geothird-linguist 2.6.1

data/lib/linguist/file_blob.rb

@@ -0,0 +1,56 @@
+require 'linguist/blob_helper'
+
+module Linguist
+  # A FileBlob is a wrapper around a File object to make it quack
+  # like a Grit::Blob. It provides the basic interface: `name`,
+  # `data`, and `size`.
+  class FileBlob
+    include BlobHelper
+
+    # Public: Initialize a new FileBlob from a path
+    #
+    # path      - A path String that exists on the file system.
+    # base_path - Optional base to relativize the path
+    #
+    # Returns a FileBlob.
+    def initialize(path, base_path = nil)
+      @path = path
+      @name = base_path ? path.sub("#{base_path}/", '') : path
+    end
+
+    # Public: Filename
+    #
+    # Examples
+    #
+    #   FileBlob.new("/path/to/linguist/lib/linguist.rb").name
+    #   # =>  "/path/to/linguist/lib/linguist.rb"
+    #
+    #   FileBlob.new("/path/to/linguist/lib/linguist.rb",
+    #                "/path/to/linguist").name
+    #   # =>  "lib/linguist.rb"
+    #
+    # Returns a String
+    attr_reader :name
+
+    # Public: Read file permissions
+    #
+    # Returns a String like '100644'
+    def mode
+      File.stat(@path).mode.to_s(8)
+    end
+
+    # Public: Read file contents.
+    #
+    # Returns a String.
+    def data
+      File.read(@path)
+    end
+
+    # Public: Get byte size
+    #
+    # Returns an Integer.
+    def size
+      File.size(@path)
+    end
+  end
+end

data/lib/linguist/generated.rb

@@ -0,0 +1,175 @@
+module Linguist
+  class Generated
+    # Public: Is the blob a generated file?
+    #
+    # name - String filename
+    # data - String blob data. A block also maybe passed in for lazy
+    #        loading. This behavior is deprecated and you should always
+    #        pass in a String.
+    #
+    # Return true or false
+    def self.generated?(name, data)
+      new(name, data).generated?
+    end
+
+    # Internal: Initialize Generated instance
+    #
+    # name - String filename
+    # data - String blob data
+    def initialize(name, data)
+      @name = name
+      @extname = File.extname(name)
+      @_data = data
+    end
+
+    attr_reader :name, :extname
+
+    # Lazy load blob data if block was passed in.
+    #
+    # Awful, awful stuff happening here.
+    #
+    # Returns String data.
+    def data
+      @data ||= @_data.respond_to?(:call) ? @_data.call() : @_data
+    end
+
+    # Public: Get each line of data
+    #
+    # Returns an Array of lines
+    def lines
+      # TODO: data should be required to be a String, no nils
+      @lines ||= data ? data.split("\n", -1) : []
+    end
+
+    # Internal: Is the blob a generated file?
+    #
+    # Generated source code is supressed in diffs and is ignored by
+    # language statistics.
+    #
+    # Please add additional test coverage to
+    # `test/test_blob.rb#test_generated` if you make any changes.
+    #
+    # Return true or false
+    def generated?
+      name == 'Gemfile.lock' ||
+        minified_javascript? ||
+        compiled_coffeescript? ||
+        xcode_project_file? ||
+        generated_net_docfile? ||
+        generated_parser? ||
+        compiled_cython_file?
+    end
+
+    # Internal: Is the blob an XCode project file?
+    #
+    # Generated if the file extension is an XCode project
+    # file extension.
+    #
+    # Returns true of false.
+    def xcode_project_file?
+      ['.xib', '.nib', '.storyboard', '.pbxproj', '.xcworkspacedata', '.xcuserstate'].include?(extname)
+    end
+
+    # Internal: Is the blob minified JS?
+    #
+    # Consider JS minified if the average line length is
+    # greater then 100c.
+    #
+    # Returns true or false.
+    def minified_javascript?
+      return unless extname == '.js'
+      if lines.any?
+        (lines.inject(0) { |n, l| n += l.length } / lines.length) > 100
+      else
+        false
+      end
+    end
+
+    # Internal: Is the blob of JS generated by CoffeeScript?
+    #
+    # CoffeeScript is meant to output JS that would be difficult to
+    # tell if it was generated or not. Look for a number of patterns
+    # output by the CS compiler.
+    #
+    # Return true or false
+    def compiled_coffeescript?
+      return false unless extname == '.js'
+
+      # CoffeeScript generated by > 1.2 include a comment on the first line
+      if lines[0] =~ /^\/\/ Generated by /
+        return true
+      end
+
+      if lines[0] == '(function() {' &&     # First line is module closure opening
+          lines[-2] == '}).call(this);' &&  # Second to last line closes module closure
+          lines[-1] == ''                   # Last line is blank
+
+        score = 0
+
+        lines.each do |line|
+          if line =~ /var /
+            # Underscored temp vars are likely to be Coffee
+            score += 1 * line.gsub(/(_fn|_i|_len|_ref|_results)/).count
+
+            # bind and extend functions are very Coffee specific
+            score += 3 * line.gsub(/(__bind|__extends|__hasProp|__indexOf|__slice)/).count
+          end
+        end
+
+        # Require a score of 3. This is fairly arbitrary. Consider
+        # tweaking later.
+        score >= 3
+      else
+        false
+      end
+    end
+
+    # Internal: Is this a generated documentation file for a .NET assembly?
+    #
+    # .NET developers often check in the XML Intellisense file along with an
+    # assembly - however, these don't have a special extension, so we have to
+    # dig into the contents to determine if it's a docfile. Luckily, these files
+    # are extremely structured, so recognizing them is easy.
+    #
+    # Returns true or false
+    def generated_net_docfile?
+      return false unless extname.downcase == ".xml"
+      return false unless lines.count > 3
+
+      # .NET Docfiles always open with <doc> and their first tag is an
+      # <assembly> tag
+      return lines[1].include?("<doc>") &&
+        lines[2].include?("<assembly>") &&
+        lines[-2].include?("</doc>")
+    end
+
+    # Internal: Is the blob of JS a parser generated by PEG.js?
+    #
+    # PEG.js-generated parsers are not meant to be consumed by humans.
+    #
+    # Return true or false
+    def generated_parser?
+      return false unless extname == '.js'
+
+      # PEG.js-generated parsers include a comment near the top  of the file
+      # that marks them as such.
+      if lines[0..4].join('') =~ /^(?:[^\/]|\/[^\*])*\/\*(?:[^\*]|\*[^\/])*Generated by PEG.js/
+        return true
+      end
+
+      false
+    end
+
+    # Internal: Is this a compiled C/C++ file from Cython?
+    #
+    # Cython-compiled C/C++ files typically contain:
+    #     /* Generated by Cython x.x.x on ... */
+    # on the first line.
+    #
+    # Return true or false
+    def compiled_cython_file?
+      return false unless ['.c', '.cpp'].include? extname
+      return lines[0].include?("Generated by Cython")
+    end
+  end
+end

data/lib/linguist/language.rb

@@ -0,0 +1,481 @@
+require 'escape_utils'
+require 'pygments'
+require 'yaml'
+
+require 'linguist/classifier'
+require 'linguist/samples'
+
+module Linguist
+  # Language names that are recognizable by GitHub. Defined languages
+  # can be highlighted, searched and listed under the Top Languages page.
+  #
+  # Languages are defined in `lib/linguist/languages.yml`.
+  class Language
+    @languages       = []
+    @index           = {}
+    @name_index      = {}
+    @alias_index     = {}
+    @extension_index = Hash.new { |h,k| h[k] = [] }
+    @filename_index  = Hash.new { |h,k| h[k] = [] }
+
+    # Valid Languages types
+    TYPES = [:data, :markup, :programming]
+
+    # Internal: Create a new Language object
+    #
+    # attributes - A hash of attributes
+    #
+    # Returns a Language object
+    def self.create(attributes = {})
+      language = new(attributes)
+
+      @languages << language
+
+      # All Language names should be unique. Raise if there is a duplicate.
+      if @name_index.key?(language.name)
+        raise ArgumentError, "Duplicate language name: #{language.name}"
+      end
+
+      # Language name index
+      @index[language.name] = @name_index[language.name] = language
+
+      language.aliases.each do |name|
+        # All Language aliases should be unique. Raise if there is a duplicate.
+        if @alias_index.key?(name)
+          raise ArgumentError, "Duplicate alias: #{name}"
+        end
+
+        @index[name] = @alias_index[name] = language
+      end
+
+      language.extensions.each do |extension|
+        if extension !~ /^\./
+          raise ArgumentError, "Extension is missing a '.': #{extension.inspect}"
+        end
+
+        @extension_index[extension] << language
+      end
+
+      language.filenames.each do |filename|
+        @filename_index[filename] << language
+      end
+
+      language
+    end
+
+    # Public: Detects the Language of the blob.
+    #
+    # name - String filename
+    # data - String blob data. A block also maybe passed in for lazy
+    #        loading. This behavior is deprecated and you should always
+    #        pass in a String.
+    # mode - Optional String mode (defaults to nil)
+    #
+    # Returns Language or nil.
+    def self.detect(name, data, mode = nil)
+      # A bit of an elegant hack. If the file is exectable but extensionless,
+      # append a "magic" extension so it can be classified with other
+      # languages that have shebang scripts.
+      if File.extname(name).empty? && mode && (mode.to_i(8) & 05) == 05
+        name += ".script!"
+      end
+
+      possible_languages = find_by_filename(name)
+
+      if possible_languages.length > 1
+        data = data.call() if data.respond_to?(:call)
+        if data.nil? || data == ""
+          nil
+        elsif result = Classifier.classify(Samples::DATA, data, possible_languages.map(&:name)).first
+          Language[result[0]]
+        end
+      else
+        possible_languages.first
+      end
+    end
+
+    # Public: Get all Languages
+    #
+    # Returns an Array of Languages
+    def self.all
+      @languages
+    end
+
+    # Public: Look up Language by its proper name.
+    #
+    # name - The String name of the Language
+    #
+    # Examples
+    #
+    #   Language.find_by_name('Ruby')
+    #   # => #<Language name="Ruby">
+    #
+    # Returns the Language or nil if none was found.
+    def self.find_by_name(name)
+      @name_index[name]
+    end
+
+    # Public: Look up Language by one of its aliases.
+    #
+    # name - A String alias of the Language
+    #
+    # Examples
+    #
+    #   Language.find_by_alias('cpp')
+    #   # => #<Language name="C++">
+    #
+    # Returns the Lexer or nil if none was found.
+    def self.find_by_alias(name)
+      @alias_index[name]
+    end
+
+    # Public: Look up Languages by filename.
+    #
+    # filename - The path String.
+    #
+    # Examples
+    #
+    #   Language.find_by_filename('foo.rb')
+    #   # => [#<Language name="Ruby">]
+    #
+    # Returns all matching Languages or [] if none were found.
+    def self.find_by_filename(filename)
+      basename, extname = File.basename(filename), File.extname(filename)
+      @filename_index[basename] + @extension_index[extname]
+    end
+
+    # Public: Look up Language by its name or lexer.
+    #
+    # name - The String name of the Language
+    #
+    # Examples
+    #
+    #   Language['Ruby']
+    #   # => #<Language name="Ruby">
+    #
+    #   Language['ruby']
+    #   # => #<Language name="Ruby">
+    #
+    # Returns the Language or nil if none was found.
+    def self.[](name)
+      @index[name]
+    end
+
+    # Public: A List of popular languages
+    #
+    # Popular languages are sorted to the top of language chooser
+    # dropdowns.
+    #
+    # This list is configured in "popular.yml".
+    #
+    # Returns an Array of Lexers.
+    def self.popular
+      @popular ||= all.select(&:popular?).sort_by { |lang| lang.name.downcase }
+    end
+
+    # Public: A List of non-popular languages
+    #
+    # Unpopular languages appear below popular ones in language
+    # chooser dropdowns.
+    #
+    # This list is created from all the languages not listed in "popular.yml".
+    #
+    # Returns an Array of Lexers.
+    def self.unpopular
+      @unpopular ||= all.select(&:unpopular?).sort_by { |lang| lang.name.downcase }
+    end
+
+    # Public: A List of languages with assigned colors.
+    #
+    # Returns an Array of Languages.
+    def self.colors
+      @colors ||= all.select(&:color).sort_by { |lang| lang.name.downcase }
+    end
+
+    # Public: A List of languages compatible with Ace.
+    #
+    # Returns an Array of Languages.
+    def self.ace_modes
+      @ace_modes ||= all.select(&:ace_mode).sort_by { |lang| lang.name.downcase }
+    end
+
+    # Internal: Initialize a new Language
+    #
+    # attributes - A hash of attributes
+    def initialize(attributes = {})
+      # @name is required
+      @name = attributes[:name] || raise(ArgumentError, "missing name")
+
+      # Set type
+      @type = attributes[:type] ? attributes[:type].to_sym : nil
+      if @type && !TYPES.include?(@type)
+        raise ArgumentError, "invalid type: #{@type}"
+      end
+
+      @color = attributes[:color]
+
+      # Set aliases
+      @aliases = [default_alias_name] + (attributes[:aliases] || [])
+
+      # Lookup Lexer object
+      @lexer = Pygments::Lexer.find_by_name(attributes[:lexer] || name) ||
+        raise(ArgumentError, "#{@name} is missing lexer")
+
+      @ace_mode = attributes[:ace_mode]
+      @wrap = attributes[:wrap] || false
+
+      # Set legacy search term
+      @search_term = attributes[:search_term] || default_alias_name
+
+      # Set extensions or default to [].
+      @extensions = attributes[:extensions] || []
+      @filenames  = attributes[:filenames]  || []
+
+      unless @primary_extension = attributes[:primary_extension]
+        raise ArgumentError, "#{@name} is missing primary extension"
+      end
+
+      # Prepend primary extension unless its already included
+      if primary_extension && !extensions.include?(primary_extension)
+        @extensions = [primary_extension] + extensions
+      end
+
+      # Set popular, and searchable flags
+      @popular    = attributes.key?(:popular)    ? attributes[:popular]    : false
+      @searchable = attributes.key?(:searchable) ? attributes[:searchable] : true
+
+      # If group name is set, save the name so we can lazy load it later
+      if attributes[:group_name]
+        @group = nil
+        @group_name = attributes[:group_name]
+
+      # Otherwise we can set it to self now
+      else
+        @group = self
+      end
+    end
+
+    # Public: Get proper name
+    #
+    # Examples
+    #
+    #   # => "Ruby"
+    #   # => "Python"
+    #   # => "Perl"
+    #
+    # Returns the name String
+    attr_reader :name
+
+    # Public: Get type.
+    #
+    # Returns a type Symbol or nil.
+    attr_reader :type
+
+    # Public: Get color.
+    #
+    # Returns a hex color String.
+    attr_reader :color
+
+    # Public: Get aliases
+    #
+    # Examples
+    #
+    #   Language['C++'].aliases
+    #   # => ["cpp"]
+    #
+    # Returns an Array of String names
+    attr_reader :aliases
+
+    # Deprecated: Get code search term
+    #
+    # Examples
+    #
+    #   # => "ruby"
+    #   # => "python"
+    #   # => "perl"
+    #
+    # Returns the name String
+    attr_reader :search_term
+
+    # Public: Get Lexer
+    #
+    # Returns the Lexer
+    attr_reader :lexer
+
+    # Public: Get Ace mode
+    #
+    # Examples
+    #
+    #  # => "text"
+    #  # => "javascript"
+    #  # => "c_cpp"
+    #
+    # Returns a String name or nil
+    attr_reader :ace_mode
+
+    # Public: Should language lines be wrapped
+    #
+    # Returns true or false
+    attr_reader :wrap
+
+    # Public: Get extensions
+    #
+    # Examples
+    #
+    #   # => ['.rb', '.rake', ...]
+    #
+    # Returns the extensions Array
+    attr_reader :extensions
+
+    # Deprecated: Get primary extension
+    #
+    # Defaults to the first extension but can be overriden
+    # in the languages.yml.
+    #
+    # The primary extension can not be nil. Tests should verify this.
+    #
+    # This attribute is only used by app/helpers/gists_helper.rb for
+    # creating the language dropdown. It really should be using `name`
+    # instead. Would like to drop primary extension.
+    #
+    # Returns the extension String.
+    attr_reader :primary_extension
+
+    # Public: Get filenames
+    #
+    # Examples
+    #
+    #   # => ['Rakefile', ...]
+    #
+    # Returns the extensions Array
+    attr_reader :filenames
+
+    # Public: Get URL escaped name.
+    #
+    # Examples
+    #
+    #   "C%23"
+    #   "C%2B%2B"
+    #   "Common%20Lisp"
+    #
+    # Returns the escaped String.
+    def escaped_name
+      EscapeUtils.escape_url(name).gsub('+', '%20')
+    end
+
+    # Internal: Get default alias name
+    #
+    # Returns the alias name String
+    def default_alias_name
+      name.downcase.gsub(/\s/, '-')
+    end
+
+    # Public: Get Language group
+    #
+    # Returns a Language
+    def group
+      @group ||= Language.find_by_name(@group_name)
+    end
+
+    # Public: Is it popular?
+    #
+    # Returns true or false
+    def popular?
+      @popular
+    end
+
+    # Public: Is it not popular?
+    #
+    # Returns true or false
+    def unpopular?
+      !popular?
+    end
+
+    # Public: Is it searchable?
+    #
+    # Unsearchable languages won't by indexed by solr and won't show
+    # up in the code search dropdown.
+    #
+    # Returns true or false
+    def searchable?
+      @searchable
+    end
+
+    # Public: Highlight syntax of text
+    #
+    # text    - String of code to be highlighted
+    # options - A Hash of options (defaults to {})
+    #
+    # Returns html String
+    def colorize(text, options = {})
+      lexer.highlight(text, options = {})
+    end
+
+    # Public: Return name as String representation
+    def to_s
+      name
+    end
+
+    def ==(other)
+      eql?(other)
+    end
+
+    def eql?(other)
+      equal?(other)
+    end
+
+    def hash
+      name.hash
+    end
+
+    def inspect
+      "#<#{self.class} name=#{name}>"
+    end
+  end
+
+  extensions = Samples::DATA['extnames']
+  filenames = Samples::DATA['filenames']
+  popular = YAML.load_file(File.expand_path("../popular.yml", __FILE__))
+
+  YAML.load_file(File.expand_path("../languages.yml", __FILE__)).each do |name, options|
+    options['extensions'] ||= []
+    options['filenames'] ||= []
+
+    if extnames = extensions[name]
+      extnames.each do |extname|
+        if !options['extensions'].include?(extname)
+          options['extensions'] << extname
+        else
+          warn "#{name} #{extname.inspect} is already defined in samples/. Remove from languages.yml."
+        end
+      end
+    end
+
+    if fns = filenames[name]
+      fns.each do |filename|
+        if !options['filenames'].include?(filename)
+          options['filenames'] << filename
+        else
+          warn "#{name} #{filename.inspect} is already defined in samples/. Remove from languages.yml."
+        end
+      end
+    end
+
+    Language.create(
+      :name              => name,
+      :color             => options['color'],
+      :type              => options['type'],
+      :aliases           => options['aliases'],
+      :lexer             => options['lexer'],
+      :ace_mode          => options['ace_mode'],
+      :wrap              => options['wrap'],
+      :group_name        => options['group'],
+      :searchable        => options.key?('searchable') ? options['searchable'] : true,
+      :search_term       => options['search_term'],
+      :extensions        => options['extensions'].sort,
+      :primary_extension => options['primary_extension'],
+      :filenames         => options['filenames'],
+      :popular           => popular.include?(name)
+    )
+  end
+end