Line data    Source code

       1             : //===-- llvm/ADT/edit_distance.h - Array edit distance function --- C++ -*-===//
       2             : //
       3             : //                     The LLVM Compiler Infrastructure
       4             : //
       5             : // This file is distributed under the University of Illinois Open Source
       6             : // License. See LICENSE.TXT for details.
       7             : //
       8             : //===----------------------------------------------------------------------===//
       9             : //
      10             : // This file defines a Levenshtein distance function that works for any two
      11             : // sequences, with each element of each sequence being analogous to a character
      12             : // in a string.
      13             : //
      14             : //===----------------------------------------------------------------------===//
      15             : 
      16             : #ifndef LLVM_ADT_EDIT_DISTANCE_H
      17             : #define LLVM_ADT_EDIT_DISTANCE_H
      18             : 
      19             : #include "llvm/ADT/ArrayRef.h"
      20             : #include <algorithm>
      21             : #include <memory>
      22             : 
      23             : namespace llvm {
      24             : 
      25             : /// Determine the edit distance between two sequences.
      26             : ///
      27             : /// \param FromArray the first sequence to compare.
      28             : ///
      29             : /// \param ToArray the second sequence to compare.
      30             : ///
      31             : /// \param AllowReplacements whether to allow element replacements (change one
      32             : /// element into another) as a single operation, rather than as two operations
      33             : /// (an insertion and a removal).
      34             : ///
      35             : /// \param MaxEditDistance If non-zero, the maximum edit distance that this
      36             : /// routine is allowed to compute. If the edit distance will exceed that
      37             : /// maximum, returns \c MaxEditDistance+1.
      38             : ///
      39             : /// \returns the minimum number of element insertions, removals, or (if
      40             : /// \p AllowReplacements is \c true) replacements needed to transform one of
      41             : /// the given sequences into the other. If zero, the sequences are identical.
      42             : template<typename T>
      43     2041854 : unsigned ComputeEditDistance(ArrayRef<T> FromArray, ArrayRef<T> ToArray,
      44             :                              bool AllowReplacements = true,
      45             :                              unsigned MaxEditDistance = 0) {
      46             :   // The algorithm implemented below is the "classic"
      47             :   // dynamic-programming algorithm for computing the Levenshtein
      48             :   // distance, which is described here:
      49             :   //
      50             :   //   http://en.wikipedia.org/wiki/Levenshtein_distance
      51             :   //
      52             :   // Although the algorithm is typically described using an m x n
      53             :   // array, only one row plus one element are used at a time, so this
      54             :   // implementation just keeps one vector for the row.  To update one entry,
      55             :   // only the entries to the left, top, and top-left are needed.  The left
      56             :   // entry is in Row[x-1], the top entry is what's in Row[x] from the last
      57             :   // iteration, and the top-left entry is stored in Previous.
      58     2041854 :   typename ArrayRef<T>::size_type m = FromArray.size();
      59     2041854 :   typename ArrayRef<T>::size_type n = ToArray.size();
      60             : 
      61             :   const unsigned SmallBufferSize = 64;
      62             :   unsigned SmallBuffer[SmallBufferSize];
      63             :   std::unique_ptr<unsigned[]> Allocated;
      64             :   unsigned *Row = SmallBuffer;
      65     2041854 :   if (n + 1 > SmallBufferSize) {
      66        8622 :     Row = new unsigned[n + 1];
      67             :     Allocated.reset(Row);
      68             :   }
      69             : 
      70    26548409 :   for (unsigned i = 1; i <= n; ++i)
      71    24506555 :     Row[i] = i;
      72             : 
      73    15194594 :   for (typename ArrayRef<T>::size_type y = 1; y <= m; ++y) {
      74    14869118 :     Row[0] = y;
      75    14869118 :     unsigned BestThisRow = Row[0];
      76             : 
      77    14869118 :     unsigned Previous = y - 1;
      78   323650513 :     for (typename ArrayRef<T>::size_type x = 1; x <= n; ++x) {
      79   308781395 :       int OldRow = Row[x];
      80   308781395 :       if (AllowReplacements) {
      81   303217303 :         Row[x] = std::min(
      82   606434606 :             Previous + (FromArray[y-1] == ToArray[x-1] ? 0u : 1u),
      83   467854951 :             std::min(Row[x-1], Row[x])+1);
      84             :       }
      85             :       else {
      86    16692276 :         if (FromArray[y-1] == ToArray[x-1]) Row[x] = Previous;
      87     5727612 :         else Row[x] = std::min(Row[x-1], Row[x]) + 1;
      88             :       }
      89             :       Previous = OldRow;
      90   308781395 :       BestThisRow = std::min(BestThisRow, Row[x]);
      91             :     }
      92             : 
      93    14869118 :     if (MaxEditDistance && BestThisRow > MaxEditDistance)
      94     1716378 :       return MaxEditDistance + 1;
      95             :   }
      96             : 
      97      325476 :   unsigned Result = Row[n];
      98      325476 :   return Result;
      99             : }
     100             : 
     101             : } // End llvm namespace
     102             : 
     103             : #endif