正则表达式从字典哈希中简单替换文档(Perl)

Question

我需要尽快从大型文档的散列中查找和替换关键字。我厌倦了以下两种方法，一种速度提高了 320%，但我确信我这样做是错误的，并且确信有更好的方法来做到这一点。

我只想替换字典哈希中存在的关键字并保留那些不存在的关键字，这样我就知道它不在字典中。

以下两种方法都扫描两次以查找和替换我认为的。我确信像向前或向后看这样的正则表达式可以更快地优化它。

#!/usr/bin/perl

use strict;
use warnings;

use Benchmark qw(:all);

my %dictionary = (
            pollack => "pollard",
            polynya => "polyoma",
            pomaces => "pomaded",
            pomades => "pomatum",
            practic => "praetor",
            prairie => "praised",
            praiser => "praises",
            prajnas => "praline",
            quakily => "quaking",
            qualify => "quality",
            quamash => "quangos",
            quantal => "quanted", 
            quantic => "quantum",
    );

my $content =qq{
        Start this is the text that contains the words to replace. {quantal} A computer {pollack} is a general {pomaces} purpose device {practic} that 
        can be {quakily} programmed to carry out a set {quantic} of arithmetic or logical operations automatically {quamash}.
        Since a {prajnas} sequence of operations can {praiser} be readily changed, the computer {pomades} can solve more than {prairie}
        one kind of problem {qualify} {doesNotExist} end.
    };

# just duplicate content many times
$content .= $content;

cmpthese(100000, {
    replacer_1 => sub {my $text = replacer1($content)},
    replacer_2 => sub {my $text = replacer2($content)},
});

print replacer1($content) , "\n--------------------------\n";
print replacer2($content) , "\n--------------------------\n";
exit;

sub replacer1 {
    my ($content) = shift;
    $content =~ s/\{(.+?)\}/exists $dictionary{$1} ? "[$dictionary{$1}]": "\{$1\}"/gex;
    return $content;
}

sub replacer2 {
    my ($content) = shift;
    my @names = $content =~ /\{(.+?)\}/g;
    foreach my $name (@names) {
        if (exists $dictionary{$name}) {
            $content =~ s/\{$name\}/\[$dictionary{$name}\]/;
        }
    }
    return $content;
}

这是基准测试结果:

              Rate replacer_2 replacer_1
replacer_2  5565/s         --       -76%
replacer_1 23397/s       320%         --

Answer 1

这是一种更快更紧凑的方法:

sub replacer3 {
    my ($content) = shift;
    $content =~ s#\{(.+?)\}#"[".($dictionary{$1} // $1)."]"#ge;
    return $content;
}

在 Perl 5.8 中，如果您的字典值都不是“false”，则可以使用 || 而不是 //。

使用已经包含大括号和方括号的字典还有一些好处:

sub replacer5 {
    my ($content) = shift;
    our %dict2;
    if (!%dict2) {
        %dict2 = map { "{".$_."}" => "[".$dictionary{$_}."]" } keys %dictionary
    }
    $content =~ s#(\{.+?\})#$dict2{$1} || $1#ge;
    return $content;
}

基准测试结果:

              Rate replacer_2 replacer_1 replacer_3 replacer_5
replacer_2  2908/s         --       -79%       -83%       -84%
replacer_1 14059/s       383%         --       -20%       -25%
replacer_3 17513/s       502%        25%         --        -7%
replacer_5 18741/s       544%        33%         7%         --