SlideCombine/BookmarkExtractor.cs

using System;
using System.Collections.Generic;
using System.IO;
using System.Text;

namespace SlideCombine
{
    public class BookmarkItem
    {
        public string Title { get; set; }
        public string Page { get; set; }
        public string FormattedContent { get; set; }
    }

    public class BookmarkExtractor
    {
        public static List<BookmarkItem> ExtractBookmarksFromBkmk(string bkmkFilePath)
        {
            var bookmarks = new List<BookmarkItem>();

            if (!File.Exists(bkmkFilePath))
            {
                throw new FileNotFoundException($"FreePic2Pdf_bkmk文件不存在: {bkmkFilePath}");
            }

            try
            {
                // 尝试用UTF-8读取，如果失败则用GBK
                string content;
                try
                {
                    content = File.ReadAllText(bkmkFilePath, Encoding.UTF8);
                }
                catch
                {
                    content = File.ReadAllText(bkmkFilePath, Encoding.GetEncoding("GBK"));
                }

                // 按行分割内容
                var lines = content.Split(new[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries);

                foreach (var line in lines)
                {
                    var trimmedLine = line.Trim();
                    if (string.IsNullOrEmpty(trimmedLine))
                        continue;

                    // 查找页码模式 - 假设格式为 "标题 页码" 或 "标题:页码"
                    var bookmark = ParseBookmarkLine(trimmedLine);
                    if (bookmark != null)
                    {
                        bookmarks.Add(bookmark);
                    }
                }
            }
            catch (Exception ex)
            {
                throw new Exception($"读取书签文件失败: {ex.Message}");
            }

            return bookmarks;
        }

        private static BookmarkItem ParseBookmarkLine(string line)
        {
            // 简单的解析逻辑，根据实际FreePic2Pdf_bkmk文件格式调整
            var parts = line.Split(new[] { ' ', '\t', ':' }, StringSplitOptions.RemoveEmptyEntries);

            if (parts.Length < 2)
                return null;

            var bookmark = new BookmarkItem();

            // 假设最后一部分是页码
            var pagePart = parts[parts.Length - 1];
            if (IsPageNumber(pagePart))
            {
                bookmark.Page = pagePart;
                bookmark.Title = string.Join(" ", parts, 0, parts.Length - 1);
            }
            else
            {
                // 如果没有明确的页码，跳过这一行
                return null;
            }

            return bookmark;
        }

        private static bool IsPageNumber(string text)
        {
            // 检查是否为数字（可以是罗马数字或阿拉伯数字）
            return System.Text.RegularExpressions.Regex.IsMatch(text, @"^\d+$") ||
                   System.Text.RegularExpressions.Regex.IsMatch(text, @"^[IVXLCDMivxlcdm]+$");
        }
    }
}