Copyright detection improvements

 * Correct detection for #168 with some new exception to proper nouns
 * Test is now passing for #191
 * New test for #130

Signed-off-by: Philippe Ombredanne <[email protected]>

Author: pombredanne

src/cluecode/copyrights.py

@@ -346,6 +346,9 @@ def detect(location):
     # exceptions to proper nouns
     (r'^(The|Commons|AUTHOR|software)$', 'NN'),
 
+    # exceptions to proper noun
+    (r"^(Natural|Docs?)$", 'NN'),
+
     # composed proper nouns, ie. Jean-Claude or ST-Microelectronics
     # FIXME: what about a variant with spaces around the dash?
     (r'^[A-Z][a-zA-Z]*\s?[\-]\s?[A-Z]?[a-zA-Z]+.?$', 'NNP'),
@@ -571,6 +574,9 @@ def detect(location):
 # XZY emails
     COMPANY: {<COMPANY> <EMAIL>+}        #1400
 
+# by the a href http wtforms.simplecodes.com WTForms Team
+    COMPANY: {<BY> <NN>+ <COMP|COMPANY>}        #1420
+
 
 # "And" some name
     ANDCO: {<CC>+ <NN> <NNP>+<UNI|COMP>?}        #1430

tests/cluecode/test_copyrights.py

@@ -3966,10 +3966,9 @@ def test_copyright_should_not_contain_leading_or_trailing_colon(self):
         expected = ['copyright (c) 2013 by Armin Ronacher.']
         check_detection(expected, test_file)
 
-    @expectedFailure
     def test_copyright_in_markup_should_not_be_truncated(self):
         test_file = self.get_test_loc('copyrights/copyright_in_html.html')
-        expected = ['(c) Copyright 2010 by the WTForms Team']
+        expected = ["(c) Copyright 2010 by the <a href http://wtforms.simplecodes.com'>WTForms Team"]
         check_detection(expected, test_file)
 
     def test_copyright_should_not_have_trailing_garbage(self):
@@ -4138,24 +4137,60 @@ def test_copyright_copr5_correct(self):
 
     def test_copyright_oracle(self):
         test_lines = ['Copyright (c) 1997-2015 Oracle and/or its affiliates. All rights reserved.']
-
         expected = ['Copyright (c) 1997-2015 Oracle and/or its affiliates.']
         check_detection(expected, test_lines)
 
     def test_copyright_windows(self):
         test_lines = ['This release supports NT-based Windows releases like Windows 2000 SP4, Windows XP, and Windows 2003.']
-
         expected = []
         check_detection(expected, test_lines)
 
     def test_copyright_in_binary_sql_server(self):
         test_lines = ['2005charchar? 7 DDLSQL Server 2005smalldatetimedatetimeLDDDDDD7']
-
         expected = []
         check_detection(expected, test_lines)
 
     def test_copyright_with_example_com_url(self):
         test_lines = ['"domain": function(c) { assert.equal(c.domain, "example.com") },']
-
         expected = []
         check_detection(expected, test_lines)
+
+    def test_copyright_various(self):
+        test_lines = '''
+            libwmf (<libwmf/api.h>): library for wmf conversion
+            Copyright (C) 2000 - various; see CREDITS, ChangeLog, and sources
+            The libwmf Library is free software; you can redistribute it and/or
+        '''.splitlines(False)
+        expected = ['Copyright (c) 2000 - various'] # ; see CREDITS, ChangeLog, and sources
+        check_detection(expected, test_lines)
+
+    def test_copyright_natural_docs(self):
+        test_lines = '''
+            // Search script generated by doxygen
+            // Copyright (C) 2009 by Dimitri van Heesch.
+            
+            // The code in this file is loosly based on main.js, part of Natural Docs,
+            // which is Copyright (C) 2003-2008 Greg Valure
+            // Natural Docs is licensed under the GPL.
+        '''.splitlines(False)
+        expected = [
+            u'Copyright (c) 2009 by Dimitri van Heesch.',
+            u'Copyright (c) 2003-2008 Greg Valure'
+        ]
+        check_detection(expected, test_lines)
+
+    def test_copyright_and_authors_mixed(self):
+        test_lines = '''
+             * Copyright (c) 1998 Softweyr LLC.  All rights reserved.
+             *
+             * strtok_r, from Berkeley strtok
+             * Oct 13, 1998 by Wes Peters <[email protected]>
+             *
+             * Copyright (c) 1988, 1993
+             *  The Regents of the University of California.  All rights reserved.
+        '''.splitlines(False)
+        expected = [
+            u'Copyright (c) 1998 Softweyr LLC.',
+            u'Copyright (c) 1988, 1993 The Regents of the University of California.'
+        ]
+        check_detection(expected, test_lines)