Extend url2purl/purl2url coverage for Git-based source hosts

src/packageurl/contrib/purl2url.py

@@ -27,6 +27,7 @@
 from packageurl import PackageURL
 from packageurl.contrib.route import NoRouteAvailable
 from packageurl.contrib.route import Router
+import re
 
 DEFAULT_MAVEN_REPOSITORY = "https://repo.maven.apache.org/maven2"
 
@@ -172,6 +173,26 @@ def build_gitlab_repo_url(purl):
         return f"https://gitlab.com/{namespace}/{name}"
 
 
+@repo_router.route("pkg:generic/.*")
+def build_generic_repo_url(purl):
+    """
+    Return a Generic repo URL from the `purl` string.
+    """
+    purl_data = PackageURL.from_string(purl)
+    name = purl_data.name
+    namespace = purl_data.namespace
+    version = purl_data.version
+
+    if namespace.startswith("git.kernel.org"):
+        return f"https://{namespace}/{name}/commit/?id={version}"
+
+    if re.match(r"^(android|aomedia)\.googlesource\.com(/.*)?$", namespace):
+        return f"https://{namespace}/{name}/+/{version}"
+
+    if namespace.startswith("sourceforge.net"):
+        return f"https://{namespace}/{name}/ci/{version}"
+
+
 @repo_router.route("pkg:(gem|rubygems)/.*")
 def build_rubygems_repo_url(purl):
     """

src/packageurl/contrib/url2purl.py

@@ -667,6 +667,94 @@ def build_bitbucket_purl(url):
     )
 
 
+@purl_router.route("https?://git\.kernel\\.org/.*")
+def build_kernel_purl(url):
+    """
+    Return a PackageURL object from Kernel `url`.
+    For example:
+    https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?id=74770b1fd2be612f9c2cf807db81fcdcc35e6560
+    """
+
+    kernel_project_pattern = (
+        r"^https?://git\.kernel\.org/"
+        r"(?P<namespace>.+)/"
+        r"(?P<name>[^/]+?)"
+        r"(?:\.git)?"
+        r"/commit/\?id="
+        r"(?P<version>[0-9a-fA-F]{7,64})/?$"
+    )
+
+    commit_matche = re.search(kernel_project_pattern, url)
+    if commit_matche:
+        namespace = "git.kernel.org/" + commit_matche.group("namespace")
+        return PackageURL(
+            type="generic",
+            namespace=namespace,
+            name=commit_matche.group("name"),
+            version=commit_matche.group("version"),
+            qualifiers={},
+            subpath="",
+        )
+
+
+@purl_router.route(r"https?://(android|aomedia)\.googlesource\.com/.*")
+def build_googlesource_purl(url):
+    """
+    Return a PackageURL object from Android `url`.
+    For example:
+    https://android.googlesource.com/platform/packages/apps/Settings/+/2968ccc911956fa5813a9a6a5e5c8970e383a60f
+    https://aomedia.googlesource.com/libavifinfo/+/43716e9c34d3389b4882fbd1a81c04543ed04fe3
+    """
+
+    commit_pattern = (
+        r"^https?://(?P<namespace>(android|aomedia)\.googlesource\.com(?:/.*)?)"
+        r"/(?P<name>[^/]+)"
+        r"/\+/"
+        r"(?P<version>[0-9a-fA-F]{7,64})"
+    )
+
+    match = re.search(commit_pattern, url)
+    if match:
+        return PackageURL(
+            type="generic",
+            namespace=match.group("namespace"),
+            name=match.group("name"),
+            version=match.group("version"),
+            qualifiers={},
+            subpath="",
+        )
+
+
+@purl_router.route("https?://sourceforge\\.net/p/.*")
+def build_sourceforge_purl(url):
+    """
+    Return a PackageURL object from sourceforge `url`.
+    For example:
+    https://sourceforge.net/p/djvu/djvulibre-git/ci/e15d51510048927f172f1bf1f27ede65907d940d
+    https://sourceforge.net/p/infrarecorder/code/ci/9361b6f267e7b1c1576c48f6dac6dec18d8a93e0/
+    """
+
+    sourceforge_pattern = (
+        r"^https?://sourceforge\.net/"
+        r"(?P<namespace>.+)/"
+        r"(?P<name>[^/]+?)"
+        r"/ci/"
+        r"(?P<version>[0-9a-fA-F]{7,64})/?$"
+    )
+
+    commit_matche = re.search(sourceforge_pattern, url)
+    if commit_matche:
+        namespace = "sourceforge.net/" + commit_matche.group("namespace")
+        return PackageURL(
+            type="generic",
+            namespace=namespace,
+            name=commit_matche.group("name"),
+            version=commit_matche.group("version"),
+            qualifiers={},
+            subpath="",
+        )
+
+
 @purl_router.route("https?://gitlab\\.com/(?!.*/archive/).*")
 def build_gitlab_purl(url):
     """

tests/contrib/data/url2purl.json

@@ -277,5 +277,10 @@
   "https://packagemanager.rstudio.com/cran/2022-06-23/src/contrib/curl_4.3.2.tar.gz": "pkg:cran/curl@4.3.2?download_url=https://packagemanager.rstudio.com/cran/2022-06-23/src/contrib/curl_4.3.2.tar.gz",
   "https://github.com/TG1999/first_repo/commit/98e516011d6e096e25247b82fc5f196bbeecff10": "pkg:github/tg1999/first_repo@98e516011d6e096e25247b82fc5f196bbeecff10",
   "https://gitlab.com/TG1999/first_repo/-/commit/bf04e5f289885cf2f20a92b387bcc6df33e30809": "pkg:gitlab/tg1999/first_repo@bf04e5f289885cf2f20a92b387bcc6df33e30809",
-  "https://bitbucket.org/TG1999/first_repo/commits/16a60c4a74ef477cd8c16ca82442eaab2fbe8c86": "pkg:bitbucket/tg1999/first_repo@16a60c4a74ef477cd8c16ca82442eaab2fbe8c86"
+  "https://git.kernel.org/pub/scm/utils/b4/b4.git/commit/?id=477734000555ffc24bf873952e40367deee26f17": "pkg:generic/git.kernel.org/pub/scm/utils/b4/b4@477734000555ffc24bf873952e40367deee26f17",
+  "https://git.kernel.org/pub/scm/docs/kernel/ksmap.git/commit/?id=e8c7bac5321ba31d63496bd7fecea3db1848e355": "pkg:generic/git.kernel.org/pub/scm/docs/kernel/ksmap@e8c7bac5321ba31d63496bd7fecea3db1848e355",
+  "https://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git/commit/?id=7457fe9541b5162f285454947448d553a5d5a531": "pkg:generic/git.kernel.org/pub/scm/virt/kvm/mst/qemu@7457fe9541b5162f285454947448d553a5d5a531",
+  "https://android.googlesource.com/platform/frameworks/base/+/b4da73a935a8c906ff5df562155824d63ac849ab": "pkg:generic/android.googlesource.com/platform/frameworks/base@b4da73a935a8c906ff5df562155824d63ac849ab",
+  "https://android.googlesource.com/device/generic/vulkan-cereal/+/240dedcb0fa917b3d2dcc4a9d4c332697c5e48a0": "pkg:generic/android.googlesource.com/device/generic/vulkan-cereal@240dedcb0fa917b3d2dcc4a9d4c332697c5e48a0",
+  "https://sourceforge.net/p/djvu/djvulibre-git/ci/e15d51510048927f172f1bf1f27ede65907d940d": "pkg:generic/sourceforge.net/p/djvu/djvulibre-git@e15d51510048927f172f1bf1f27ede65907d940d"
 }

tests/contrib/test_purl2url.py

@@ -70,6 +70,11 @@ def test_purl2url_get_repo_url():
         "pkg:cocoapods/MapsIndoors@3.24.0": "https://cocoapods.org/pods/MapsIndoors",
         "pkg:maven/org.apache.commons/commons-io@1.3.2": "https://repo.maven.apache.org/maven2/org/apache/commons/commons-io/1.3.2",
         "pkg:maven/org.apache.commons/commons-io@1.3.2?repository_url=https://repo1.maven.org/maven2": "https://repo1.maven.org/maven2/org/apache/commons/commons-io/1.3.2",
+        "pkg:generic/git.kernel.org/pub/scm/libs/liba2i/liba2i.git@4fc8196d7811c26abefaf3a3ae6b5c67c4c9cbc9": "https://git.kernel.org/pub/scm/libs/liba2i/liba2i.git/commit/?id=4fc8196d7811c26abefaf3a3ae6b5c67c4c9cbc9",
+        "pkg:generic/git.kernel.org/pub/scm/linux/kernel/git/a.hindborg/configfs.git@bc3372351d0c8b2726b7d4229b878342e3e6b0e8": "https://git.kernel.org/pub/scm/linux/kernel/git/a.hindborg/configfs.git/commit/?id=bc3372351d0c8b2726b7d4229b878342e3e6b0e8",
+        "pkg:generic/android.googlesource.com/accessories/manifest@9ad7ef740dc39834a88bf95c69f35f18b8f45543": "https://android.googlesource.com/accessories/manifest/+/9ad7ef740dc39834a88bf95c69f35f18b8f45543",
+        "pkg:generic/aomedia.googlesource.com/libavifinfo@43716e9c34d3389b4882fbd1a81c04543ed04fe3": "https://aomedia.googlesource.com/libavifinfo/+/43716e9c34d3389b4882fbd1a81c04543ed04fe3",
+        "pkg:generic/sourceforge.net/p/infrarecorder/code@8fab704119ff23691f075f6a281521b6c7d7e55f": "https://sourceforge.net/p/infrarecorder/code/ci/8fab704119ff23691f075f6a281521b6c7d7e55f",
     }
 
     for purl, url in purls_url.items():