python/highlighter.cpp

#include "highlighter.h"
#include <algorithm>
#include <regex>

std::string Highlighter::highlight_text(
    const std::string& text,
    const std::vector<std::string>& keywords,
    const std::string& pre_tag,
    const std::string& post_tag
) {
    auto spans = find_keyword_positions(text, keywords);
    std::string result = text;
    
    // 从后向前插入标签，避免位置偏移
    std::sort(spans.begin(), spans.end(),
        [](const HighlightSpan& a, const HighlightSpan& b) {
            return a.start > b.start;
        });
    
    for (const auto& span : spans) {
        result.insert(span.start + span.length, post_tag);
        result.insert(span.start, pre_tag);
    }
    
    return result;
}

std::vector<Highlighter::HighlightSpan> Highlighter::find_keyword_positions(
    const std::string& text,
    const std::vector<std::string>& keywords
) {
    std::vector<HighlightSpan> spans;
    
    for (const auto& keyword : keywords) {
        std::regex word_regex(keyword, std::regex::icase);
        auto words_begin = std::sregex_iterator(
            text.begin(), text.end(), word_regex
        );
        auto words_end = std::sregex_iterator();
        
        for (auto i = words_begin; i != words_end; ++i) {
            spans.push_back({
                static_cast<size_t>(i->position()),
                static_cast<size_t>(i->length())
            });
        }
    }
    
    return spans;
}