mirror of
				https://github.com/KevinMidboe/linguist.git
				synced 2025-10-29 17:50:22 +00:00 
			
		
		
		
	Merge pull request #1788 from github/refactor-heuristics
Refactor heuristics (updated)
This commit is contained in:
		@@ -1,8 +1,6 @@
 | 
			
		||||
module Linguist
 | 
			
		||||
  # A collection of simple heuristics that can be used to better analyze languages.
 | 
			
		||||
  class Heuristics
 | 
			
		||||
    ACTIVE = true
 | 
			
		||||
 | 
			
		||||
    # Public: Use heuristics to detect language of the blob.
 | 
			
		||||
    #
 | 
			
		||||
    # blob               - An object that quacks like a blob.
 | 
			
		||||
@@ -14,177 +12,123 @@ module Linguist
 | 
			
		||||
    #     Language["Ruby"], Language["Python"]
 | 
			
		||||
    #   ])
 | 
			
		||||
    #
 | 
			
		||||
    # Returns an Array with one Language if a heuristic matched, or empty if
 | 
			
		||||
    # none matched or were inconclusive.
 | 
			
		||||
    # Returns an Array of languages, or empty if none matched or were inconclusive.
 | 
			
		||||
    def self.call(blob, languages)
 | 
			
		||||
      find_by_heuristics(blob.data, languages.map(&:name))
 | 
			
		||||
    end
 | 
			
		||||
      data = blob.data
 | 
			
		||||
 | 
			
		||||
    # Public: Given an array of String language names,
 | 
			
		||||
    # apply heuristics against the given data and return an array
 | 
			
		||||
    # of matching languages, or nil.
 | 
			
		||||
    #
 | 
			
		||||
    # data      - Array of tokens or String data to analyze.
 | 
			
		||||
    # languages - Array of language name Strings to restrict to.
 | 
			
		||||
    #
 | 
			
		||||
    # Returns an array of Languages or []
 | 
			
		||||
    def self.find_by_heuristics(data, languages)
 | 
			
		||||
      if active?
 | 
			
		||||
        result = []
 | 
			
		||||
 | 
			
		||||
        if languages.all? { |l| ["Perl", "Prolog"].include?(l) }
 | 
			
		||||
          result = disambiguate_pl(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["ECL", "Prolog"].include?(l) }
 | 
			
		||||
          result = disambiguate_ecl(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["IDL", "Prolog"].include?(l) }
 | 
			
		||||
          result = disambiguate_pro(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["Common Lisp", "OpenCL"].include?(l) }
 | 
			
		||||
          result = disambiguate_cl(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["Hack", "PHP"].include?(l) }
 | 
			
		||||
          result = disambiguate_hack(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["Scala", "SuperCollider"].include?(l) }
 | 
			
		||||
          result = disambiguate_sc(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["AsciiDoc", "AGS Script"].include?(l) }
 | 
			
		||||
          result = disambiguate_asc(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["FORTRAN", "Forth"].include?(l) }
 | 
			
		||||
          result = disambiguate_f(data)
 | 
			
		||||
        end
 | 
			
		||||
        if languages.all? { |l| ["F#", "Forth", "GLSL"].include?(l) }
 | 
			
		||||
          result = disambiguate_fs(data)
 | 
			
		||||
        end
 | 
			
		||||
        return result
 | 
			
		||||
      @heuristics.each do |heuristic|
 | 
			
		||||
        return Array(heuristic.call(data)) if heuristic.matches?(languages)
 | 
			
		||||
      end
 | 
			
		||||
 | 
			
		||||
      [] # No heuristics matched
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    # .h extensions are ambiguous between C, C++, and Objective-C.
 | 
			
		||||
    # We want to shortcut look for Objective-C _and_ now C++ too!
 | 
			
		||||
    # Internal: Define a new heuristic.
 | 
			
		||||
    #
 | 
			
		||||
    # Returns an array of Languages or []
 | 
			
		||||
    def self.disambiguate_c(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
      if data.include?("@interface")
 | 
			
		||||
        matches << Language["Objective-C"]
 | 
			
		||||
      elsif data.include?("#include <cstdint>")
 | 
			
		||||
        matches << Language["C++"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    # languages - String names of languages to disambiguate.
 | 
			
		||||
    # heuristic - Block which takes data as an argument and returns a Language or nil.
 | 
			
		||||
    #
 | 
			
		||||
    # Examples
 | 
			
		||||
    #
 | 
			
		||||
    #     disambiguate "Perl", "Prolog" do |data|
 | 
			
		||||
    #       if data.include?("use strict")
 | 
			
		||||
    #         Language["Perl"]
 | 
			
		||||
    #       elsif data.include?(":-")
 | 
			
		||||
    #         Language["Prolog"]
 | 
			
		||||
    #       end
 | 
			
		||||
    #     end
 | 
			
		||||
    #
 | 
			
		||||
    def self.disambiguate(*languages, &heuristic)
 | 
			
		||||
      @heuristics << new(languages, &heuristic)
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_pl(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
    # Internal: Array of defined heuristics
 | 
			
		||||
    @heuristics = []
 | 
			
		||||
 | 
			
		||||
    # Internal
 | 
			
		||||
    def initialize(languages, &heuristic)
 | 
			
		||||
      @languages = languages
 | 
			
		||||
      @heuristic = heuristic
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    # Internal: Check if this heuristic matches the candidate languages.
 | 
			
		||||
    def matches?(candidates)
 | 
			
		||||
      candidates.all? { |l| @languages.include?(l.name) }
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    # Internal: Perform the heuristic
 | 
			
		||||
    def call(data)
 | 
			
		||||
      @heuristic.call(data)
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    disambiguate "Perl", "Prolog" do |data|
 | 
			
		||||
      if data.include?("use strict")
 | 
			
		||||
        matches << Language["Perl"]
 | 
			
		||||
        Language["Perl"]
 | 
			
		||||
      elsif data.include?(":-")
 | 
			
		||||
        matches << Language["Prolog"]
 | 
			
		||||
        Language["Prolog"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_ecl(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
    disambiguate "ECL", "Prolog" do |data|
 | 
			
		||||
      if data.include?(":-")
 | 
			
		||||
        matches << Language["Prolog"]
 | 
			
		||||
        Language["Prolog"]
 | 
			
		||||
      elsif data.include?(":=")
 | 
			
		||||
        matches << Language["ECL"]
 | 
			
		||||
        Language["ECL"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_pro(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
      if (data.include?(":-"))
 | 
			
		||||
        matches << Language["Prolog"]
 | 
			
		||||
    disambiguate "IDL", "Prolog" do |data|
 | 
			
		||||
      if data.include?(":-")
 | 
			
		||||
        Language["Prolog"]
 | 
			
		||||
      else
 | 
			
		||||
        matches << Language["IDL"]
 | 
			
		||||
        Language["IDL"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_ts(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
      if (data.include?("</translation>"))
 | 
			
		||||
        matches << Language["XML"]
 | 
			
		||||
      else
 | 
			
		||||
        matches << Language["TypeScript"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_cl(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
    disambiguate "Common Lisp", "OpenCL" do |data|
 | 
			
		||||
      if data.include?("(defun ")
 | 
			
		||||
        matches << Language["Common Lisp"]
 | 
			
		||||
        Language["Common Lisp"]
 | 
			
		||||
      elsif /\/\* |\/\/ |^\}/.match(data)
 | 
			
		||||
        matches << Language["OpenCL"]
 | 
			
		||||
        Language["OpenCL"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_r(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
      matches << Language["Rebol"] if /\bRebol\b/i.match(data)
 | 
			
		||||
      matches << Language["R"] if data.include?("<-")
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_hack(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
    disambiguate "Hack", "PHP" do |data|
 | 
			
		||||
      if data.include?("<?hh")
 | 
			
		||||
        matches << Language["Hack"]
 | 
			
		||||
        Language["Hack"]
 | 
			
		||||
      elsif /<?[^h]/.match(data)
 | 
			
		||||
        matches << Language["PHP"]
 | 
			
		||||
        Language["PHP"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_sc(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
      if (/\^(this|super)\./.match(data) || /^\s*(\+|\*)\s*\w+\s*{/.match(data) || /^\s*~\w+\s*=\./.match(data))
 | 
			
		||||
        matches << Language["SuperCollider"]
 | 
			
		||||
    disambiguate "Scala", "SuperCollider" do |data|
 | 
			
		||||
      if /\^(this|super)\./.match(data) || /^\s*(\+|\*)\s*\w+\s*{/.match(data) || /^\s*~\w+\s*=\./.match(data)
 | 
			
		||||
        Language["SuperCollider"]
 | 
			
		||||
      elsif /^\s*import (scala|java)\./.match(data) || /^\s*val\s+\w+\s*=/.match(data) || /^\s*class\b/.match(data)
 | 
			
		||||
        Language["Scala"]
 | 
			
		||||
      end
 | 
			
		||||
      if (/^\s*import (scala|java)\./.match(data) || /^\s*val\s+\w+\s*=/.match(data) || /^\s*class\b/.match(data))
 | 
			
		||||
        matches << Language["Scala"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_asc(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
      matches << Language["AsciiDoc"] if /^=+(\s|\n)/.match(data)
 | 
			
		||||
      matches
 | 
			
		||||
    disambiguate "AsciiDoc", "AGS Script" do |data|
 | 
			
		||||
      Language["AsciiDoc"] if /^=+(\s|\n)/.match(data)
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_f(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
    disambiguate "FORTRAN", "Forth" do |data|
 | 
			
		||||
      if /^: /.match(data)
 | 
			
		||||
        matches << Language["Forth"]
 | 
			
		||||
        Language["Forth"]
 | 
			
		||||
      elsif /^([c*][^a-z]|      subroutine\s)/i.match(data)
 | 
			
		||||
        matches << Language["FORTRAN"]
 | 
			
		||||
        Language["FORTRAN"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.disambiguate_fs(data)
 | 
			
		||||
      matches = []
 | 
			
		||||
    disambiguate "F#", "Forth", "GLSL" do |data|
 | 
			
		||||
      if /^(: |new-device)/.match(data)
 | 
			
		||||
        matches << Language["Forth"]
 | 
			
		||||
        Language["Forth"]
 | 
			
		||||
      elsif /^(#light|import|let|module|namespace|open|type)/.match(data)
 | 
			
		||||
        matches << Language["F#"]
 | 
			
		||||
        Language["F#"]
 | 
			
		||||
      elsif /^(#include|#pragma|precision|uniform|varying|void)/.match(data)
 | 
			
		||||
        matches << Language["GLSL"]
 | 
			
		||||
        Language["GLSL"]
 | 
			
		||||
      end
 | 
			
		||||
      matches
 | 
			
		||||
    end
 | 
			
		||||
 | 
			
		||||
    def self.active?
 | 
			
		||||
      !!ACTIVE
 | 
			
		||||
    end
 | 
			
		||||
  end
 | 
			
		||||
end
 | 
			
		||||
 
 | 
			
		||||
		Reference in New Issue
	
	Block a user