feat: complete SPDX license expression (#425)

Signed-off-by: Jan Kowalleck <[email protected]>

Author: jkowalleck

cyclonedx/spdx.py

@@ -14,11 +14,19 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
-__all__ = ['is_supported_id', 'fixup_id', 'is_compound_expression']
+__all__ = [
+    'is_supported_id', 'fixup_id',
+    'is_compound_expression'
+]
 
 from json import load as json_load
 from os.path import dirname, join as path_join
-from typing import Dict, Optional, Set
+from typing import TYPE_CHECKING, Dict, Optional, Set
+
+from license_expression import get_spdx_licensing  # type: ignore
+
+if TYPE_CHECKING:
+    from license_expression import Licensing
 
 # region init
 # python's internal module loader will assure that this init-part runs only once.
@@ -30,9 +38,11 @@
 
 __IDS_LOWER_MAP: Dict[str, str] = dict((id_.lower(), id_) for id_ in __IDS)
 
+__SPDX_EXPRESSION_LICENSING: 'Licensing' = get_spdx_licensing()
 
 # endregion
 
+
 def is_supported_id(value: str) -> bool:
     """Validate a SPDX-ID according to current spec."""
     return value in __IDS
@@ -50,11 +60,12 @@ def is_compound_expression(value: str) -> bool:
     """Validate compound expression.
 
     .. note::
-        Uses a best-effort detection of SPDX compound expression according to `SPDX license expression spec`_.
+        Utilizes `license-expression library`_ to
+        validate SPDX compound expression according to `SPDX license expression spec`_.
 
     .. _SPDX license expression spec: https://spdx.github.io/spdx-spec/v2.3/SPDX-license-expressions/
+    .. _license-expression library: https://github.com/nexB/license-expression
     """
-    # shortest known valid expression: (A or B) - 8 characters long
-    return len(value) >= 8 \
-        and value.startswith('(') \
-        and value.endswith(')')
+    return 0 == len(
+        __SPDX_EXPRESSION_LICENSING.validate(value).errors
+    )

poetry.lock

@@ -55,6 +55,7 @@ python = "^3.7"
 packageurl-python = ">= 0.11"
 py-serializable = "^0.11.1"
 sortedcontainers = "^2.4.0"
+license-expression = "^30"
 
 [tool.poetry.dev-dependencies]
 ddt = "^1.6.0"

pyproject.toml

@@ -33,8 +33,8 @@
 
 VALID_COMPOUND_EXPRESSIONS = {
     # for valid test data see the spec: https://spdx.github.io/spdx-spec/v2.3/SPDX-license-expressions/
-    '(MIT WITH Apache-2.0)',
-    '(BSD-2-Clause OR Apache-2.0)',
+    '(MIT AND Apache-2.0)',
+    'BSD-2-Clause OR Apache-2.0',
 }
 
 
@@ -89,8 +89,9 @@ def test_positive(self, valid_expression: str) -> None:
         self.assertTrue(actual)
 
     @data(
+        'MIT AND Apache-2.0 OR something-unknown'
         'something invalid',
-        '(c) John Doe'
+        '(c) John Doe',
     )
     def test_negative(self, invalid_expression: str) -> None:
         actual = spdx.is_compound_expression(invalid_expression)