Add tests for purl hash

Signed-off-by: Keshav Priyadarshi <[email protected]>

aboutcode/hashid/__init__.py

@@ -72,7 +72,7 @@ def get_vcid_yml_file_path(vcid: str):
     return Path(VULNERABILITY_REPO_NAME) / vulnerability_yml_path(vcid)
 
 
-# This cuxstom 32 characters alphabet is designed to avoid visually easily confusable characters:
+# This custom 32 characters alphabet is designed to avoid visually easily confusable characters:
 # i and l
 # 0 and o
 _base32_alphabet = b"abcdefghjkmnpqrstuvwxyz123456789"
@@ -164,7 +164,7 @@ def get_package_vulnerabilities_yml_file_path(purl: Union[PackageURL, str]):
 
 # We use a 4-tier system for storing package metadata.
 # The tiers are as follows:
-# 1. Super Large Ecosystem (~5M packages): 2^10 = 1,028 git repositories
+# 1. Super Large Ecosystem (~5M packages): 2^10 = 1,024 git repositories
 # 2. Large Ecosystem (~500K packages): 2^7 = 128 git repositories
 # 3. Medium Ecosystem (~50K packages): 2^5 = 32 git repositories
 # 4. Small Ecosystem (~2K packages): 2^0 = 1 git repository
@@ -188,6 +188,7 @@ def get_package_vulnerabilities_yml_file_path(purl: Union[PackageURL, str]):
     "composer": 5,
     "deb": 5,
     "docker": 5,
+    "gem": 5,
     "generic": 5,
     "huggingface": 5,
     "mlflow": 5,
@@ -200,7 +201,6 @@ def get_package_vulnerabilities_yml_file_path(purl: Union[PackageURL, str]):
     "conda": 0,
     "cpan": 0,
     "cran": 0,
-    "gem": 0,
     "hackage": 0,
     "hex": 0,
     "luarocks": 0,

aboutcode/hashid/tests/test_hashid.py

@@ -0,0 +1,54 @@
+#
+# Copyright (c) nexB Inc. and others. All rights reserved.
+# Portions Copyright (c) The Python Software Foundation
+# VulnerableCode is a trademark of nexB Inc.
+# SPDX-License-Identifier: Apache-2.0 and Python-2.0
+# See http://www.apache.org/licenses/LICENSE-2.0 for the license text.
+# See https://github.com/nexB/vulnerablecode for support or download.
+# See https://aboutcode.org for more information about nexB OSS projects.
+#
+
+import pytest
+
+from aboutcode.hashid import package_path_elements
+
+
+@pytest.mark.parametrize(
+    "purl, purl_hash",
+    [
+        ("pkg:maven/org.apache.commons/io", "4f"),
+        ("pkg:GOLANG/google.golang.org/genproto@abcdedf#/googleapis/api/annotations/", "4a"),
+        ("pkg:golang/github.com/nats-io/nats-server/v2/[email protected]", "22"),
+        ("pkg:bitbucket/birKenfeld/pyGments-main@244fd47e07d1014f0aed9c", "03"),
+        ("pkg:github/Package-url/purl-Spec@244fd47e07d1004f0aed9c", "095"),
+        ("pkg:deb/debian/[email protected]?arch=i386&distro=jessie", "19"),
+        (
+            "pkg:docker/customer/dockerimage@sha256:244fd47e07d1004f0aed9c?repository_url=gcr.io",
+            "10",
+        ),
+        ("pkg:gem/[email protected]?Platform=java", "1e"),
+        (
+            "pkg:Maven/org.apache.xmlgraphics/[email protected]?repositorY_url=repo.spring.io/release&classifier=sources",
+            "28",
+        ),
+        (
+            "pkg:Maven/org.apache.xmlgraphics/[email protected]?repositorY_url=repo.spring.io/release&extension=pom",
+            "28",
+        ),
+        ("pkg:Maven/net.sf.jacob-project/[email protected]?type=dll&classifier=x86", "17"),
+        ("pkg:npm/%40angular/[email protected]", "323"),
+        ("pkg:Nuget/[email protected]", "63"),
+        ("pkg:PYPI/[email protected]", "00"),
+        ("pkg:composer/guzzlehttp/[email protected]", "1d"),
+        ("pkg:Rpm/fedora/[email protected]?Arch=i386&Distro=fedora-25", "16"),
+        ("pkg:maven/HTTPClient/[email protected]", "4d"),
+        ("pkg:maven/mygroup/[email protected]%20Final?mykey=my%20value", "6f"),
+        ("pkg:npm/@babel/core#/googleapis/api/annotations/", "0dc"),
+        ("pkg:npm/@babel/[email protected]#/googleapis/api/annotations/", "0dc"),
+        ("pkg:npm/[email protected]#/googleapis/api/annotations/", "23b"),
+        ("pkg:npm/core#/googleapis/api/annotations/", "23b"),
+    ],
+)
+def test_purl_hash(purl, purl_hash):
+    result_hash, *_ = package_path_elements(purl)
+    assert result_hash == purl_hash

pyproject-aboutcode.hashid.toml

@@ -66,6 +66,7 @@ excludes = [
     "**/*.bak",
     "**/.ipynb_checkpoints",
     "aboutcode/hashid/python.LICENSE",
+    "aboutcode/hashid/tests/**/*",
 ]
 
 metadata_files = ["apache-2.0.LICENSE", "NOTICE",  "aboutcode/hashid/python.LICENSE"]