validators.is_purl),
+    ("lsid", validators.is_lsid),
+    ("urn", validators.is_urn),
+    ("ads", validators.is_ads),
+    ("arxiv", validators.is_arxiv),
+    ("ascl", validators.is_ascl),
+    ("hal", validators.is_hal),
+    ("pmcid", validators.is_pmcid),
+    ("isbn", validators.is_isbn),
+    ("issn", validators.is_issn),
+    ("orcid", validators.is_orcid),
+    ("isni", validators.is_isni),
+    ("ean13", validators.is_ean13),
+    ("ean8", validators.is_ean8),
+    ("istc", validators.is_istc),
+    ("gnd", validators.is_gnd),
+    ("ror", validators.is_ror),
+    ("pmid", validators.is_pmid),
+    ("url", validators.is_url),
+    ("sra", validators.is_sra),
+    ("bioproject", validators.is_bioproject),
+    ("biosample", validators.is_biosample),
+    ("ensembl", validators.is_ensembl),
+    ("uniprot", validators.is_uniprot),
+    ("refseq", validators.is_refseq),
+    ("genome", validators.is_genome),
+    ("geo", validators.is_geo),
+    ("arrayexpress_array", validators.is_arrayexpress_array),
+    ("arrayexpress_experiment", validators.is_arrayexpress_experiment),
+    ("swh", validators.is_swh),
+    ("viaf", validators.is_viaf),
+]
+"""Definition of scheme name and associated test function.
+
+Order of list is important, as identifier scheme detection will test in the
+order given by this list."""
+
+
+IDUTILS_SCHEME_FILTER_CONFIG = [
+    (
+        "url",
+        # None these can have URLs, in which case we exclude them
+        ["isbn", "istc", "urn", "lsid", "issn", "ean8", "viaf"],
+    ),
+    ("ean8", ["gnd", "pmid", "viaf"]),
+    ("ean13", ["gnd", "pmid"]),
+    ("isbn", ["gnd", "pmid"]),
+    ("orcid", ["gnd", "pmid"]),
+    ("isni", ["gnd", "pmid"]),
+    (
+        "issn",
+        [
+            "gnd",
+            "viaf",
+        ],
+    ),
+    ("pmid", ["viaf"]),
+]
+"""(present_scheme, [list of schemes to remove if present_scheme found])."""

idutils/detectors.py

@@ -14,70 +14,17 @@
 """Functions for detecting the persistent identifier."""
 
 from . import validators
-from .proxies import current_idutils
+from .config import IDUTILS_PID_SCHEMES_CONFIG, IDUTILS_SCHEME_FILTER_CONFIG
+from .proxies import custom_schemes_registry
 
-IDUTILS_PID_SCHEMES = [
-    ("doi", validators.is_doi),
-    ("ark", validators.is_ark),
-    ("handle", validators.is_handle),
-    ("purl", validators.is_purl),
-    ("lsid", validators.is_lsid),
-    ("urn", validators.is_urn),
-    ("ads", validators.is_ads),
-    ("arxiv", validators.is_arxiv),
-    ("ascl", validators.is_ascl),
-    ("hal", validators.is_hal),
-    ("pmcid", validators.is_pmcid),
-    ("isbn", validators.is_isbn),
-    ("issn", validators.is_issn),
-    ("orcid", validators.is_orcid),
-    ("isni", validators.is_isni),
-    ("ean13", validators.is_ean13),
-    ("ean8", validators.is_ean8),
-    ("istc", validators.is_istc),
-    ("gnd", validators.is_gnd),
-    ("ror", validators.is_ror),
-    ("pmid", validators.is_pmid),
-    ("url", validators.is_url),
-    ("sra", validators.is_sra),
-    ("bioproject", validators.is_bioproject),
-    ("biosample", validators.is_biosample),
-    ("ensembl", validators.is_ensembl),
-    ("uniprot", validators.is_uniprot),
-    ("refseq", validators.is_refseq),
-    ("genome", validators.is_genome),
-    ("geo", validators.is_geo),
-    ("arrayexpress_array", validators.is_arrayexpress_array),
-    ("arrayexpress_experiment", validators.is_arrayexpress_experiment),
-    ("swh", validators.is_swh),
-    ("viaf", validators.is_viaf),
-]
+IDUTILS_PID_SCHEMES = IDUTILS_PID_SCHEMES_CONFIG
 """Definition of scheme name and associated test function.
 
 Order of list is important, as identifier scheme detection will test in the
 order given by this list."""
 
 
-IDUTILS_SCHEME_FILTER = [
-    (
-        "url",
-        # None these can have URLs, in which case we exclude them
-        ["isbn", "istc", "urn", "lsid", "issn", "ean8", "viaf"],
-    ),
-    ("ean8", ["gnd", "pmid", "viaf"]),
-    ("ean13", ["gnd", "pmid"]),
-    ("isbn", ["gnd", "pmid"]),
-    ("orcid", ["gnd", "pmid"]),
-    ("isni", ["gnd", "pmid"]),
-    (
-        "issn",
-        [
-            "gnd",
-            "viaf",
-        ],
-    ),
-    ("pmid", ["viaf"]),
-]
+IDUTILS_SCHEME_FILTER = IDUTILS_SCHEME_FILTER_CONFIG
 """(present_scheme, [list of schemes to remove if present_scheme found])."""
 
 
@@ -87,7 +34,7 @@ def detect_identifier_schemes(val):
     .. note:: Some schemes like PMID are very generic.
     """
     schemes = []
-    scheme_validators = IDUTILS_PID_SCHEMES + current_idutils.pick_scheme_key(
+    scheme_validators = IDUTILS_PID_SCHEMES + custom_schemes_registry().pick_scheme_key(
         "validator"
     )
     for scheme, test in scheme_validators:
@@ -111,7 +58,9 @@ def detect_identifier_schemes(val):
             if val.startswith(viaf_url):
                 schemes.remove("handle")
 
-    scheme_filter = IDUTILS_SCHEME_FILTER + current_idutils.pick_scheme_key("filter")
+    scheme_filter = IDUTILS_SCHEME_FILTER + custom_schemes_registry().pick_scheme_key(
+        "filter"
+    )
     for first, remove_schemes in scheme_filter:
         if first in schemes:
             schemes = list(filter(lambda x: x not in remove_schemes, schemes))

idutils/ext.py

@@ -13,10 +13,11 @@
 
 """Invenio IDUtils module for managing persistent identifiers used in scholarly communication."""
 
+from threading import Lock
+
 from importlib_metadata import entry_points
 
-from .detectors import IDUTILS_PID_SCHEMES
-from .proxies import current_idutils
+from .config import IDUTILS_PID_SCHEMES_CONFIG
 
 
 def _set_default_custom_scheme_config(scheme_config):
@@ -33,21 +34,26 @@ def _set_default_custom_scheme_config(scheme_config):
         scheme_key in default_config.keys() for scheme_key in scheme_config.keys()
     )
 
+    # Merge the provided scheme config with defaults
     return {**default_config, **scheme_config}
 
 
-class IDUtils(object):
-    """Invenio extension."""
+class CustomSchemesRegistry:
+    """Singleton class for loading and storing custom schemes from entry points."""
 
-    def __init__(self, app=None):
-        """Extension initialization."""
-        if app:
-            self.init_app(app)
+    _instance = None
+    _lock = Lock()  # To ensure thread-safe singleton creation
 
-    def init_app(self, app):
-        """Flask application initialiation."""
-        self.init_idutils_registry()
-        app.extensions["idutils"] = self
+    def __new__(cls):
+        """Create a new instance."""
+        with cls._lock:
+            if cls._instance is None:
+                cls._instance = super(CustomSchemesRegistry, cls).__new__(cls)
+                cls._instance._custom_schemes_registry = (
+                    {}
+                )  # Internal dictionary to store schemes
+                cls._instance._load_entry_points("idutils.custom_schemes")
+        return cls._instance
 
     @property
     def custom_schemes(self):
@@ -66,10 +72,8 @@ def custom_schemes(self):
                 "url_generator": lambda scheme, normalized_pid: "normalized_url",
             }
         }
-
-        See examples in `idutils.validators` file.
         """
-        return self._custom_schemes
+        return self._custom_schemes_registry
 
     def pick_scheme_key(self, key):
         """Serialize the registered custom registered schemes by key.
@@ -78,40 +82,26 @@ def pick_scheme_key(self, key):
         """
         return [(scheme, config[key]) for scheme, config in self.custom_schemes.items()]
 
-    def init_idutils_registry(self):
-        """Initialize custom schemes registries."""
-        self._custom_schemes = {}
-        self._load_entry_point(
-            self._custom_schemes,
-            "idutils.custom_schemes",
-        )
-
-    def _load_entry_point(self, registry, ep_name):
-        """Load entry points inton the given registry."""
-        existing_id_names = set(scheme[0] for scheme in IDUTILS_PID_SCHEMES)
+    def _load_entry_points(self, ep_name):
+        """Load entry points into the internal registry."""
+        existing_id_names = set(scheme[0] for scheme in IDUTILS_PID_SCHEMES_CONFIG)
+
+        # Load entry points from the specified group
         for ep in set(entry_points(group=ep_name)):
             name = ep.name
-            # Assert that the custom scheme is not overriding any existing scheme
-            assert name not in existing_id_names
 
+            # Ensure no custom scheme overrides existing ones
+            assert name not in existing_id_names, f"Scheme {name} already exists!"
+
+            # Load the function from entry point
             scheme_register_func = ep.load()
-            assert callable(scheme_register_func)
+            assert callable(scheme_register_func), f"{name} must be callable!"
 
+            # Call the function to get the scheme config
             scheme_config = scheme_register_func()
-            scheme_config = _set_default_custom_scheme_config(scheme_config)
-            registry.setdefault(name, scheme_config)
-
-
-def finalize_app(app):
-    """Finalize app."""
-    init(app)
-
-
-def api_finalize_app(app):
-    """Finalize app."""
-    init(app)
 
+            # Set default config values if needed
+            scheme_config = _set_default_custom_scheme_config(scheme_config)
 
-def init(app):
-    """Init app."""
-    ext = app.extensions["idutils"]
+            # Store in the registry
+            self._custom_schemes_registry.setdefault(name, scheme_config)

idutils/normalizers.py

@@ -15,7 +15,7 @@
 
 import isbnlib
 
-from .proxies import current_idutils
+from .proxies import custom_schemes_registry
 from .utils import *
 from .validators import is_arxiv_post_2007, is_arxiv_pre_2007
 
@@ -172,7 +172,9 @@ def normalize_pid(val, scheme):
     elif scheme == "viaf":
         return normalize_viaf(val)
     else:
-        for custom_scheme, normalizer in current_idutils.pick_scheme_key("normalizer"):
+        for custom_scheme, normalizer in custom_schemes_registry().pick_scheme_key(
+            "normalizer"
+        ):
             if scheme == custom_scheme:
                 return normalizer(val)
     return val
@@ -234,7 +236,7 @@ def to_url(val, scheme, url_scheme="http"):
     elif scheme in ["purl", "url"]:
         return pid
     else:
-        for custom_scheme, url_generator in current_idutils.pick_scheme_key(
+        for custom_scheme, url_generator in custom_schemes_registry().pick_scheme_key(
             "url_generator"
         ):
             if scheme == custom_scheme:

idutils/proxies.py

@@ -13,8 +13,7 @@
 
 """Proxy definitions."""
 
-from flask import current_app
-from werkzeug.local import LocalProxy
+from .ext import CustomSchemesRegistry
 
-current_idutils = LocalProxy(lambda: current_app.extensions["idutils"])
-"""Proxy to the extension."""
+custom_schemes_registry = lambda: CustomSchemesRegistry()
+"""Proxy to the custom scheme registrty."""

setup.cfg

@@ -37,7 +37,6 @@ install_requires =
 
 [options.extras_require]
 tests =
-    invenio-app>=1.4.0
     pytest-black-ng>=0.4.0
     pytest-cache>=1.0
     pytest-runner>=2.6.2

tests/conftest.py

@@ -14,20 +14,6 @@
 """Pytest configuration."""
 
 import pytest
-from invenio_app.factory import create_api
-
-
-@pytest.fixture(scope="module")
-def create_app(entry_points):
-    """Application factory fixture."""
-    return create_api
-
-
-@pytest.fixture(scope="module")
-def base_app(base_app):
-    """Application factory fixture."""
-    with base_app.app_context():
-        yield base_app
 
 
 @pytest.fixture(scope="module")

tests/test_custom_scheme_registry.py

@@ -0,0 +1,27 @@
+# -*- coding: utf-8 -*-
+#
+# This file is part of IDUtils
+# Copyright (C) 2024 CERN.
+#
+# IDUtils is free software; you can redistribute it and/or modify
+# it under the terms of the Revised BSD License; see LICENSE file for
+# more details.
+#
+# In applying this license, CERN does not waive the privileges and immunities
+# granted to it by virtue of its status as an Intergovernmental Organization
+# or submit itself to any jurisdiction.
+
+"""Persistent identifier utilities tests."""
+
+import pytest
+
+from idutils.proxies import custom_schemes_registry
+
+
+def test_custom_registry_singleton(entry_points):
+    """Test that the registry is instantiated only once."""
+    instance1 = custom_schemes_registry()
+
+    instance2 = custom_schemes_registry()
+
+    assert instance1 is instance2