Add support for more kind of tarball URLs

Hit issues with those urls on ERP5 SR because they were not recognized previously.

nxdbom/__init__.py

@@ -32,6 +32,7 @@ import sys, configparser, re
 from os.path import basename
 from glob import glob
 from collections import namedtuple
+from urllib.parse import unquote
 
 
 # PkgInfo represents information about a package
@@ -194,28 +195,57 @@ def bom_node(XXX):
 #
 #     http://www.python.org/ftp/python/2.7.18/Python-2.7.18.tar.xz  ->  ('Python', '2.7.18')
 _gitweb_re = re.compile(r'/gitweb/\?p=(?P<name>\w+)\.git;a=snapshot;h=(?P<rev>\w+)')
-_github_re = re.compile(r'github.com/\w+/(?P<name>\w+)/archive/(?P<rev>.+)$')
+_github_re = re.compile(r'/github.com/[\w\-]+/(?P<name>[\w\-]+)/archive/(refs/tags/)?(?P<rev>.+)$')
+_github_rre= re.compile(r'/raw.githubusercontent.com/[\w\-]+/(?P<name>[\w\-]+)/(?P<rev>[\w\.\-]+)/')
+_SF_re     = re.compile(r'/sourceforge.net/.+/(?P<name>[\w\_]+)-(?P<rev>[\w\.]+)/download$')
+_go_re     = re.compile(r'/golang.org/dl/(?P<name>go)(?P<rev>[\w\.]+).src$')
+_osdn_f    = re.compile(r'/osdn.net/frs/redir.php\?f=(?P<f>.+)$')
 def namever(url): # -> (name, ver)
-    for tail in ('.tgz', '.tar.gz', '.tbz', '.tar.bz2', '.tar.xz', '.tar.lz',
-                 '-py2.7.egg', '-py2.7-linux-x86_64.egg'):    # FIXME -> re to cover py2/py3 x86/arm ...
-        url = removesuffix(url, tail)
+    # http://www.ijg.org/files/jpegsrc.v9d.tar.gz  ->  http://www.ijg.org/files/jpegsrc.v9d
+    def del_tgztail(s):
+        for tail in ('.tgz', '.tar.gz', '.tbz', '.tar.bz2', '.tar.xz', '.tar.lz', '.zip',
+                     '-py2.7.egg', '-py2.7-linux-x86_64.egg'):    # FIXME -> re to cover py2/py3 x86/arm ...
+            s = removesuffix(s, tail)
+        return s
+    url = del_tgztail(url)
+
     name, ver = _namever(url)
-    ver = removeprefix(ver, 'v')
+    if ver is not None:
+        # swig-3.0.12.tar.gz  ->  swig-3.0.12
+        # originally from https://sourceforge.net/projects/swig/files/swig/swig-3.0.12/swig-3.0.12.tar.gz/download
+        ver = del_tgztail(ver)
+        ver = removeprefix(ver, 'v')
     return name, ver
 
 def _namever(url):
-    for r in (_gitweb_re, _github_re):
+    for r in (_gitweb_re, _github_re, _github_rre, _SF_re, _go_re):
         m = r.search(url)
         if m is not None:
             return m.group('name'), m.group('rev')
 
+    m = _osdn_f.search(url)
+    if m is not None:
+        url = unquote(m.group('f'))
+
     filename = basename(url)
     # re.rsplit([-_], filename, 1)
     m = re.search('[-_][^-_]*$', filename)
-    assert m is not None
-    name = filename[:m.start()]
-    ver  = filename[m.start()+1:]
-    return name, ver
+    if m is not None:
+        name = filename[:m.start()]
+        ver  = filename[m.start()+1:]
+        return name, ver
+
+    m = re.search(r'\.v.+$', filename)  # jpegsrc.v9d
+    if m is not None:
+        name = filename[:m.start()]
+        ver  = filename[m.start()+1:]
+        return name, ver
+
+    m = re.match(r'(?P<name>.*[^0-9])(?P<rev>[0-9]+)$', filename) # IPAexfont00201
+    if m is not None:
+        return m.group('name'), m.group('rev')
+
+    raise RuntimeError('Unsupported url: %r' % (url,))
 
 
 # isconf returns whether url points to data related to configuration file

nxdbom/nxdbom_test.py

@@ -27,8 +27,17 @@ from os.path import dirname
 
 @pytest.mark.parametrize('url,nameok,verok', [
     ('http://www.python.org/ftp/python/2.7.18/Python-2.7.18.tar.xz',  'Python',  '2.7.18'),
+    ('http://www.ijg.org/files/jpegsrc.v9d.tar.gz',                   'jpegsrc', '9d'),
     ('https://github.com/nghttp2/nghttp2/archive/v1.40.0.tar.gz',     'nghttp2', '1.40.0'),
+    ('https://golang.org/dl/go1.18.9.src',                            'go',      '1.18.9'),
+    ('https://github.com/tesseract-ocr/tesseract/archive/refs/tags/4.1.1.tar.gz',  'tesseract', '4.1.1'),
+    ('https://raw.githubusercontent.com/zuphilip/ocropy-models/master/en-default.pyrnn.gz',  'ocropy-models', 'master'),
+    ('https://sourceforge.net/projects/swig/files/swig/swig-3.0.12/swig-3.0.12.tar.gz/download',  'swig', '3.0.12'),
     ('https://git.savannah.gnu.org/gitweb/?p=config.git;a=snapshot;h=5e531d39;sf=tgz',  'config', '5e531d39'),
+    ('/ROOT/develop-eggs/mysqlclient-1.3.12-py2.7-linux-x86_64.egg',  'mysqlclient', '1.3.12'),
+    ('https://osdn.net/frs/redir.php?f=ipafonts%2F57330%2FIPAexfont00201.zip', 'IPAexfont', '00201'),
+    ('https://osdn.net/frs/redir.php?f=ipafonts%2F51868%2FIPAfont00303.zip',   'IPAfont',   '00303'),
+    ('https://osdn.net/frs/redir.php?f=tsukurimashou%2F56948%2Focr-0.2.zip',   'ocr',       '0.2'),
 ])
 def test_namever(url, nameok, verok):
     assert nxdbom.namever(url) == (nameok, verok)