Class: Gonzui::Indexer

Inherits:
Object
  • Object
show all
Includes:
Util
Defined in:
lib/gonzui/indexer.rb

Constant Summary collapse

BinaryRegexp = 
Regexp.new(pattern)
@@performance_monitor = 
PerformanceMonitor.new

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Util

assert, assert_equal, assert_equal_all, assert_non_nil, assert_not_reached, benchmark, command_exist?, commify, eprintf, format_bytes, program_name, protect_from_signals, require_command, set_verbosity, shell_escape, unix?, vprintf, windows?, wprintf

Constructor Details

#initialize(config, dbm, source_uri, normalized_path, content, options = {}) ⇒ Indexer

Returns a new instance of Indexer.



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/gonzui/indexer.rb', line 41

def initialize(config, dbm, source_uri, normalized_path, content, 
               options = {})
  @config = config
  @dbm = dbm
  @normalized_path = normalized_path
  @source_uri = source_uri
  @content = content
  @content_hash = Digest::MD5.hexdigest(content.text)
  @noindex_formats = (options[:noindex_formats] or @config.noindex_formats)

  @package_name = get_package_name
  @seqno      = 0

  @word_cache = {}
  @wordinfo_cache = {}
  @digest_cache = []

  # to be initialized
  @format_id  = nil
  @license_id = nil
  @license_abbrev = nil
  @encoding = nil
  @nlines = nil
  @package_id = nil
  @path_id = nil
  @bols = [] # positions of beginning of lines
  @indexed_p = false

  initialize_profilers_if_necessary
end

Class Method Details

.statistics ⇒ Object 



24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
 
# File 'lib/gonzui/indexer.rb', line 24

def self.statistics 
  return "" if @@performance_monitor.empty?
  pm = @@performance_monitor
  summary = "Performance statistics:\n"
  summary << pm.heading
  summary << pm.format([Indexer, :index],
                       [Indexer, :read_content],
                       [Indexer, :add_license],
                       [Indexer, :index_content])
  labels = LangScan.modules.map {|m|
    [m, :scan]
  }.push([Indexer, :add_fragment],
         [Indexer, :flush_cache])
  summary << pm.format([Indexer, :index_content], *labels)
  return summary
end

Instance Method Details

#add_binary_content ⇒ Object 



276
277
278
 
# File 'lib/gonzui/indexer.rb', line 276

def add_binary_content
  add_content_common("binary", "Binary")
end

#add_content ⇒ Object 



300
301
302
303
304
305
306
 
# File 'lib/gonzui/indexer.rb', line 300

def add_content
  if @encoding == "binary"
    add_binary_content
  else
    add_content_with_indexing
  end
end

#add_content_common(format_abbrev, format_name) ⇒ Object 



265
266
267
268
269
270
271
272
273
274
 
# File 'lib/gonzui/indexer.rb', line 265

def add_content_common(format_abbrev, format_name)
  add_format(format_abbrev, format_name)
  add_license
  @dbm.pathid_pkgid[@path_id] = @package_id
  @dbm.pathid_content[@path_id] = @content.text
  @dbm.pathid_info[@path_id] = make_content_info
  @dbm.pathid_hash[@path_id] = @content_hash
  vprintf("added (%s): %s (%s)", format_abbrev, 
          @normalized_path, @license_abbrev)
end

#add_content_with_indexing ⇒ Object 



290
291
292
293
294
295
296
297
298
 
# File 'lib/gonzui/indexer.rb', line 290

def add_content_with_indexing
  scanner = make_scanner
  if indexable?(scanner)
    index_content(scanner)
  else
    vprintf("skip indexing: %s", @normalized_path)
  end
  add_content_common(scanner.abbrev, scanner.name)
end

#add_format(format_abbrev, format_name) ⇒ Object 



184
185
186
187
188
189
190
 
# File 'lib/gonzui/indexer.rb', line 184

def add_format(format_abbrev, format_name)
  @format_id = add_property(format_abbrev, 
                            format_name,
                            :format_id_counter,
                            :make_ncontents_by_format_key,
                            :pkgid_fmtids)
end

#add_fragment(fragment) ⇒ Object 



150
151
152
153
154
155
156
157
158
159
 
# File 'lib/gonzui/indexer.rb', line 150

def add_fragment(fragment)
  type_id = @dbm.get_type_id(fragment.type)
  if LangScan::Type.splittable?(fragment.type)
    add_text(fragment, type_id)
  else
    add_word(fragment.text, fragment.byteno, type_id)
  end

  @digest_cache.push(fragment.byteno, fragment.text.length, type_id)
end

#add_license ⇒ Object 



192
193
194
195
196
197
198
199
200
201
 
# File 'lib/gonzui/indexer.rb', line 192

def add_license
  detector = LicenseDetector.new(@content.text)
  license = detector.detect
  @license_id = add_property(license.abbrev, 
                             license.name,
                             :license_id_counter,
                             :make_ncontents_by_license_key,
                             :pkgid_lcsids)
  @license_abbrev = license.abbrev
end

#add_package_if_necessary ⇒ Object 



228
229
230
231
232
233
234
235
236
237
238
 
# File 'lib/gonzui/indexer.rb', line 228

def add_package_if_necessary
  if @dbm.has_package?(@package_name)
    @package_id = @dbm.get_package_id(@package_name)
  else
    @package_id = @dbm.package_id_counter.make_new_id
    @dbm.pkg_pkgid[@package_name] = @package_id
    @dbm.pkgid_pkg[@package_id] = @package_name
    @dbm.pkgid_src[@package_id] = @source_uri.to_s
    @dbm.put_package_options(@package_id)
  end
end

#add_path ⇒ Object 



203
204
205
206
207
208
209
 
# File 'lib/gonzui/indexer.rb', line 203

def add_path
  assert_equal(false, @dbm.path_pathid.include?(@normalized_path))
  @path_id = @dbm.path_id_counter.make_new_id
  @dbm.path_pathid[@normalized_path] = @path_id
  @dbm.pathid_path[@path_id] = @normalized_path
  @dbm.pkgid_pathids[@package_id] = @path_id
end

#add_property(abbrev, name, counter, make_key, pkgid_ids) ⇒ Object 



177
178
179
180
181
182
 
# File 'lib/gonzui/indexer.rb', line 177

def add_property(abbrev, name, counter, make_key, pkgid_ids)
  id = @dbm.send(counter).get_id2(abbrev, name)
  @dbm.send(pkgid_ids)[@package_id] = id
  @dbm.increase_counter(@dbm.send(make_key, abbrev))
  return id
end

#add_text(fragment, type_id) ⇒ Object 



142
143
144
145
146
147
148
 
# File 'lib/gonzui/indexer.rb', line 142

def add_text(fragment, type_id)
  text = fragment.text
  byteno = fragment.byteno
  TextTokenizer.each_word(text) {|word, pos|
    add_word(word, byteno + pos, type_id)
  }
end

#add_word(word, byteno, type_id) ⇒ Object 



221
222
223
224
225
226
 
# File 'lib/gonzui/indexer.rb', line 221

def add_word(word, byteno, type_id)
  word_id = @dbm.word_id_counter.get_id(word)
  array = (@wordinfo_cache[word_id] ||= [])
  array.push(@seqno, byteno, type_id)
  @seqno += 1
end

#binary_content?(content) ⇒ Boolean

Returns:

  • (Boolean) 


105
106
107
 
# File 'lib/gonzui/indexer.rb', line 105

def binary_content?(content)
  BinaryRegexp.match(content)
end

#convert_to_utf8(content) ⇒ Object 



109
110
111
112
113
114
115
116
117
118
119
 
# File 'lib/gonzui/indexer.rb', line 109

def convert_to_utf8(content)
  encoding = "ascii"
  if binary_content?(content)
    encoding = "binary"
  else 
    if @config.utf8
      content, encoding = UTF8.to_utf8(content)
    end
  end
  return content, encoding
end

#flush_cache ⇒ Object 



161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
 
# File 'lib/gonzui/indexer.rb', line 161

def flush_cache
  all_word_ids = @wordinfo_cache.keys.sort!
  all_word_ids.each {|word_id|
    path_word_id = AutoPack.pack_id2(@path_id, word_id)
    @dbm.pathwordid_info[path_word_id] = 
      DeltaDumper.dump_tuples(WordInfo, @wordinfo_cache[word_id])
  }
  @dbm.put_pathid_wordids(@package_id, @path_id, all_word_ids)
  @dbm.pathid_wordids[@path_id] = DeltaDumper.dump_ids(all_word_ids)
  @dbm.pathid_digest[@path_id] = 
    DeltaDumper.dump_tuples(DigestInfo, @digest_cache)
  @dbm.pathid_bols[@path_id] = DeltaDumper.dump_fixnums(@bols)
  @wordinfo_cache.clear
  @dbm.word_id_counter.flush
end

#get_fragments(scanner) ⇒ Object 



211
212
213
214
215
216
217
218
219
 
# File 'lib/gonzui/indexer.rb', line 211

def get_fragments(scanner)
  @@performance_monitor.profile(scanner, :scan) if @config.verbose
  fragments = []
  scanner.scan(@content.text) {|fragment|
    fragments.push(fragment) if LangScan::Type.include?(fragment.type)
  }
  fragments = fragments.sort_by {|fragment| fragment.byteno }
  return fragments
end

#get_package_name ⇒ Object 



130
131
132
133
134
135
136
137
138
139
140
 
# File 'lib/gonzui/indexer.rb', line 130

def get_package_name
  parts = @normalized_path.split("/")
  if parts.length < 2
    raise IndexerError.new("normalized path should not be flat")
  end
  package_name = parts.first
  if package_name.size == 0 || package_name == "." || package_name == ".."
    package_name = File.basename(@source_uri.path)
  end
  return package_name
end

#index ⇒ Object 



309
310
311
312
313
314
 
# File 'lib/gonzui/indexer.rb', line 309

def index
  read_content
  add_package_if_necessary
  add_path
  add_content
end

#index_content(scanner) ⇒ Object 



246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
 
# File 'lib/gonzui/indexer.rb', line 246

def index_content(scanner)
  fragments = []
  begin
    fragments = get_fragments(scanner)
  rescue 
    # fallback to the text scanner
    unless scanner == LangScan::Text
      vprintf("#{@normalized_path}: fallback to LangScan::Text")
      scanner = LangScan::Text
      retry
    end
  end
  fragments.each {|fragment| add_fragment(fragment) }
  flush_cache
  @dbm.increase_counter(:ncontents_indexed)
  @dbm.increase_counter(:nlines_indexed, @nlines)
  @indexed_p = true
end

#indexable?(scanner) ⇒ Boolean

Returns:

  • (Boolean) 


286
287
288
 
# File 'lib/gonzui/indexer.rb', line 286

def indexable?(scanner)
  not @noindex_formats.include?(scanner.abbrev)
end

#initialize_profilers_if_necessary ⇒ Object 



72
73
74
75
76
77
78
79
80
81
82
 
# File 'lib/gonzui/indexer.rb', line 72

def initialize_profilers_if_necessary
  # profiler
  if @config.verbose
    @@performance_monitor.profile(Indexer, :index)
    @@performance_monitor.profile(Indexer, :read_content)
    @@performance_monitor.profile(Indexer, :index_content)
    @@performance_monitor.profile(Indexer, :add_fragment)
    @@performance_monitor.profile(Indexer, :add_license)
    @@performance_monitor.profile(Indexer, :flush_cache)
  end
end

#make_content_info ⇒ Object 



240
241
242
243
244
 
# File 'lib/gonzui/indexer.rb', line 240

def make_content_info
  ContentInfo.dump(@content.length, @content.mtime.to_i,
                   Time.now.to_i, @format_id, @license_id,
                   @nlines, @indexed_p)
end

#make_scanner ⇒ Object 



280
281
282
283
284
 
# File 'lib/gonzui/indexer.rb', line 280

def make_scanner
  scanner = LangScan.choose(@normalized_path, @content.text)
  scanner = LangScan::Text if scanner.nil?
  return scanner
end

#normalize_content(content) ⇒ Object 



121
122
123
124
125
126
127
128
 
# File 'lib/gonzui/indexer.rb', line 121

def normalize_content(content)
  content, encoding = convert_to_utf8(content)
  unless encoding == "binary"
    content = content.untabify
    content.gsub!(/\r\n?/, "\n")
  end
  return content, encoding
end

#read_content ⇒ Object 



84
85
86
87
88
89
90
91
92
93
94
 
# File 'lib/gonzui/indexer.rb', line 84

def read_content
  content, @encoding = normalize_content(@content.text)
  @content.text = content
  @nlines = 0
  pos = 0
  @content.text.each_line {|line| 
    @bols.push(pos)
    @nlines += 1
    pos += line.length
  }
end