Improve support for SourceForge URLs in url2purl #139

Signed-off-by: tdruez <tdruez@nexb.com>

Author: tdruez

CHANGELOG.rst

@@ -1,6 +1,12 @@
 Changelog
 =========
 
+0.13.3 (2024-01-04)
+-------------------
+
+- Improve support for SourceForge URLs in `url2purl`.
+  https://github.com/package-url/packageurl-python/issues/139
+
 0.13.2 (2024-01-04)
 -------------------
 

README.rst

@@ -121,7 +121,7 @@ Make a new release
     git tag -a $VERSION -m "Tag $VERSION"
     git push origin $VERSION
 
-- Review and publish the "draft" release created by the workflow at
+- Review the GitHub release created by the workflow at
   https://github.com/package-url/packageurl-python/releases
 
 .. |ci-tests| image:: https://github.com/package-url/packageurl-python/actions/workflows/ci.yml/badge.svg?branch=main

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = packageurl-python
-version = 0.13.2
+version = 0.13.3
 license = MIT
 description = A purl aka. Package URL parser and builder
 long_description = file:README.rst

src/packageurl/contrib/url2purl.py

@@ -344,7 +344,24 @@ def build_pypi_purl(uri):
 register_pattern("nuget", nuget_api_pattern)
 
 
-@purl_router.route("https?://.*sourceforge.net/projects?/.*")
+# https://sourceforge.net/projects/turbovnc/files/3.1/turbovnc-3.1.tar.gz/download
+# https://sourceforge.net/projects/scribus/files/scribus/1.6.0/scribus-1.6.0.tar.gz/download
+# https://sourceforge.net/projects/ventoy/files/v1.0.96/Ventoy%201.0.96%20release%20source%20code.tar.gz/download
+# https://sourceforge.net/projects/geoserver/files/GeoServer/2.23.4/geoserver-2.23.4-war.zip/download
+sourceforge_download_pattern = (
+    r"^https?://.*sourceforge.net/projects/"
+    r"(?P<name>.+)/"
+    r"files/"
+    r"(?i:(?P=name)/)?"  # optional case-insensitive name segment repeated
+    r"v?(?P<version>[0-9\.]+)/"  # version restricted to digits and dots
+    r"(?i:(?P=name)).*(?P=version).*"  # case-insensitive matching for {name}-{version}
+    r"(/download)$"  # ending with "/download"
+)
+
+register_pattern("sourceforge", sourceforge_download_pattern)
+
+
+@purl_router.route("https?://.*sourceforge.net/project/.*")
 def build_sourceforge_purl(uri):
     # We use a more general route pattern instead of using `sourceforge_pattern`
     # below by itself because we want to capture all sourceforge download URLs,
@@ -353,14 +370,13 @@ def build_sourceforge_purl(uri):
     # URL that we can't handle.
 
     # http://master.dl.sourceforge.net/project/libpng/zlib/1.2.3/zlib-1.2.3.tar.bz2
-    # https://sourceforge.net/projects/scribus/files/scribus/1.6.0/scribus-1.6.0.tar.gz/download
     sourceforge_pattern = (
         r"^https?://.*sourceforge.net/projects?/"
         r"(?P<namespace>([^/]+))/"  # do not allow more "/" segments
-        r"(files/)?"  # optional segment for "*/download" type URLs
+        r"(OldFiles/)?"
         r"(?P<name>.+)/"
-        r"(?P<version>[0-9\.]+)/"  # version restricted to digits and dots
-        r"(?P=name)-(?P=version).*"  # {name}-{version} repeated in the filename
+        r"(?P<version>[v0-9\.]+)/"  # version restricted to digits and dots
+        r"(?P=name).*(?P=version).*"  # {name}-{version} repeated in the filename
         r"[^/]$"  # not ending with "/"
     )
 

tests/contrib/data/url2purl.json

@@ -160,18 +160,21 @@
   "http://master.dl.sourceforge.net/project/zinnia/zinnia-win32/0.06/zinnia-win32-0.06.zip": "pkg:sourceforge/zinnia/zinnia-win32@0.06",
   "http://iweb.dl.sourceforge.net/project/findbugs/findbugs/1.3.4/findbugs-1.3.4.tar.gz/": "pkg:sourceforge/findbugs?download_url=http://iweb.dl.sourceforge.net/project/findbugs/findbugs/1.3.4/findbugs-1.3.4.tar.gz/",
   "http://master.dl.sourceforge.net/project/arestc/net/sf/arestc/arestc/0.1.4/arestc-0.1.4-javadoc.jar": "pkg:sourceforge/arestc?download_url=http://master.dl.sourceforge.net/project/arestc/net/sf/arestc/arestc/0.1.4/arestc-0.1.4-javadoc.jar",
-  "http://master.dl.sourceforge.net/project/intraperson/OldFiles/intraperson/0.28/intraperson-0.28.tar.gz": "pkg:sourceforge/intraperson?download_url=http://master.dl.sourceforge.net/project/intraperson/OldFiles/intraperson/0.28/intraperson-0.28.tar.gz",
+  "http://master.dl.sourceforge.net/project/intraperson/OldFiles/intraperson/0.28/intraperson-0.28.tar.gz": "pkg:sourceforge/intraperson/intraperson@0.28",
   "http://master.dl.sourceforge.net/project/pwiki/pwiki/0.1.2/0.1.2.zip": "pkg:sourceforge/pwiki?download_url=http://master.dl.sourceforge.net/project/pwiki/pwiki/0.1.2/0.1.2.zip",
   "http://master.dl.sourceforge.net/project/iswraid/iswraid/0.1.4.3/2.4.28-pre3-iswraid.patch.gz": "pkg:sourceforge/iswraid?download_url=http://master.dl.sourceforge.net/project/iswraid/iswraid/0.1.4.3/2.4.28-pre3-iswraid.patch.gz",
   "http://master.dl.sourceforge.net/project/aloyscore/aloyscore/0.1a1%20stable/0.1a1_stable_AloysCore.zip": "pkg:sourceforge/aloyscore?download_url=http://master.dl.sourceforge.net/project/aloyscore/aloyscore/0.1a1%2520stable/0.1a1_stable_AloysCore.zip",
   "http://master.dl.sourceforge.net/project/myenterprise/OldFiles/1.0.0.2.MyEnterprise.Source.zip": "pkg:sourceforge/myenterprise?download_url=http://master.dl.sourceforge.net/project/myenterprise/OldFiles/1.0.0.2.MyEnterprise.Source.zip",
   "http://master.dl.sourceforge.net/project/wxhaskell/wxhaskell/wxhaskell-0.9/wxhaskell-src-0.9.zip": "pkg:sourceforge/wxhaskell?download_url=http://master.dl.sourceforge.net/project/wxhaskell/wxhaskell/wxhaskell-0.9/wxhaskell-src-0.9.zip",
   "http://master.dl.sourceforge.net/project/a2freedom/A2/1.2/a2freedom-1.2.zip": "pkg:sourceforge/a2freedom?download_url=http://master.dl.sourceforge.net/project/a2freedom/A2/1.2/a2freedom-1.2.zip",
-  "http://master.dl.sourceforge.net/project/tinyos/OldFiles/tinyos/1.1.0/tinyos-1.1.0.tar.gz": "pkg:sourceforge/tinyos?download_url=http://master.dl.sourceforge.net/project/tinyos/OldFiles/tinyos/1.1.0/tinyos-1.1.0.tar.gz",
+  "http://master.dl.sourceforge.net/project/tinyos/OldFiles/tinyos/1.1.0/tinyos-1.1.0.tar.gz": "pkg:sourceforge/tinyos/tinyos@1.1.0",
   "http://master.dl.sourceforge.net/project/urlchecker/lu/ng/urlchecker/urlchecker/1.7/urlchecker-1.7-javadoc.jar": "pkg:sourceforge/urlchecker?download_url=http://master.dl.sourceforge.net/project/urlchecker/lu/ng/urlchecker/urlchecker/1.7/urlchecker-1.7-javadoc.jar",
   "http://master.dl.sourceforge.net/project/zclasspath/maven2/org/zclasspath/zclasspath/1.5/zclasspath-1.5.jar": "pkg:sourceforge/zclasspath?download_url=http://master.dl.sourceforge.net/project/zclasspath/maven2/org/zclasspath/zclasspath/1.5/zclasspath-1.5.jar",
   "http://master.dl.sourceforge.net/project/googleimagedown/project/v1.1/GoogleImageDownloader-v1.1-src.tar.bz2": "pkg:sourceforge/googleimagedown?download_url=http://master.dl.sourceforge.net/project/googleimagedown/project/v1.1/GoogleImageDownloader-v1.1-src.tar.bz2",
-  "https://sourceforge.net/projects/scribus/files/scribus/1.6.0/scribus-1.6.0.tar.gz/download": "pkg:sourceforge/scribus/scribus@1.6.0",
+  "https://sourceforge.net/projects/scribus/files/scribus/1.6.0/scribus-1.6.0.tar.gz/download": "pkg:sourceforge/scribus@1.6.0",
+  "https://sourceforge.net/projects/turbovnc/files/3.1/turbovnc-3.1.tar.gz/download": "pkg:sourceforge/turbovnc@3.1",
+  "https://sourceforge.net/projects/ventoy/files/v1.0.96/Ventoy%201.0.96%20release%20source%20code.tar.gz/download": "pkg:sourceforge/ventoy@1.0.96",
+  "https://sourceforge.net/projects/geoserver/files/GeoServer/2.23.4/geoserver-2.23.4-war.zip/download": "pkg:sourceforge/geoserver@2.23.4",
   "https://crates.io/api/v1/crates/rand/0.7.2/download": "pkg:cargo/rand@0.7.2",
   "https://crates.io/api/v1/crates/clap/2.33.0/download": "pkg:cargo/clap@2.33.0",
   "https://crates.io/api/v1/crates/structopt/0.3.11/download": "pkg:cargo/structopt@0.3.11",