add support for detecting non-licensed projects

lib/licensee.rb

@@ -10,6 +10,7 @@ require_relative "licensee/license_file"
 require_relative "licensee/project"
 require_relative "licensee/matcher"
 require_relative "licensee/matchers/exact_matcher"
+require_relative "licensee/matchers/copyright_matcher"
 require_relative "licensee/matchers/git_matcher"
 require_relative "licensee/matchers/levenshtein_matcher"
 
@@ -41,7 +42,7 @@ class Licensee
     # Array of matchers to use, in order of preference
     # The order should be decending order of anticipated speed to match
     def matchers
-      [Licensee::ExactMatcher, Licensee::GitMatcher, Licensee::LevenshteinMatcher]
+      [Licensee::ExactMatcher, Licensee::CopyrightMatcher, Licensee::GitMatcher, Licensee::LevenshteinMatcher]
     end
   end
 end

lib/licensee/license.rb

@@ -55,6 +55,8 @@ class Licensee
     def hashsig
       @hashsig ||= Rugged::Blob::HashSignature.new(
         body, Rugged::Blob::HashSignature::WHITESPACE_SMART)
+    rescue Rugged::InvalidError
+      nil
     end
 
     def inspect

lib/licensee/matchers/copyright_matcher.rb

@@ -0,0 +1,18 @@
+class Licensee
+  class CopyrightMatcher < Matcher
+
+    REGEX = /^(©|\(c\)|Copyright) \d{4}(.*)$/i
+
+    def no_license
+      @no_license ||= Licensee::Licenses.find("no-license")
+    end
+
+    def match
+      no_license if file.content_normalized =~ REGEX
+    end
+
+    def confidence
+      100
+    end
+  end
+end

lib/licensee/matchers/git_matcher.rb

@@ -16,7 +16,7 @@ class Licensee
     end
 
     def similarity(other)
-      file.blob.similarity(other.hashsig)
+      other.hashsig ? file.blob.similarity(other.hashsig) : 0
     end
 
     # Pulled out for easier testing

script/vendor-licenses

@@ -1,3 +1,2 @@
 rm -Rf vendor/choosealicense.com
 bower install github/choosealicense.com#gh-pages
-rm -f vendor/choosealicense.com/_licenses/no-license.txt

test/functions.rb

@@ -30,11 +30,13 @@ class FakeBlob
   end
 
   def similarity(other)
-    Rugged::Blob::HashSignature.compare(self.hashsig, other)
+    self.hashsig ? Rugged::Blob::HashSignature.compare(self.hashsig, other) : 0
   end
 
   def hashsig(options = 0)
     @hashsig ||= Rugged::Blob::HashSignature.new(content, options)
+  rescue Rugged::InvalidError
+    nil
   end
 end
 

test/test_licensee.rb

@@ -3,7 +3,7 @@ require 'helper'
 class TestLicensee < Minitest::Test
   should "know the licenses" do
     assert_equal Array, Licensee.licenses.class
-    assert_equal 15, Licensee.licenses.size
+    assert_equal 16, Licensee.licenses.size
     assert_equal Licensee::License, Licensee.licenses.first.class
   end
 

test/test_licensee_copyright_matcher.rb

@@ -0,0 +1,39 @@
+require 'helper'
+
+class TestLicenseeCopyrightMatcher < Minitest::Test
+
+  def setup
+    text = "Copyright 2015 Ben Balter"
+    blob = FakeBlob.new(text)
+    @file = Licensee::LicenseFile.new(blob)
+  end
+
+  should "match the license" do
+    assert_equal "no-license", Licensee::CopyrightMatcher.match(@file).key
+  end
+
+  should "know the match confidence" do
+    assert_equal 100, Licensee::CopyrightMatcher.new(@file).confidence
+  end
+
+  should "match (C) copyright notices" do
+    text = "(C) 2015 Ben Balter"
+    blob = FakeBlob.new(text)
+    file = Licensee::LicenseFile.new(blob)
+    assert_equal "no-license", Licensee::CopyrightMatcher.match(file).key
+  end
+
+  should "match © copyright notices" do
+    text = "© 2015 Ben Balter"
+    blob = FakeBlob.new(text)
+    file = Licensee::LicenseFile.new(blob)
+    assert_equal "no-license", Licensee::CopyrightMatcher.match(file).key
+  end
+
+  should "not false positive" do
+    text = File.open(Licensee::Licenses.find("mit").path).read.split("---").last
+    blob = FakeBlob.new(text)
+    file = Licensee::LicenseFile.new(blob)
+    assert_equal nil, Licensee::CopyrightMatcher.match(file)
+  end
+end

test/test_licensee_licenses.rb

@@ -4,12 +4,12 @@ class TestLicenseeLicenses < Minitest::Test
 
   should "know license names" do
     assert_equal Array, Licensee::Licenses.keys.class
-    assert_equal 15, Licensee::Licenses.keys.size
+    assert_equal 16, Licensee::Licenses.keys.size
   end
 
   should "load the licenses" do
     assert_equal Array, Licensee::Licenses.list.class
-    assert_equal 15, Licensee::Licenses.list.size
+    assert_equal 16, Licensee::Licenses.list.size
     assert_equal Licensee::License, Licensee::Licenses.list.first.class
   end
 

test/test_licensee_vendor.rb

@@ -11,21 +11,21 @@ class TestLicenseeVendor < Minitest::Test
   should "detect each vendored license when modified" do
     licenses = Dir["#{Licensee::Licenses.base}/*"].shuffle
     licenses.each do |license|
-      verify_license_file(license, true)
+      verify_license_file(license, true) unless license =~ /no-license\.txt$/
     end
   end
 
   should "detect each vendored license with different line lengths" do
     licenses = Dir["#{Licensee::Licenses.base}/*"].shuffle
     licenses.each do |license|
-      verify_license_file(license, false, 50)
+      verify_license_file(license, false, 50) unless license =~ /no-license\.txt$/
     end
   end
 
   should "detect each vendored license with different line lengths when modified" do
     licenses = Dir["#{Licensee::Licenses.base}/*"].shuffle
     licenses.each do |license|
-      verify_license_file(license, true, 50)
+      verify_license_file(license, true, 50) unless license =~ /no-license\.txt$/
     end
   end
 end

vendor/choosealicense.com/_licenses/no-license.txt

@@ -0,0 +1,28 @@
+---
+layout: license
+permalink: /licenses/no-license/
+category: No License
+class: license-types
+title: No License
+
+description: You retain all rights and do not permit distribution, reproduction, or derivative works. You may grant some rights in cases where you publish your source code to a site that requires accepting terms of service. For example, publishing code in a public repository on GitHub requires that you allow others to view and fork your code.
+
+note: This option may be subject to the Terms Of Use of the site where you publish your source code.
+
+how: Simply do nothing, though including a copyright notice is recommended.
+
+required:
+  - include-copyright
+
+permitted:
+  - commercial-use
+  - private-use
+
+forbidden:
+  - modifications
+  - distribution
+  - sublicense
+
+---
+
+Copyright [year] [fullname]