Module: Sequence::StringLike

Included in:
File, IO, OfString
Defined in:
lib/sequence/stringlike.rb

Defined Under Namespace

Classes: CorrectedMatchData

Constant Summary collapse

FFS_4BITTABLE = 
[nil,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0]
@@anchor_cache = 
{}

Instance Method Summary collapse

Instance Method Details

#_anchor(str, backwards = false, cache = true) ⇒ Object

add an anchor to a Regexp-string. normally,



216
217
218
219
220
221
222
223
 
# File 'lib/sequence/stringlike.rb', line 216

def _anchor(str,backwards=false,cache=true)
  cache and result=@@anchor_cache[[str,backwards]] and return result
  result=backwards ? "(?:#{str})\\Z" : "\\A(?:#{str})"
  cache and return @@anchor_cache[[str,backwards]]||=Regexp.new( result, 0, 'n' )
    #is it really correct to force regexp to be binary encoding here? not sure
    #maybe encoding of str should be detected and reused?
  return result
end

#data_class ⇒ Object 



12
 
# File 'lib/sequence/stringlike.rb', line 12

def data_class; String end

#ffs ⇒ Object 



18
19
20
21
22
23
24
25
26
27
 
# File 'lib/sequence/stringlike.rb', line 18

def ffs
  holding{
    begin!
    zeros=read_til_charset(/[^\0]/)
    byte=read1
    lo=byte&0xF
    rem=FFS_4BITTABLE[lo]||FFS_4BITTABLE[byte>>4]+4
    return zeros.size<<3+rem
  }
end

#fixup_match_result(matchdata, addedgroups, pos_adjust, namelet, &body) ⇒ Object 



226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
 
# File 'lib/sequence/stringlike.rb', line 226

def fixup_match_result(matchdata,addedgroups,pos_adjust,namelet,&body)

  #remove extra capture results from () we inserted from MatchData
  #..first extract groups, begin and end idxs from old
  groups=matchdata.to_a
  begins=[]
  ends=[]
  matchdata.to_a.each_with_index{|substr,i| 
    if substr
      begins<<matchdata.begin(i)+pos_adjust
      ends<<matchdata.end(i)+pos_adjust
    else
      begins<<nil
      ends<<nil
    end
  }
  
  #..remove data at group indexes we added above
  addedgroups.reverse_each{|groupidx| 
    [groups,begins,ends].each{|arr| arr.delete_at groupidx }
  }
  
  #..now change matchdata to use fixed-up arrays
  result=CorrectedMatchData.new
  result.begins=begins
  result.ends=ends
  result.groups=groups
  if namelet==:pre
    result.set_pre_match_body( &body)
    result.set_post_match_body {matchdata.post_match}
  else
    result.set_pre_match_body {matchdata.pre_match}
    result.set_post_match_body( &body)
  end
  result.pos=pos_adjust
  
  result
end

#fns(bitnum) ⇒ Object 



30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
 
# File 'lib/sequence/stringlike.rb', line 30

def fns(bitnum)
  holding{
    goto bitnum>>3
    bitnum&=0x7
    byte=read1
    byte&=~((1<<(bitnum+1))-1)
    if byte.nonzero?
      zeros_size=0
    else
      zeros_size=read_til_charset(/[^\0]/).size
      byte=read1
    end
    lo=byte&0xF
    rem=FFS_4BITTABLE[lo]||FFS_4BITTABLE[byte>>4]+4
    return zeros_size<<3+rem
  }
end

#group_anchors(rex, backwards, anchored = false) ⇒ Object

if not backwards: replace Z with (?!) replace $ with (?=n) if backwards: replace A with (?!) replace ^ with (^) (and adjust addedgroups) there’s no lookback in ruby regexp (yet) so, ^ in reverse regexp will perhaps lead to unexpected results. some matches with ^ in them will fail, when they should have succeeded even if the ^ couldn’t match. you should be pretty much ok if you don’t use ^ within alternation (|) in backwards match. if anchored, an implicit anchor is added at the end (begin if backwards) there’s also a nice cache,so that the cost of regexp rebuilding is reduced returns: the modified regex and addedgroups



166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
 
# File 'lib/sequence/stringlike.rb', line 166

def group_anchors(rex,backwards,anchored=false)
  @@fs_cache||={}
  result=@@fs_cache[[rex,backwards,anchored]] and return result
  if backwards 
    caret,dollar,buffanchor='^',nil,'A'
  else 
    caret,dollar,buffanchor=nil,'$','Z' 
  end
  newrex=(anchored ?  _anchor(rex,backwards,false) : rex.to_s)

  rewritten=incclass=false
  groupnum=0
  addedgroups=[]
  result=''
  (frags=newrex.split( /((?:[^\\(\[\]$^]+|\\(?:[CM]-)*[^CMZA])*)/ )).each_index{|i|
    frag=frags[i]
    case frag
      when "\\"
        if !incclass and frags[i+1][0,1]==buffanchor
          frags[i+1].slice! 0
          frag='(?!)'
          rewritten=true
        end
      when caret 
        unless incclass
          addedgroups<<(groupnum+=1)
          frag="(^)"
          rewritten=true
        end
      when dollar 
        unless incclass
          frag="(?=\n)"
          rewritten=true
        end
      when "("; incclass or frags[i+1][0]==?? or groupnum+=1
      when "["; incclass=true #ignore stuff til ]
      when "]"; incclass=false #stop ignoring stuff
    end
    result<<frag
  }
  
  newrex=rewritten ? Regexp.new(result) : rex
  
  @@fs_cache[[rex,backwards,anchored]]=[newrex,addedgroups]
end

#index(pat, pos = 0) ⇒ Object 



427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
 
# File 'lib/sequence/stringlike.rb', line 427

def index pat,pos=0
  posi= self.begin()
  until_buffer_len=4*maxmatchlen(false)
  if Regexp===pat
    until_step_len=3*maxmatchlen(false)
      until posi.eof?
        if m=posi.match(pat,false,until_buffer_len)
          range=0...m.begin(0)
          pre=subseq(range)
          m.set_pre_match_body { pre } 
          self.last_match=Thread.current[:last_match]=m
          return m.begin(0) #return match and what preceded it
        end
        posi.move until_step_len
      end
  #elsif SetOfChar===pat; ...
  else
    until_step_len=until_buffer_len
    String===pat and until_step_len-=pat.size-1          
      until posi.eof?
        buf=posi.readahead(until_buffer_len)
        if i=buf.index( pat)
          result=posi.pos+i
          return result
        end
        posi.move until_step_len
      end
  end
  return nil
ensure
  posi.close
end

#like ⇒ Object 



14
 
# File 'lib/sequence/stringlike.rb', line 14

def like; StringLike end

#match(rex, anchored = true, len = maxmatchlen(false)) ⇒ Object

like match_fast, but anchors work correctly and post_match is set to something, if not exactly what you expected. (an Sequence, not String.) 2nd parameter determines if match is anchored on the left side to the current position or not.



130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
 
# File 'lib/sequence/stringlike.rb', line 130

def match(rex,anchored=true, len=maxmatchlen(false))
  
  newrex=nearend(len)? rex : group_anchors(rex,false,false).first

  #do the match against what input we have
  matchdata=match_fast(newrex,false,len)

  if !matchdata or anchored && matchdata.begin(0).nonzero?
    self.last_match=Thread.current[:last_match]=nil
    return
  end
  posi=position;posi.move matchdata.end(0)
  result=fixup_match_result(matchdata,[],pos,:post) { posi.subseq(posi.pos..-1) }
      #note: post_match is a SubSeq

      #rex.last_match=
      self.last_match=Thread.current[:last_match]=result
end

#match_fast(rex, backwards = false, len = maxmatchlen(backwards)) ⇒ Object

this version is fast and simple, but anchors do not work right, matches are NOT implicitly anchored to the current position, and the file position is not advanced. post_match (or pre_match if going backwards) is always nil.



76
77
78
79
80
81
82
83
84
85
86
 
# File 'lib/sequence/stringlike.rb', line 76

def match_fast(rex,backwards=false,len=maxmatchlen(backwards))
  str=send backwards ? :readbehind  : :readahead, len
    if result=rex.match(str)
      if backwards
        def result.pre_match; end
      else
        def result.post_match ; end
      end
    end
  return result
end

#matchback(rex, anchored = true, len = maxmatchlen(true)) ⇒ Object

like match, but goes backwards



92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
 
# File 'lib/sequence/stringlike.rb', line 92

def matchback(rex,anchored=true, len=maxmatchlen(true))
  nearbegin=nearbegin(len) 
  newrex,addedgroups=
  if nearbegin && !anchored
   [rex,[]]
  else group_anchors(rex,:back,anchored)
  end
  #do the match against what input we have
  
  matchdata=match_fast(newrex,true,len)
  #fail if any  ^ or \A matched at begin of buffer, 
  #but buffer isn't begin of file
  if !matchdata or #not actually a match
    addedgroups.find{|i| matchdata.end(i)==0 } && !nearbegin
      self.last_match=Thread.current[:last_match]=nil
      return
  end

  matchpos=pos-len
  matchpos>=0 or matchpos=0
  assert(matchpos>=0)
  match1st=position matchpos+matchdata.begin(0)
  result=fixup_match_result(matchdata,addedgroups,matchpos,:pre) do
        result=SubSeq.new(self,0,match1st.pos)
        result.pos=match1st.pos
        result
      end
      #note: pre_match is a subseq.

      #rex.last_match=
      self.last_match=Thread.current[:last_match]=result
end

#new_data ⇒ Object

Return an empty object used for returning a sequence of elements. The only method required of this object is << (append to the sequence). usually [] or “”



10
 
# File 'lib/sequence/stringlike.rb', line 10

def new_data; '' end

#push(str) ⇒ Object 



417
418
419
420
 
# File 'lib/sequence/stringlike.rb', line 417

def push(str)
  Integer===str and str=str.chr
  insert size, str
end

#read_til_charset(charrex, blocksize = 16) ⇒ Object

read until a character in a user-supplied set is found. charrex must be a regexp that contains only a single character class



51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
 
# File 'lib/sequence/stringlike.rb', line 51

def read_til_charset(charrex,blocksize=16)
   blocks=[]
   m=nil
   until eof?
      block=read blocksize
      #if near eof, less than a full block may have been read
  
      if m=charrex .match(block)
         self.pos-=m.post_match.length+1
         #'self.' shouldn't be needed... but is
  
         blocks<<m.pre_match if m.pre_match.length>0
         break
      end
      blocks<<block
   end
   return blocks.join
end

#rindex(pat, pos = size-1) ⇒ Object 



460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
 
# File 'lib/sequence/stringlike.rb', line 460

def rindex pat,pos=size-1
  posi= self.end()
  until_buffer_len=4*maxmatchlen(false)
  if Regexp===pat
    until_step_len=3*maxmatchlen(false)
      until posi.pos.zero?
        if m=posi.matchback(pat,false,until_buffer_len)
          range=m.end(0)+1..-1
          post=subseq(range)
          m.set_post_match_body { post } 
          self.last_match=Thread.current[:last_match]=m
          posi.close
          return m.begin(0) #return match and what preceded it
        end
        posi.move( -until_step_len )
      end
  #elsif SetOfChar===pat; ...
  else
    until_step_len=until_buffer_len
    String===pat and until_step_len-=pat.size-1          
      until posi.pos.zero?
        buf=posi.readbehind(until_buffer_len)
        if i=buf.rindex( pat)
          result=posi.pos-until_buffer_len+i
          posi.close
          return result
        end
        posi.move( -until_step_len )
      end
  end
  return nil
ensure
  posi.close
end

#scan(pat) ⇒ Object 



311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
 
# File 'lib/sequence/stringlike.rb', line 311

def scan(pat)
  holding? {case pat
    when Integer 
      pat==read1 and pat.chr
    #when SetOfChar; ...
    when String
      pat==read(pat.size) and pat
    when Regexp 
      if m=match(pat,true)
        goto m.end(0) 
        m.to_s
      end
    else raise ArgumentError.new("bad scan pattern for Sequence::StringLike")
  end}
end

#scan_until(pat) ⇒ Object 



343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
 
# File 'lib/sequence/stringlike.rb', line 343

def scan_until(pat)
  at=index( pat,pos) or return
  newpos=case pat
    when Regexp 
      m=last_match
      s=slice(pos...m.begin(0))
      m.set_pre_match_body{s}
      m.end(0)
    when String; at+pat.size
    when Integer; at+1
    #when SetOfChar; huh
    else raise ArgumentError
  end
  return( read newpos-pos)

=begin    
  holding? {
    if Regexp===pat
      until_buffer_len=4*maxmatchlen(false)
      until_step_len=3*maxmatchlen(false)
      holding_position{|posi|
        until posi.eof?
          if m=posi.match(pat,false,until_buffer_len)
            pre=read(posi.pos-pos)+m.pre_match
            m.set_prematch_body {pre} #readjust matchdata to include data between my own pos and posi
            goto m.end(0)  #advance my own position to end of match
            return m.pre_match+m.to_s #return match and what preceded it
          end
          posi.move until_step_len
        end
        nil
      }
    #elsif SetOfChar===pat; ...
    else #string or integer
      i=index(pat,pos)
      result=read(i-pos)<<pat
      move(pat.is_a?( Integer ) ? 1 : pat.size)
      result
    end
  }
=end
end

#scanback(pat) ⇒ Object 



327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
 
# File 'lib/sequence/stringlike.rb', line 327

def scanback(pat)
  holding? {case pat
    when Integer 
      pat==readback1 and pat.chr
    #when SetOfChar; ...
    when String
      pat==readback(pat.size) and pat
    when Regexp 
      if m=matchback(pat,true) 
        goto m.begin(0) 
        m.to_s
      end
    else raise ArgumentError.new("bad scan pattern for Sequence::StringLike")
  end}
end

#scanback_until(pat) ⇒ Object 



386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
 
# File 'lib/sequence/stringlike.rb', line 386

def scanback_until(pat)
  at=rindex( pat,pos) or return
  newpos=
    if Regexp===pat
      m=last_match
      s=slice(m.end(0)+1..pos)
      m.set_post_match_body{s}
      m.begin(0)
    else at
    end
  assert(newpos<=pos)
  return( readback pos-newpos)

=begin
  holding? {
    if Regexp===pat
      huh #need to scan til eof, like #scan_until does
      m=matchback(pat,false) or break
      goto= m.begin(0)
      m.to_s+m.post_match
    #elsif SetOfChar===pat; ...
    else #string or integer
      i=rindex(pat,pos)
      result=readback(pos-i-pat.size)<<pat
      move( -(pat.is_a? Integer ? 1 : pat.size))
      result
    end
  }
=end
end

#unshift(str) ⇒ Object 



422
423
424
425
 
# File 'lib/sequence/stringlike.rb', line 422

def unshift(str)
  Integer===str and str=str.chr
  insert 0, str
end