Module: Linguist::Samples

Defined in:
lib/linguist/samples.rb

Overview

Model for accessing classifier training data.

Constant Summary collapse

ROOT =

Path to samples root directory

File.expand_path("../../../samples", __FILE__)
PATH =

Path for serialized samples db

File.expand_path('../samples.json', __FILE__)

Class Method Summary collapse

Class Method Details

.cache ⇒ Object

Hash of serialized samples object



20
21
22
23
24
25
 
# File 'lib/linguist/samples.rb', line 20

def self.cache
  @cache ||= begin
    serializer = defined?(JSON) ? JSON : YAML
    serializer.load(File.read(PATH))
  end
end

.data ⇒ Object

Public: Build Classifier from all samples.

Returns trained Classifier.



75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
 
# File 'lib/linguist/samples.rb', line 75

def self.data
  db = {}
  db['extnames'] = {}
  db['interpreters'] = {}
  db['filenames'] = {}

  each do |sample|
    language_name = sample[:language]

    if sample[:extname]
      db['extnames'][language_name] ||= []
      if !db['extnames'][language_name].include?(sample[:extname])
        db['extnames'][language_name] << sample[:extname]
        db['extnames'][language_name].sort!
      end
    end

    if sample[:interpreter]
      db['interpreters'][language_name] ||= []
      if !db['interpreters'][language_name].include?(sample[:interpreter])
        db['interpreters'][language_name] << sample[:interpreter]
        db['interpreters'][language_name].sort!
      end
    end

    if sample[:filename]
      db['filenames'][language_name] ||= []
      db['filenames'][language_name] << sample[:filename]
      db['filenames'][language_name].sort!
    end

    data = File.read(sample[:path])
    Classifier.train!(db, language_name, data)
  end

  db['md5'] = Linguist::MD5.hexdigest(db)

  db
end

.each(&block) ⇒ Object

Public: Iterate over each sample.

&block - Yields Sample to block

Returns nothing.



32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/linguist/samples.rb', line 32

def self.each(&block)
  Dir.entries(ROOT).sort!.each do |category|
    next if category == '.' || category == '..'

    # Skip text and binary for now
    # Possibly reconsider this later
    next if category == 'Text' || category == 'Binary'

    dirname = File.join(ROOT, category)
    Dir.entries(dirname).each do |filename|
      next if filename == '.' || filename == '..'

      if filename == 'filenames'
        Dir.entries(File.join(dirname, filename)).each do |subfilename|
          next if subfilename == '.' || subfilename == '..'

          yield({
            :path    => File.join(dirname, filename, subfilename),
            :language => category,
            :filename => subfilename
          })
        end
      else
        if File.extname(filename) == ""
          raise "#{File.join(dirname, filename)} is missing an extension, maybe it belongs in filenames/ subdir"
        end

        yield({
          :path     => File.join(dirname, filename),
          :language => category,
          :interpreter => File.exist?(filename) ? Linguist.interpreter_from_shebang(File.read(filename)) : nil,
          :extname  => File.extname(filename)
        })
      end
    end
  end

  nil
end