Perl开发：如何按Bit拆分Description文本并将段落末尾语句存入另一数组？

Got it, let's break down how to solve this Perl task. You need two core functionalities: parsing the <Description> text to map bit numbers (including ranges) to their descriptions, and extracting the trailing standalone paragraph into a separate array. Here's a robust implementation:

First, we'll use a proper XML parser (instead of fragile regex) to safely extract the <Description> content, then process the text to split out bit entries and the trailing paragraph.

use strict;
use warnings;
use XML::LibXML;

# Your input XML fragment (expand with full content as needed)
my $xml_input = q{
<Register>
  <Name>abc</Name>
  <Abstract></Abstract>
  <Description>Bit 6 random description Bit 5 msg octet 2 Bit 4-1 Bit 0 msg octet 4 These registers contain part of the Upstream Message. They should be written only after being cleared by hardware. </Description>
  <Field>
    <Name>qwe</Name>
    <Description>...</Description>
  </Field>
</Register>
};

# Step 1: Extract the Description text via XML parsing
my $xml_parser = XML::LibXML->new();
my $doc = $xml_parser->parse_string($xml_input);
my $desc_content = $doc->findvalue('//Register/Description');

# Step 2: Initialize storage structures
my %bit_descriptions;  # Maps individual bit numbers to their text
my @trailing_sentences; # Stores the non-bit-related paragraph sentences

# Step 3: Split content into bit entries and trailing text
my @all_segments = $desc_content =~ /(Bit \d+(?:-\d+)? .*?)(?= Bit |$)/g;

# Separate the trailing paragraph (it won't start with "Bit")
if (@all_segments && $all_segments[-1] !~ /^Bit/) {
    my $trailing_text = pop @all_segments;
    # Split into sentences, clean up whitespace and empty entries
    @trailing_sentences = grep { $_ ne '' } map { 
        s/^\s+|\s+$//g; $_ 
    } split /(?<=[.!?])\s+/, $trailing_text;
}

# Step 4: Parse bit entries (handle single bits and ranges)
foreach my $segment (@all_segments) {
    $segment =~ s/^\s+|\s+$//g; # Trim extra whitespace
    next unless $segment =~ /^Bit (\d+(?:-\d+)?)(?: (.*))?$/;
    
    my $bit_spec = $1;
    my $desc_text = $2 // ''; # Default to empty string if no description
    
    # Handle bit ranges (e.g., Bit 4-1)
    if ($bit_spec =~ /^(\d+)-(\d+)$/) {
        my ($start, $end) = ($1, $2);
        # Cover all bits in the range, regardless of order
        my ($min_bit, $max_bit) = sort { $a <=> $b } ($start, $end);
        for my $bit ($min_bit .. $max_bit) {
            $bit_descriptions{$bit} = $desc_text;
        }
    } else {
        # Single bit entry
        $bit_descriptions{$bit_spec} = $desc_text;
    }
}

# Verify the results (remove this block in production)
print "### Bit Mapping Results:\n";
foreach my $bit (sort { $b <=> $a } keys %bit_descriptions) {
    print "Bit $bit: ", $bit_descriptions{$bit} || "No description provided\n";
}

print "\n### Trailing Paragraph Sentences:\n";
foreach my $sentence (@trailing_sentences) {
    print "- $sentence\n";
}

• XML Parsing: We use XML::LibXML because regex can't reliably handle XML structure (e.g., nested tags, whitespace variations). If you don't have it installed, run cpanm XML::LibXML to get it.
• Bit Range Handling: The code handles ranges like 4-1 by sorting the start/end values, ensuring we capture every bit in between—this works even if docs list higher bits first (common in hardware specs).
• Trailing Paragraph Split: The regex /(?<=[.!?])\s+/ splits sentences on whitespace following punctuation, which is more robust than just splitting on . . Adjust this if your text uses unusual sentence endings.
• Edge Cases: The code handles empty descriptions (like Bit 4-1 with no text) and trims extra whitespace from all segments.

• If your XML has multiple <Register> tags, wrap the processing logic in a loop over $doc->findnodes('//Register').
• Adjust the bit matching regex if your docs use uppercase BIT instead of Bit, or other formatting (e.g., Bit[6]).

内容的提问来源于stack exchange，提问作者Nee