path: root/Doc/lib/libdifflib.tex

\section{\module{difflib} ---
         Helpers for computing deltas}

\declaremodule{standard}{difflib}
\modulesynopsis{Helpers for computing differences between objects.}
\moduleauthor{Tim Peters}{tim.one@home.com}
\sectionauthor{Tim Peters}{tim.one@home.com}
% LaTeXification by Fred L. Drake, Jr. <fdrake@acm.org>.

\versionadded{2.1}


\begin{funcdesc}{get_close_matches}{word, possibilities\optional{,
                 n\optional{, cutoff}}}
  Return a list of the best ``good enough'' matches.  \var{word} is a
  sequence for which close matches are desired (typically a string),
  and \var{possibilities} is a list of sequences against which to
  match \var{word} (typically a list of strings).

  Optional argument \var{n} (default \code{3}) is the maximum number
  of close matches to return; \var{n} must be greater than \code{0}.

  Optional argument \var{cutoff} (default \code{0.6}) is a float in
  the range [0, 1].  Possibilities that don't score at least that
  similar to \var{word} are ignored.

  The best (no more than \var{n}) matches among the possibilities are
  returned in a list, sorted by similarity score, most similar first.

\begin{verbatim}
>>> get_close_matches('appel', ['ape', 'apple', 'peach', 'puppy'])
['apple', 'ape']
>>> import keyword
>>> get_close_matches('wheel', keyword.kwlist)
['while']
>>> get_close_matches('apple', keyword.kwlist)
[]
>>> get_close_matches('accept', keyword.kwlist)
['except']
\end{verbatim}
\end{funcdesc}

\begin{classdesc}{SequenceMatcher}{\unspecified}
  This is a flexible class for comparing pairs of sequences of any
  type, so long as the sequence elements are hashable.  The basic
  algorithm predates, and is a little fancier than, an algorithm
  published in the late 1980's by Ratcliff and Obershelp under the
  hyperbolic name ``gestalt pattern matching.''  The idea is to find
  the longest contiguous matching subsequence that contains no
  ``junk'' elements (the Ratcliff and Obershelp algorithm doesn't
  address junk).  The same idea is then applied recursively to the
  pieces of the sequences to the left and to the right of the matching
  subsequence.  This does not yield minimal edit sequences, but does
  tend to yield matches that ``look right'' to people.

  \strong{Timing:} The basic Ratcliff-Obershelp algorithm is cubic
  time in the worst case and quadratic time in the expected case.
  \class{SequenceMatcher} is quadratic time for the worst case and has
  expected-case behavior dependent in a complicated way on how many
  elements the sequences have in common; best case time is linear.
\end{classdesc}


\begin{seealso}
  \seetitle{Pattern Matching: The Gestalt Approach}{Discussion of a
            similar algorithm by John W. Ratcliff and D. E. Metzener.
            This was published in
            \citetitle[http://www.ddj.com/]{Dr. Dobb's Journal} in
            July, 1988.}
\end{seealso}


\subsection{SequenceMatcher Objects \label{sequence-matcher}}

\begin{classdesc}{SequenceMatcher}{\optional{isjunk\optional{,
                                   a\optional{, b}}}}
  Optional argument \var{isjunk} must be \code{None} (the default) or
  a one-argument function that takes a sequence element and returns
  true if and only if the element is ``junk'' and should be ignored.
  \code{None} is equivalent to passing \code{lambda x: 0}, i.e.\ no
  elements are ignored.  For example, pass

\begin{verbatim}
lambda x: x in " \t"
\end{verbatim}

  if you're comparing lines as sequences of characters, and don't want
  to synch up on blanks or hard tabs.

  The optional arguments \var{a} and \var{b} are sequences to be
  compared; both default to empty strings.  The elements of both
  sequences must be hashable.
\end{classdesc}


\class{SequenceMatcher} objects have the following methods:

\begin{methoddesc}{set_seqs}{a, b}
  Set the two sequences to be compared.
\end{methoddesc}

\class{SequenceMatcher} computes and caches detailed information about
the second sequence, so if you want to compare one sequence against
many sequences, use \method{set_seq2()} to set the commonly used
sequence once and call \method{set_seq1()} repeatedly, once for each
of the other sequences.

\begin{methoddesc}{set_seq1}{a}
  Set the first sequence to be compared.  The second sequence to be
  compared is not changed.
\end{methoddesc}

\begin{methoddesc}{set_seq2}{b}
  Set the second sequence to be compared.  The first sequence to be
  compared is not changed.
\end{methoddesc}

\begin{methoddesc}{find_longest_match}{alo, ahi, blo, bhi}
  Find longest matching block in \code{\var{a}[\var{alo}:\var{ahi}]}
  and \code{\var{b}[\var{blo}:\var{bhi}]}.

  If \var{isjunk} was omitted or \code{None},
  \method{get_longest_match()} returns \code{(\var{i}, \var{j},
  \var{k})} such that \code{\var{a}[\var{i}:\var{i}+\var{k}]} is equal
  to \code{\var{b}[\var{j}:\var{j}+\var{k}]}, where 
      \code{\var{alo} <= \var{i} <= \var{i}+\var{k} <= \var{ahi}} and
      \code{\var{blo} <= \var{j} <= \var{j}+\var{k} <= \var{bhi}}.
  For all \code{(\var{i'}, \var{j'}, \var{k'})} meeting those
  conditions, the additional conditions
      \code{\var{k} >= \var{k'}},
      \code{\var{i} <= \var{i'}},
      and if \code{\var{i} == \var{i'}}, \code{\var{j} <= \var{j'}}
  are also met.
  In other words, of all maximal matching blocks, return one that
  starts earliest in \var{a}, and of all those maximal matching blocks
  that start earliest in \var{a}, return the one that starts earliest
  in \var{b}.

\begin{verbatim}
>>> s = SequenceMatcher(None, " abcd", "abcd abcd")
>>> s.find_longest_match(0, 5, 0, 9)
(0, 4, 5)
\end{verbatim}

  If \var{isjunk} was provided, first the longest matching block is
  determined as above, but with the additional restriction that no
  junk element appears in the block.  Then that block is extended as
  far as possible by matching (only) junk elements on both sides.
  So the resulting block never matches on junk except as identical
  junk happens to be adjacent to an interesting match.

  Here's the same example as before, but considering blanks to be junk.
  That prevents \code{' abcd'} from matching the \code{' abcd'} at the
  tail end of the second sequence directly.  Instead only the
  \code{'abcd'} can match, and matches the leftmost \code{'abcd'} in
  the second sequence:

\begin{verbatim}
>>> s = SequenceMatcher(lambda x: x==" ", " abcd", "abcd abcd")
>>> s.find_longest_match(0, 5, 0, 9)
(1, 0, 4)
\end{verbatim}

  If no blocks match, this returns \code{(\var{alo}, \var{blo}, 0)}.
\end{methoddesc}

\begin{methoddesc}{get_matching_blocks}{}
  Return list of triples describing matching subsequences.
  Each triple is of the form \code{(\var{i}, \var{j}, \var{n})}, and
  means that \code{\var{a}[\var{i}:\var{i}+\var{n}] ==
  \var{b}[\var{j}:\var{j}+\var{n}]}.  The triples are monotonically
  increasing in \var{i} and \var{j}.

  The last triple is a dummy, and has the value \code{(len(\var{a}),
  len(\var{b}), 0)}.  It is the only triple with \code{\var{n} == 0}.
  % Explain why a dummy is used!

\begin{verbatim}
>>> s = SequenceMatcher(None, "abxcd", "abcd")
>>> s.get_matching_blocks()
[(0, 0, 2), (3, 2, 2), (5, 4, 0)]
\end{verbatim}
\end{methoddesc}

\begin{methoddesc}{get_opcodes}{}
  Return list of 5-tuples describing how to turn \var{a} into \var{b}.
  Each tuple is of the form \code{(\var{tag}, \var{i1}, \var{i2},
  \var{j1}, \var{j2})}.  The first tuple has \code{\var{i1} ==
  \var{j1} == 0}, and remaining tuples have \var{i1} equal to the
  \var{i2} from the preceeding tuple, and, likewise, \var{j1} equal to
  the previous \var{j2}.

  The \var{tag} values are strings, with these meanings:

\begin{tableii}{l|l}{code}{Value}{Meaning}
  \lineii{'replace'}{\code{\var{a}[\var{i1}:\var{i2}]} should be
                     replaced by \code{\var{b}[\var{j1}:\var{j2}]}.}
  \lineii{'delete'}{\code{\var{a}[\var{i1}:\var{i2}]} should be
                    deleted.  Note that \code{\var{j1} == \var{j2}} in
                    this case.}
  \lineii{'insert'}{\code{\var{b}[\var{j1}:\var{j2}]} should be
                    inserted at \code{\var{a}[\var{i1}:\var{i1}]}. 
                    Note that \code{\var{i1} == \var{i2}} in this
                    case.}
  \lineii{'equal'}{\code{\var{a}[\var{i1}:\var{i2}] ==
                   \var{b}[\var{j1}:\var{j2}]} (the sub-sequences are
                   equal).}
\end{tableii}

For example:

\begin{verbatim}
>>> a = "qabxcd"
>>> b = "abycdf"
>>> s = SequenceMatcher(None, a, b)
>>> for tag, i1, i2, j1, j2 in s.get_opcodes():
...    print ("%7s a[%d:%d] (%s) b[%d:%d] (%s)" %
...           (tag, i1, i2, a[i1:i2], j1, j2, b[j1:j2]))
 delete a[0:1] (q) b[0:0] ()
  equal a[1:3] (ab) b[0:2] (ab)
replace a[3:4] (x) b[2:3] (y)
  equal a[4:6] (cd) b[3:5] (cd)
 insert a[6:6] () b[5:6] (f)
\end{verbatim}
\end{methoddesc}

\begin{methoddesc}{ratio}{}
  Return a measure of the sequences' similarity as a float in the
  range [0, 1].

  Where T is the total number of elements in both sequences, and M is
  the number of matches, this is 2.0*M / T. Note that this is \code{1.}
  if the sequences are identical, and \code{0.} if they have nothing in
  common.

  This is expensive to compute if \method{get_matching_blocks()} or
  \method{get_opcodes()} hasn't already been called, in which case you
  may want to try \method{quick_ratio()} or
  \method{real_quick_ratio()} first to get an upper bound.
\end{methoddesc}

\begin{methoddesc}{quick_ratio}{}
  Return an upper bound on \method{ratio()} relatively quickly.

  This isn't defined beyond that it is an upper bound on
  \method{ratio()}, and is faster to compute.
\end{methoddesc}

\begin{methoddesc}{real_quick_ratio}{}
  Return an upper bound on \method{ratio()} very quickly.

  This isn't defined beyond that it is an upper bound on
  \method{ratio()}, and is faster to compute than either
  \method{ratio()} or \method{quick_ratio()}.
\end{methoddesc}

The three methods that return the ratio of matching to total characters
can give different results due to differing levels of approximation,
although \method{quick_ratio()} and \method{real_quick_ratio()} are always
at least as large as \method{ratio()}:

\begin{verbatim}
>>> s = SequenceMatcher(None, "abcd", "bcde")
>>> s.ratio()
0.75
>>> s.quick_ratio()
0.75
>>> s.real_quick_ratio()
1.0
\end{verbatim}


\subsection{Examples \label{difflib-examples}}


This example compares two strings, considering blanks to be ``junk:''

\begin{verbatim}
>>> s = SequenceMatcher(lambda x: x == " ",
...                     "private Thread currentThread;",
...                     "private volatile Thread currentThread;")
\end{verbatim}

\method{ratio()} returns a float in [0, 1], measuring the similarity
of the sequences.  As a rule of thumb, a \method{ratio()} value over
0.6 means the sequences are close matches:

\begin{verbatim}
>>> print round(s.ratio(), 3)
0.866
\end{verbatim}

If you're only interested in where the sequences match,
\method{get_matching_blocks()} is handy:

\begin{verbatim}
>>> for block in s.get_matching_blocks():
...     print "a[%d] and b[%d] match for %d elements" % block
a[0] and b[0] match for 8 elements
a[8] and b[17] match for 6 elements
a[14] and b[23] match for 15 elements
a[29] and b[38] match for 0 elements
\end{verbatim}

Note that the last tuple returned by \method{get_matching_blocks()} is
always a dummy, \code{(len(\var{a}), len(\var{b}), 0)}, and this is
the only case in which the last tuple element (number of elements
matched) is \code{0}.

If you want to know how to change the first sequence into the second,
use \method{get_opcodes()}:

\begin{verbatim}
>>> for opcode in s.get_opcodes():
...     print "%6s a[%d:%d] b[%d:%d]" % opcode
 equal a[0:8] b[0:8]
insert a[8:8] b[8:17]
 equal a[8:14] b[17:23]
 equal a[14:29] b[23:38]
\end{verbatim}

See \file{Tools/scripts/ndiff.py} from the Python source distribution
for a fancy human-friendly file differencer, which uses
\class{SequenceMatcher} both to view files as sequences of lines, and
lines as sequences of characters.

See also the function \function{get_close_matches()} in this module,
which shows how simple code building on \class{SequenceMatcher} can be
used to do useful work.