Source file src/index/suffixarray/sais.go

     1  // Copyright 2019 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  // Suffix array construction by induced sorting (SAIS).
     6  // See Ge Nong, Sen Zhang, and Wai Hong Chen,
     7  // "Two Efficient Algorithms for Linear Time Suffix Array Construction",
     8  // especially section 3 (https://ieeexplore.ieee.org/document/5582081).
     9  // See also http://zork.net/~st/jottings/sais.html.
    10  //
    11  // With optimizations inspired by Yuta Mori's sais-lite
    12  // (https://sites.google.com/site/yuta256/sais).
    13  //
    14  // And with other new optimizations.
    15  
    16  // Many of these functions are parameterized by the sizes of
    17  // the types they operate on. The generator gen.go makes
    18  // copies of these functions for use with other sizes.
    19  // Specifically:
    20  //
    21  // - A function with a name ending in _8_32 takes []byte and []int32 arguments
    22  //   and is duplicated into _32_32, _8_64, and _64_64 forms.
    23  //   The _32_32 and _64_64_ suffixes are shortened to plain _32 and _64.
    24  //   Any lines in the function body that contain the text "byte-only" or "256"
    25  //   are stripped when creating _32_32 and _64_64 forms.
    26  //   (Those lines are typically 8-bit-specific optimizations.)
    27  //
    28  // - A function with a name ending only in _32 operates on []int32
    29  //   and is duplicated into a _64 form. (Note that it may still take a []byte,
    30  //   but there is no need for a version of the function in which the []byte
    31  //   is widened to a full integer array.)
    32  
    33  // The overall runtime of this code is linear in the input size:
    34  // it runs a sequence of linear passes to reduce the problem to
    35  // a subproblem at most half as big, invokes itself recursively,
    36  // and then runs a sequence of linear passes to turn the answer
    37  // for the subproblem into the answer for the original problem.
    38  // This gives T(N) = O(N) + T(N/2) = O(N) + O(N/2) + O(N/4) + ... = O(N).
    39  //
    40  // The outline of the code, with the forward and backward scans
    41  // through O(N)-sized arrays called out, is:
    42  //
    43  // sais_I_N
    44  //	placeLMS_I_B
    45  //		bucketMax_I_B
    46  //			freq_I_B
    47  //				<scan +text> (1)
    48  //			<scan +freq> (2)
    49  //		<scan -text, random bucket> (3)
    50  //	induceSubL_I_B
    51  //		bucketMin_I_B
    52  //			freq_I_B
    53  //				<scan +text, often optimized away> (4)
    54  //			<scan +freq> (5)
    55  //		<scan +sa, random text, random bucket> (6)
    56  //	induceSubS_I_B
    57  //		bucketMax_I_B
    58  //			freq_I_B
    59  //				<scan +text, often optimized away> (7)
    60  //			<scan +freq> (8)
    61  //		<scan -sa, random text, random bucket> (9)
    62  //	assignID_I_B
    63  //		<scan +sa, random text substrings> (10)
    64  //	map_B
    65  //		<scan -sa> (11)
    66  //	recurse_B
    67  //		(recursive call to sais_B_B for a subproblem of size at most 1/2 input, often much smaller)
    68  //	unmap_I_B
    69  //		<scan -text> (12)
    70  //		<scan +sa> (13)
    71  //	expand_I_B
    72  //		bucketMax_I_B
    73  //			freq_I_B
    74  //				<scan +text, often optimized away> (14)
    75  //			<scan +freq> (15)
    76  //		<scan -sa, random text, random bucket> (16)
    77  //	induceL_I_B
    78  //		bucketMin_I_B
    79  //			freq_I_B
    80  //				<scan +text, often optimized away> (17)
    81  //			<scan +freq> (18)
    82  //		<scan +sa, random text, random bucket> (19)
    83  //	induceS_I_B
    84  //		bucketMax_I_B
    85  //			freq_I_B
    86  //				<scan +text, often optimized away> (20)
    87  //			<scan +freq> (21)
    88  //		<scan -sa, random text, random bucket> (22)
    89  //
    90  // Here, _B indicates the suffix array size (_32 or _64) and _I the input size (_8 or _B).
    91  //
    92  // The outline shows there are in general 22 scans through
    93  // O(N)-sized arrays for a given level of the recursion.
    94  // In the top level, operating on 8-bit input text,
    95  // the six freq scans are fixed size (256) instead of potentially
    96  // input-sized. Also, the frequency is counted once and cached
    97  // whenever there is room to do so (there is nearly always room in general,
    98  // and always room at the top level), which eliminates all but
    99  // the first freq_I_B text scans (that is, 5 of the 6).
   100  // So the top level of the recursion only does 22 - 6 - 5 = 11
   101  // input-sized scans and a typical level does 16 scans.
   102  //
   103  // The linear scans do not cost anywhere near as much as
   104  // the random accesses to the text made during a few of
   105  // the scans (specifically #6, #9, #16, #19, #22 marked above).
   106  // In real texts, there is not much but some locality to
   107  // the accesses, due to the repetitive structure of the text
   108  // (the same reason Burrows-Wheeler compression is so effective).
   109  // For random inputs, there is no locality, which makes those
   110  // accesses even more expensive, especially once the text
   111  // no longer fits in cache.
   112  // For example, running on 50 MB of Go source code, induceSubL_8_32
   113  // (which runs only once, at the top level of the recursion)
   114  // takes 0.44s, while on 50 MB of random input, it takes 2.55s.
   115  // Nearly all the relative slowdown is explained by the text access:
   116  //
   117  //		c0, c1 := text[k-1], text[k]
   118  //
   119  // That line runs for 0.23s on the Go text and 2.02s on random text.
   120  
   121  //go:generate go run gen.go
   122  
   123  package suffixarray
   124  
   125  // text_32 returns the suffix array for the input text.
   126  // It requires that len(text) fit in an int32
   127  // and that the caller zero sa.
   128  func text_32(text []byte, sa []int32) {
   129  	if int(int32(len(text))) != len(text) || len(text) != len(sa) {
   130  		panic("suffixarray: misuse of text_32")
   131  	}
   132  	sais_8_32(text, 256, sa, make([]int32, 2*256))
   133  }
   134  
   135  // sais_8_32 computes the suffix array of text.
   136  // The text must contain only values in [0, textMax).
   137  // The suffix array is stored in sa, which the caller
   138  // must ensure is already zeroed.
   139  // The caller must also provide temporary space tmp
   140  // with len(tmp) ≥ textMax. If len(tmp) ≥ 2*textMax
   141  // then the algorithm runs a little faster.
   142  // If sais_8_32 modifies tmp, it sets tmp[0] = -1 on return.
   143  func sais_8_32(text []byte, textMax int, sa, tmp []int32) {
   144  	if len(sa) != len(text) || len(tmp) < textMax {
   145  		panic("suffixarray: misuse of sais_8_32")
   146  	}
   147  
   148  	// Trivial base cases. Sorting 0 or 1 things is easy.
   149  	if len(text) == 0 {
   150  		return
   151  	}
   152  	if len(text) == 1 {
   153  		sa[0] = 0
   154  		return
   155  	}
   156  
   157  	// Establish slices indexed by text character
   158  	// holding character frequency and bucket-sort offsets.
   159  	// If there's only enough tmp for one slice,
   160  	// we make it the bucket offsets and recompute
   161  	// the character frequency each time we need it.
   162  	var freq, bucket []int32
   163  	if len(tmp) >= 2*textMax {
   164  		freq, bucket = tmp[:textMax], tmp[textMax:2*textMax]
   165  		freq[0] = -1 // mark as uninitialized
   166  	} else {
   167  		freq, bucket = nil, tmp[:textMax]
   168  	}
   169  
   170  	// The SAIS algorithm.
   171  	// Each of these calls makes one scan through sa.
   172  	// See the individual functions for documentation
   173  	// about each's role in the algorithm.
   174  	numLMS := placeLMS_8_32(text, sa, freq, bucket)
   175  	if numLMS <= 1 {
   176  		// 0 or 1 items are already sorted. Do nothing.
   177  	} else {
   178  		induceSubL_8_32(text, sa, freq, bucket)
   179  		induceSubS_8_32(text, sa, freq, bucket)
   180  		length_8_32(text, sa, numLMS)
   181  		maxID := assignID_8_32(text, sa, numLMS)
   182  		if maxID < numLMS {
   183  			map_32(sa, numLMS)
   184  			recurse_32(sa, tmp, numLMS, maxID)
   185  			unmap_8_32(text, sa, numLMS)
   186  		} else {
   187  			// If maxID == numLMS, then each LMS-substring
   188  			// is unique, so the relative ordering of two LMS-suffixes
   189  			// is determined by just the leading LMS-substring.
   190  			// That is, the LMS-suffix sort order matches the
   191  			// (simpler) LMS-substring sort order.
   192  			// Copy the original LMS-substring order into the
   193  			// suffix array destination.
   194  			copy(sa, sa[len(sa)-numLMS:])
   195  		}
   196  		expand_8_32(text, freq, bucket, sa, numLMS)
   197  	}
   198  	induceL_8_32(text, sa, freq, bucket)
   199  	induceS_8_32(text, sa, freq, bucket)
   200  
   201  	// Mark for caller that we overwrote tmp.
   202  	tmp[0] = -1
   203  }
   204  
   205  // freq_8_32 returns the character frequencies
   206  // for text, as a slice indexed by character value.
   207  // If freq is nil, freq_8_32 uses and returns bucket.
   208  // If freq is non-nil, freq_8_32 assumes that freq[0] >= 0
   209  // means the frequencies are already computed.
   210  // If the frequency data is overwritten or uninitialized,
   211  // the caller must set freq[0] = -1 to force recomputation
   212  // the next time it is needed.
   213  func freq_8_32(text []byte, freq, bucket []int32) []int32 {
   214  	if freq != nil && freq[0] >= 0 {
   215  		return freq // already computed
   216  	}
   217  	if freq == nil {
   218  		freq = bucket
   219  	}
   220  
   221  	freq = freq[:256] // eliminate bounds check for freq[c] below
   222  	clear(freq)
   223  	for _, c := range text {
   224  		freq[c]++
   225  	}
   226  	return freq
   227  }
   228  
   229  // bucketMin_8_32 stores into bucket[c] the minimum index
   230  // in the bucket for character c in a bucket-sort of text.
   231  func bucketMin_8_32(text []byte, freq, bucket []int32) {
   232  	freq = freq_8_32(text, freq, bucket)
   233  	freq = freq[:256]     // establish len(freq) = 256, so 0 ≤ i < 256 below
   234  	bucket = bucket[:256] // eliminate bounds check for bucket[i] below
   235  	total := int32(0)
   236  	for i, n := range freq {
   237  		bucket[i] = total
   238  		total += n
   239  	}
   240  }
   241  
   242  // bucketMax_8_32 stores into bucket[c] the maximum index
   243  // in the bucket for character c in a bucket-sort of text.
   244  // The bucket indexes for c are [min, max).
   245  // That is, max is one past the final index in that bucket.
   246  func bucketMax_8_32(text []byte, freq, bucket []int32) {
   247  	freq = freq_8_32(text, freq, bucket)
   248  	freq = freq[:256]     // establish len(freq) = 256, so 0 ≤ i < 256 below
   249  	bucket = bucket[:256] // eliminate bounds check for bucket[i] below
   250  	total := int32(0)
   251  	for i, n := range freq {
   252  		total += n
   253  		bucket[i] = total
   254  	}
   255  }
   256  
   257  // The SAIS algorithm proceeds in a sequence of scans through sa.
   258  // Each of the following functions implements one scan,
   259  // and the functions appear here in the order they execute in the algorithm.
   260  
   261  // placeLMS_8_32 places into sa the indexes of the
   262  // final characters of the LMS substrings of text,
   263  // sorted into the rightmost ends of their correct buckets
   264  // in the suffix array.
   265  //
   266  // The imaginary sentinel character at the end of the text
   267  // is the final character of the final LMS substring, but there
   268  // is no bucket for the imaginary sentinel character,
   269  // which has a smaller value than any real character.
   270  // The caller must therefore pretend that sa[-1] == len(text).
   271  //
   272  // The text indexes of LMS-substring characters are always ≥ 1
   273  // (the first LMS-substring must be preceded by one or more L-type
   274  // characters that are not part of any LMS-substring),
   275  // so using 0 as a “not present” suffix array entry is safe,
   276  // both in this function and in most later functions
   277  // (until induceL_8_32 below).
   278  func placeLMS_8_32(text []byte, sa, freq, bucket []int32) int {
   279  	bucketMax_8_32(text, freq, bucket)
   280  
   281  	numLMS := 0
   282  	lastB := int32(-1)
   283  	bucket = bucket[:256] // eliminate bounds check for bucket[c1] below
   284  
   285  	// The next stanza of code (until the blank line) loop backward
   286  	// over text, stopping to execute a code body at each position i
   287  	// such that text[i] is an L-character and text[i+1] is an S-character.
   288  	// That is, i+1 is the position of the start of an LMS-substring.
   289  	// These could be hoisted out into a function with a callback,
   290  	// but at a significant speed cost. Instead, we just write these
   291  	// seven lines a few times in this source file. The copies below
   292  	// refer back to the pattern established by this original as the
   293  	// "LMS-substring iterator".
   294  	//
   295  	// In every scan through the text, c0, c1 are successive characters of text.
   296  	// In this backward scan, c0 == text[i] and c1 == text[i+1].
   297  	// By scanning backward, we can keep track of whether the current
   298  	// position is type-S or type-L according to the usual definition:
   299  	//
   300  	//	- position len(text) is type S with text[len(text)] == -1 (the sentinel)
   301  	//	- position i is type S if text[i] < text[i+1], or if text[i] == text[i+1] && i+1 is type S.
   302  	//	- position i is type L if text[i] > text[i+1], or if text[i] == text[i+1] && i+1 is type L.
   303  	//
   304  	// The backward scan lets us maintain the current type,
   305  	// update it when we see c0 != c1, and otherwise leave it alone.
   306  	// We want to identify all S positions with a preceding L.
   307  	// Position len(text) is one such position by definition, but we have
   308  	// nowhere to write it down, so we eliminate it by untruthfully
   309  	// setting isTypeS = false at the start of the loop.
   310  	c0, c1, isTypeS := byte(0), byte(0), false
   311  	for i := len(text) - 1; i >= 0; i-- {
   312  		c0, c1 = text[i], c0
   313  		if c0 < c1 {
   314  			isTypeS = true
   315  		} else if c0 > c1 && isTypeS {
   316  			isTypeS = false
   317  
   318  			// Bucket the index i+1 for the start of an LMS-substring.
   319  			b := bucket[c1] - 1
   320  			bucket[c1] = b
   321  			sa[b] = int32(i + 1)
   322  			lastB = b
   323  			numLMS++
   324  		}
   325  	}
   326  
   327  	// We recorded the LMS-substring starts but really want the ends.
   328  	// Luckily, with two differences, the start indexes and the end indexes are the same.
   329  	// The first difference is that the rightmost LMS-substring's end index is len(text),
   330  	// so the caller must pretend that sa[-1] == len(text), as noted above.
   331  	// The second difference is that the first leftmost LMS-substring start index
   332  	// does not end an earlier LMS-substring, so as an optimization we can omit
   333  	// that leftmost LMS-substring start index (the last one we wrote).
   334  	//
   335  	// Exception: if numLMS <= 1, the caller is not going to bother with
   336  	// the recursion at all and will treat the result as containing LMS-substring starts.
   337  	// In that case, we don't remove the final entry.
   338  	if numLMS > 1 {
   339  		sa[lastB] = 0
   340  	}
   341  	return numLMS
   342  }
   343  
   344  // induceSubL_8_32 inserts the L-type text indexes of LMS-substrings
   345  // into sa, assuming that the final characters of the LMS-substrings
   346  // are already inserted into sa, sorted by final character, and at the
   347  // right (not left) end of the corresponding character bucket.
   348  // Each LMS-substring has the form (as a regexp) /S+L+S/:
   349  // one or more S-type, one or more L-type, final S-type.
   350  // induceSubL_8_32 leaves behind only the leftmost L-type text
   351  // index for each LMS-substring. That is, it removes the final S-type
   352  // indexes that are present on entry, and it inserts but then removes
   353  // the interior L-type indexes too.
   354  // (Only the leftmost L-type index is needed by induceSubS_8_32.)
   355  func induceSubL_8_32(text []byte, sa, freq, bucket []int32) {
   356  	// Initialize positions for left side of character buckets.
   357  	bucketMin_8_32(text, freq, bucket)
   358  	bucket = bucket[:256] // eliminate bounds check for bucket[cB] below
   359  
   360  	// As we scan the array left-to-right, each sa[i] = j > 0 is a correctly
   361  	// sorted suffix array entry (for text[j:]) for which we know that j-1 is type L.
   362  	// Because j-1 is type L, inserting it into sa now will sort it correctly.
   363  	// But we want to distinguish a j-1 with j-2 of type L from type S.
   364  	// We can process the former but want to leave the latter for the caller.
   365  	// We record the difference by negating j-1 if it is preceded by type S.
   366  	// Either way, the insertion (into the text[j-1] bucket) is guaranteed to
   367  	// happen at sa[i´] for some i´ > i, that is, in the portion of sa we have
   368  	// yet to scan. A single pass therefore sees indexes j, j-1, j-2, j-3,
   369  	// and so on, in sorted but not necessarily adjacent order, until it finds
   370  	// one preceded by an index of type S, at which point it must stop.
   371  	//
   372  	// As we scan through the array, we clear the worked entries (sa[i] > 0) to zero,
   373  	// and we flip sa[i] < 0 to -sa[i], so that the loop finishes with sa containing
   374  	// only the indexes of the leftmost L-type indexes for each LMS-substring.
   375  	//
   376  	// The suffix array sa therefore serves simultaneously as input, output,
   377  	// and a miraculously well-tailored work queue.
   378  
   379  	// placeLMS_8_32 left out the implicit entry sa[-1] == len(text),
   380  	// corresponding to the identified type-L index len(text)-1.
   381  	// Process it before the left-to-right scan of sa proper.
   382  	// See body in loop for commentary.
   383  	k := len(text) - 1
   384  	c0, c1 := text[k-1], text[k]
   385  	if c0 < c1 {
   386  		k = -k
   387  	}
   388  
   389  	// Cache recently used bucket index:
   390  	// we're processing suffixes in sorted order
   391  	// and accessing buckets indexed by the
   392  	// byte before the sorted order, which still
   393  	// has very good locality.
   394  	// Invariant: b is cached, possibly dirty copy of bucket[cB].
   395  	cB := c1
   396  	b := bucket[cB]
   397  	sa[b] = int32(k)
   398  	b++
   399  
   400  	for i := 0; i < len(sa); i++ {
   401  		j := int(sa[i])
   402  		if j == 0 {
   403  			// Skip empty entry.
   404  			continue
   405  		}
   406  		if j < 0 {
   407  			// Leave discovered type-S index for caller.
   408  			sa[i] = int32(-j)
   409  			continue
   410  		}
   411  		sa[i] = 0
   412  
   413  		// Index j was on work queue, meaning k := j-1 is L-type,
   414  		// so we can now place k correctly into sa.
   415  		// If k-1 is L-type, queue k for processing later in this loop.
   416  		// If k-1 is S-type (text[k-1] < text[k]), queue -k to save for the caller.
   417  		k := j - 1
   418  		c0, c1 := text[k-1], text[k]
   419  		if c0 < c1 {
   420  			k = -k
   421  		}
   422  
   423  		if cB != c1 {
   424  			bucket[cB] = b
   425  			cB = c1
   426  			b = bucket[cB]
   427  		}
   428  		sa[b] = int32(k)
   429  		b++
   430  	}
   431  }
   432  
   433  // induceSubS_8_32 inserts the S-type text indexes of LMS-substrings
   434  // into sa, assuming that the leftmost L-type text indexes are already
   435  // inserted into sa, sorted by LMS-substring suffix, and at the
   436  // left end of the corresponding character bucket.
   437  // Each LMS-substring has the form (as a regexp) /S+L+S/:
   438  // one or more S-type, one or more L-type, final S-type.
   439  // induceSubS_8_32 leaves behind only the leftmost S-type text
   440  // index for each LMS-substring, in sorted order, at the right end of sa.
   441  // That is, it removes the L-type indexes that are present on entry,
   442  // and it inserts but then removes the interior S-type indexes too,
   443  // leaving the LMS-substring start indexes packed into sa[len(sa)-numLMS:].
   444  // (Only the LMS-substring start indexes are processed by the recursion.)
   445  func induceSubS_8_32(text []byte, sa, freq, bucket []int32) {
   446  	// Initialize positions for right side of character buckets.
   447  	bucketMax_8_32(text, freq, bucket)
   448  	bucket = bucket[:256] // eliminate bounds check for bucket[cB] below
   449  
   450  	// Analogous to induceSubL_8_32 above,
   451  	// as we scan the array right-to-left, each sa[i] = j > 0 is a correctly
   452  	// sorted suffix array entry (for text[j:]) for which we know that j-1 is type S.
   453  	// Because j-1 is type S, inserting it into sa now will sort it correctly.
   454  	// But we want to distinguish a j-1 with j-2 of type S from type L.
   455  	// We can process the former but want to leave the latter for the caller.
   456  	// We record the difference by negating j-1 if it is preceded by type L.
   457  	// Either way, the insertion (into the text[j-1] bucket) is guaranteed to
   458  	// happen at sa[i´] for some i´ < i, that is, in the portion of sa we have
   459  	// yet to scan. A single pass therefore sees indexes j, j-1, j-2, j-3,
   460  	// and so on, in sorted but not necessarily adjacent order, until it finds
   461  	// one preceded by an index of type L, at which point it must stop.
   462  	// That index (preceded by one of type L) is an LMS-substring start.
   463  	//
   464  	// As we scan through the array, we clear the worked entries (sa[i] > 0) to zero,
   465  	// and we flip sa[i] < 0 to -sa[i] and compact into the top of sa,
   466  	// so that the loop finishes with the top of sa containing exactly
   467  	// the LMS-substring start indexes, sorted by LMS-substring.
   468  
   469  	// Cache recently used bucket index:
   470  	cB := byte(0)
   471  	b := bucket[cB]
   472  
   473  	top := len(sa)
   474  	for i := len(sa) - 1; i >= 0; i-- {
   475  		j := int(sa[i])
   476  		if j == 0 {
   477  			// Skip empty entry.
   478  			continue
   479  		}
   480  		sa[i] = 0
   481  		if j < 0 {
   482  			// Leave discovered LMS-substring start index for caller.
   483  			top--
   484  			sa[top] = int32(-j)
   485  			continue
   486  		}
   487  
   488  		// Index j was on work queue, meaning k := j-1 is S-type,
   489  		// so we can now place k correctly into sa.
   490  		// If k-1 is S-type, queue k for processing later in this loop.
   491  		// If k-1 is L-type (text[k-1] > text[k]), queue -k to save for the caller.
   492  		k := j - 1
   493  		c1 := text[k]
   494  		c0 := text[k-1]
   495  		if c0 > c1 {
   496  			k = -k
   497  		}
   498  
   499  		if cB != c1 {
   500  			bucket[cB] = b
   501  			cB = c1
   502  			b = bucket[cB]
   503  		}
   504  		b--
   505  		sa[b] = int32(k)
   506  	}
   507  }
   508  
   509  // length_8_32 computes and records the length of each LMS-substring in text.
   510  // The length of the LMS-substring at index j is stored at sa[j/2],
   511  // avoiding the LMS-substring indexes already stored in the top half of sa.
   512  // (If index j is an LMS-substring start, then index j-1 is type L and cannot be.)
   513  // There are two exceptions, made for optimizations in name_8_32 below.
   514  //
   515  // First, the final LMS-substring is recorded as having length 0, which is otherwise
   516  // impossible, instead of giving it a length that includes the implicit sentinel.
   517  // This ensures the final LMS-substring has length unequal to all others
   518  // and therefore can be detected as different without text comparison
   519  // (it is unequal because it is the only one that ends in the implicit sentinel,
   520  // and the text comparison would be problematic since the implicit sentinel
   521  // is not actually present at text[len(text)]).
   522  //
   523  // Second, to avoid text comparison entirely, if an LMS-substring is very short,
   524  // sa[j/2] records its actual text instead of its length, so that if two such
   525  // substrings have matching “length,” the text need not be read at all.
   526  // The definition of “very short” is that the text bytes must pack into a uint32,
   527  // and the unsigned encoding e must be ≥ len(text), so that it can be
   528  // distinguished from a valid length.
   529  func length_8_32(text []byte, sa []int32, numLMS int) {
   530  	end := 0 // index of current LMS-substring end (0 indicates final LMS-substring)
   531  
   532  	// The encoding of N text bytes into a “length” word
   533  	// adds 1 to each byte, packs them into the bottom
   534  	// N*8 bits of a word, and then bitwise inverts the result.
   535  	// That is, the text sequence A B C (hex 41 42 43)
   536  	// encodes as ^uint32(0x42_43_44).
   537  	// LMS-substrings can never start or end with 0xFF.
   538  	// Adding 1 ensures the encoded byte sequence never
   539  	// starts or ends with 0x00, so that present bytes can be
   540  	// distinguished from zero-padding in the top bits,
   541  	// so the length need not be separately encoded.
   542  	// Inverting the bytes increases the chance that a
   543  	// 4-byte encoding will still be ≥ len(text).
   544  	// In particular, if the first byte is ASCII (<= 0x7E, so +1 <= 0x7F)
   545  	// then the high bit of the inversion will be set,
   546  	// making it clearly not a valid length (it would be a negative one).
   547  	//
   548  	// cx holds the pre-inverted encoding (the packed incremented bytes).
   549  	cx := uint32(0) // byte-only
   550  
   551  	// This stanza (until the blank line) is the "LMS-substring iterator",
   552  	// described in placeLMS_8_32 above, with one line added to maintain cx.
   553  	c0, c1, isTypeS := byte(0), byte(0), false
   554  	for i := len(text) - 1; i >= 0; i-- {
   555  		c0, c1 = text[i], c0
   556  		cx = cx<<8 | uint32(c1+1) // byte-only
   557  		if c0 < c1 {
   558  			isTypeS = true
   559  		} else if c0 > c1 && isTypeS {
   560  			isTypeS = false
   561  
   562  			// Index j = i+1 is the start of an LMS-substring.
   563  			// Compute length or encoded text to store in sa[j/2].
   564  			j := i + 1
   565  			var code int32
   566  			if end == 0 {
   567  				code = 0
   568  			} else {
   569  				code = int32(end - j)
   570  				if code <= 32/8 && ^cx >= uint32(len(text)) { // byte-only
   571  					code = int32(^cx) // byte-only
   572  				} // byte-only
   573  			}
   574  			sa[j>>1] = code
   575  			end = j + 1
   576  			cx = uint32(c1 + 1) // byte-only
   577  		}
   578  	}
   579  }
   580  
   581  // assignID_8_32 assigns a dense ID numbering to the
   582  // set of LMS-substrings respecting string ordering and equality,
   583  // returning the maximum assigned ID.
   584  // For example given the input "ababab", the LMS-substrings
   585  // are "aba", "aba", and "ab", renumbered as 2 2 1.
   586  // sa[len(sa)-numLMS:] holds the LMS-substring indexes
   587  // sorted in string order, so to assign numbers we can
   588  // consider each in turn, removing adjacent duplicates.
   589  // The new ID for the LMS-substring at index j is written to sa[j/2],
   590  // overwriting the length previously stored there (by length_8_32 above).
   591  func assignID_8_32(text []byte, sa []int32, numLMS int) int {
   592  	id := 0
   593  	lastLen := int32(-1) // impossible
   594  	lastPos := int32(0)
   595  	for _, j := range sa[len(sa)-numLMS:] {
   596  		// Is the LMS-substring at index j new, or is it the same as the last one we saw?
   597  		n := sa[j/2]
   598  		if n != lastLen {
   599  			goto New
   600  		}
   601  		if uint32(n) >= uint32(len(text)) {
   602  			// “Length” is really encoded full text, and they match.
   603  			goto Same
   604  		}
   605  		{
   606  			// Compare actual texts.
   607  			n := int(n)
   608  			this := text[j:][:n]
   609  			last := text[lastPos:][:n]
   610  			for i := 0; i < n; i++ {
   611  				if this[i] != last[i] {
   612  					goto New
   613  				}
   614  			}
   615  			goto Same
   616  		}
   617  	New:
   618  		id++
   619  		lastPos = j
   620  		lastLen = n
   621  	Same:
   622  		sa[j/2] = int32(id)
   623  	}
   624  	return id
   625  }
   626  
   627  // map_32 maps the LMS-substrings in text to their new IDs,
   628  // producing the subproblem for the recursion.
   629  // The mapping itself was mostly applied by assignID_8_32:
   630  // sa[i] is either 0, the ID for the LMS-substring at index 2*i,
   631  // or the ID for the LMS-substring at index 2*i+1.
   632  // To produce the subproblem we need only remove the zeros
   633  // and change ID into ID-1 (our IDs start at 1, but text chars start at 0).
   634  //
   635  // map_32 packs the result, which is the input to the recursion,
   636  // into the top of sa, so that the recursion result can be stored
   637  // in the bottom of sa, which sets up for expand_8_32 well.
   638  func map_32(sa []int32, numLMS int) {
   639  	w := len(sa)
   640  	for i := len(sa) / 2; i >= 0; i-- {
   641  		j := sa[i]
   642  		if j > 0 {
   643  			w--
   644  			sa[w] = j - 1
   645  		}
   646  	}
   647  }
   648  
   649  // recurse_32 calls sais_32 recursively to solve the subproblem we've built.
   650  // The subproblem is at the right end of sa, the suffix array result will be
   651  // written at the left end of sa, and the middle of sa is available for use as
   652  // temporary frequency and bucket storage.
   653  func recurse_32(sa, oldTmp []int32, numLMS, maxID int) {
   654  	dst, saTmp, text := sa[:numLMS], sa[numLMS:len(sa)-numLMS], sa[len(sa)-numLMS:]
   655  
   656  	// Set up temporary space for recursive call.
   657  	// We must pass sais_32 a tmp buffer with at least maxID entries.
   658  	//
   659  	// The subproblem is guaranteed to have length at most len(sa)/2,
   660  	// so that sa can hold both the subproblem and its suffix array.
   661  	// Nearly all the time, however, the subproblem has length < len(sa)/3,
   662  	// in which case there is a subproblem-sized middle of sa that
   663  	// we can reuse for temporary space (saTmp).
   664  	// When recurse_32 is called from sais_8_32, oldTmp is length 512
   665  	// (from text_32), and saTmp will typically be much larger, so we'll use saTmp.
   666  	// When deeper recursions come back to recurse_32, now oldTmp is
   667  	// the saTmp from the top-most recursion, it is typically larger than
   668  	// the current saTmp (because the current sa gets smaller and smaller
   669  	// as the recursion gets deeper), and we keep reusing that top-most
   670  	// large saTmp instead of the offered smaller ones.
   671  	//
   672  	// Why is the subproblem length so often just under len(sa)/3?
   673  	// See Nong, Zhang, and Chen, section 3.6 for a plausible explanation.
   674  	// In brief, the len(sa)/2 case would correspond to an SLSLSLSLSLSL pattern
   675  	// in the input, perfect alternation of larger and smaller input bytes.
   676  	// Real text doesn't do that. If each L-type index is randomly followed
   677  	// by either an L-type or S-type index, then half the substrings will
   678  	// be of the form SLS, but the other half will be longer. Of that half,
   679  	// half (a quarter overall) will be SLLS; an eighth will be SLLLS, and so on.
   680  	// Not counting the final S in each (which overlaps the first S in the next),
   681  	// This works out to an average length 2×½ + 3×¼ + 4×⅛ + ... = 3.
   682  	// The space we need is further reduced by the fact that many of the
   683  	// short patterns like SLS will often be the same character sequences
   684  	// repeated throughout the text, reducing maxID relative to numLMS.
   685  	//
   686  	// For short inputs, the averages may not run in our favor, but then we
   687  	// can often fall back to using the length-512 tmp available in the
   688  	// top-most call. (Also a short allocation would not be a big deal.)
   689  	//
   690  	// For pathological inputs, we fall back to allocating a new tmp of length
   691  	// max(maxID, numLMS/2). This level of the recursion needs maxID,
   692  	// and all deeper levels of the recursion will need no more than numLMS/2,
   693  	// so this one allocation is guaranteed to suffice for the entire stack
   694  	// of recursive calls.
   695  	tmp := oldTmp
   696  	if len(tmp) < len(saTmp) {
   697  		tmp = saTmp
   698  	}
   699  	if len(tmp) < numLMS {
   700  		// TestSAIS/forcealloc reaches this code.
   701  		n := maxID
   702  		if n < numLMS/2 {
   703  			n = numLMS / 2
   704  		}
   705  		tmp = make([]int32, n)
   706  	}
   707  
   708  	// sais_32 requires that the caller arrange to clear dst,
   709  	// because in general the caller may know dst is
   710  	// freshly-allocated and already cleared. But this one is not.
   711  	clear(dst)
   712  	sais_32(text, maxID, dst, tmp)
   713  }
   714  
   715  // unmap_8_32 unmaps the subproblem back to the original.
   716  // sa[:numLMS] is the LMS-substring numbers, which don't matter much anymore.
   717  // sa[len(sa)-numLMS:] is the sorted list of those LMS-substring numbers.
   718  // The key part is that if the list says K that means the K'th substring.
   719  // We can replace sa[:numLMS] with the indexes of the LMS-substrings.
   720  // Then if the list says K it really means sa[K].
   721  // Having mapped the list back to LMS-substring indexes,
   722  // we can place those into the right buckets.
   723  func unmap_8_32(text []byte, sa []int32, numLMS int) {
   724  	unmap := sa[len(sa)-numLMS:]
   725  	j := len(unmap)
   726  
   727  	// "LMS-substring iterator" (see placeLMS_8_32 above).
   728  	c0, c1, isTypeS := byte(0), byte(0), false
   729  	for i := len(text) - 1; i >= 0; i-- {
   730  		c0, c1 = text[i], c0
   731  		if c0 < c1 {
   732  			isTypeS = true
   733  		} else if c0 > c1 && isTypeS {
   734  			isTypeS = false
   735  
   736  			// Populate inverse map.
   737  			j--
   738  			unmap[j] = int32(i + 1)
   739  		}
   740  	}
   741  
   742  	// Apply inverse map to subproblem suffix array.
   743  	sa = sa[:numLMS]
   744  	for i := 0; i < len(sa); i++ {
   745  		sa[i] = unmap[sa[i]]
   746  	}
   747  }
   748  
   749  // expand_8_32 distributes the compacted, sorted LMS-suffix indexes
   750  // from sa[:numLMS] into the tops of the appropriate buckets in sa,
   751  // preserving the sorted order and making room for the L-type indexes
   752  // to be slotted into the sorted sequence by induceL_8_32.
   753  func expand_8_32(text []byte, freq, bucket, sa []int32, numLMS int) {
   754  	bucketMax_8_32(text, freq, bucket)
   755  	bucket = bucket[:256] // eliminate bound check for bucket[c] below
   756  
   757  	// Loop backward through sa, always tracking
   758  	// the next index to populate from sa[:numLMS].
   759  	// When we get to one, populate it.
   760  	// Zero the rest of the slots; they have dead values in them.
   761  	x := numLMS - 1
   762  	saX := sa[x]
   763  	c := text[saX]
   764  	b := bucket[c] - 1
   765  	bucket[c] = b
   766  
   767  	for i := len(sa) - 1; i >= 0; i-- {
   768  		if i != int(b) {
   769  			sa[i] = 0
   770  			continue
   771  		}
   772  		sa[i] = saX
   773  
   774  		// Load next entry to put down (if any).
   775  		if x > 0 {
   776  			x--
   777  			saX = sa[x] // TODO bounds check
   778  			c = text[saX]
   779  			b = bucket[c] - 1
   780  			bucket[c] = b
   781  		}
   782  	}
   783  }
   784  
   785  // induceL_8_32 inserts L-type text indexes into sa,
   786  // assuming that the leftmost S-type indexes are inserted
   787  // into sa, in sorted order, in the right bucket halves.
   788  // It leaves all the L-type indexes in sa, but the
   789  // leftmost L-type indexes are negated, to mark them
   790  // for processing by induceS_8_32.
   791  func induceL_8_32(text []byte, sa, freq, bucket []int32) {
   792  	// Initialize positions for left side of character buckets.
   793  	bucketMin_8_32(text, freq, bucket)
   794  	bucket = bucket[:256] // eliminate bounds check for bucket[cB] below
   795  
   796  	// This scan is similar to the one in induceSubL_8_32 above.
   797  	// That one arranges to clear all but the leftmost L-type indexes.
   798  	// This scan leaves all the L-type indexes and the original S-type
   799  	// indexes, but it negates the positive leftmost L-type indexes
   800  	// (the ones that induceS_8_32 needs to process).
   801  
   802  	// expand_8_32 left out the implicit entry sa[-1] == len(text),
   803  	// corresponding to the identified type-L index len(text)-1.
   804  	// Process it before the left-to-right scan of sa proper.
   805  	// See body in loop for commentary.
   806  	k := len(text) - 1
   807  	c0, c1 := text[k-1], text[k]
   808  	if c0 < c1 {
   809  		k = -k
   810  	}
   811  
   812  	// Cache recently used bucket index.
   813  	cB := c1
   814  	b := bucket[cB]
   815  	sa[b] = int32(k)
   816  	b++
   817  
   818  	for i := 0; i < len(sa); i++ {
   819  		j := int(sa[i])
   820  		if j <= 0 {
   821  			// Skip empty or negated entry (including negated zero).
   822  			continue
   823  		}
   824  
   825  		// Index j was on work queue, meaning k := j-1 is L-type,
   826  		// so we can now place k correctly into sa.
   827  		// If k-1 is L-type, queue k for processing later in this loop.
   828  		// If k-1 is S-type (text[k-1] < text[k]), queue -k to save for the caller.
   829  		// If k is zero, k-1 doesn't exist, so we only need to leave it
   830  		// for the caller. The caller can't tell the difference between
   831  		// an empty slot and a non-empty zero, but there's no need
   832  		// to distinguish them anyway: the final suffix array will end up
   833  		// with one zero somewhere, and that will be a real zero.
   834  		k := j - 1
   835  		c1 := text[k]
   836  		if k > 0 {
   837  			if c0 := text[k-1]; c0 < c1 {
   838  				k = -k
   839  			}
   840  		}
   841  
   842  		if cB != c1 {
   843  			bucket[cB] = b
   844  			cB = c1
   845  			b = bucket[cB]
   846  		}
   847  		sa[b] = int32(k)
   848  		b++
   849  	}
   850  }
   851  
   852  func induceS_8_32(text []byte, sa, freq, bucket []int32) {
   853  	// Initialize positions for right side of character buckets.
   854  	bucketMax_8_32(text, freq, bucket)
   855  	bucket = bucket[:256] // eliminate bounds check for bucket[cB] below
   856  
   857  	cB := byte(0)
   858  	b := bucket[cB]
   859  
   860  	for i := len(sa) - 1; i >= 0; i-- {
   861  		j := int(sa[i])
   862  		if j >= 0 {
   863  			// Skip non-flagged entry.
   864  			// (This loop can't see an empty entry; 0 means the real zero index.)
   865  			continue
   866  		}
   867  
   868  		// Negative j is a work queue entry; rewrite to positive j for final suffix array.
   869  		j = -j
   870  		sa[i] = int32(j)
   871  
   872  		// Index j was on work queue (encoded as -j but now decoded),
   873  		// meaning k := j-1 is L-type,
   874  		// so we can now place k correctly into sa.
   875  		// If k-1 is S-type, queue -k for processing later in this loop.
   876  		// If k-1 is L-type (text[k-1] > text[k]), queue k to save for the caller.
   877  		// If k is zero, k-1 doesn't exist, so we only need to leave it
   878  		// for the caller.
   879  		k := j - 1
   880  		c1 := text[k]
   881  		if k > 0 {
   882  			if c0 := text[k-1]; c0 <= c1 {
   883  				k = -k
   884  			}
   885  		}
   886  
   887  		if cB != c1 {
   888  			bucket[cB] = b
   889  			cB = c1
   890  			b = bucket[cB]
   891  		}
   892  		b--
   893  		sa[b] = int32(k)
   894  	}
   895  }
   896  

View as plain text