Learn.VideoAnalysis/VideoAnalysisCore/Common/JsonExtractor.cs

using System;
using System.Collections.Generic;
using System.Text;
using System.Text.Json;
using System.Text.RegularExpressions;
using static System.Runtime.InteropServices.JavaScript.JSType;

namespace VideoAnalysisCore.Common
{
    public static class JsonExtractor
    {
        /// <summary>
        /// 修复字符串中不规范的反斜杠转义，使其符合 JSON 规范。
        /// 特别适用于包含 LaTeX 公式（如 \overrightarrow, \unit）的非标准 JSON 数据。
        /// </summary>
        public static string ToSafeJsonString(this string json)
        {
            if (string.IsNullOrEmpty(json)) return json;

            // 预分配稍大一点的空间，避免频繁扩容
            StringBuilder sb = new StringBuilder(json.Length + (json.Length / 10));
            int i = 0;
            int len = json.Length;

            while (i < len)
            {
                char c = json[i];

                if (c == '\\')
                {
                    int start = i;
                    while (i < len && json[i] == '\\')
                    {
                        i++;
                    }
                    int count = i - start;

                    // 只有奇数个反斜杠才需要检查“尾巴”是否合法
                    if (count % 2 != 0)
                    {
                        // 检查这最后一个反斜杠后面跟的是不是合法的 JSON 转义字符
                        if (i >= len || !IsValidJsonEscape(json, i))
                        {
                            // 非法转义，补齐它
                            count++;
                        }
                    }

                    // 性能优化：直接添加指定数量的字符
                    sb.Append('\\', count);
                }
                else
                {
                    sb.Append(c);
                    i++;
                }
            }

            return sb.ToString();
        }

        private static bool IsValidJsonEscape(string text, int nextCharIndex)
        {
            char nextChar = text[nextCharIndex];

            // 标准 JSON 简单转义
            if (nextChar == '"' || nextChar == '\\' || nextChar == '/' ||
                nextChar == 'b' || nextChar == 'f' || nextChar == 'n' ||
                nextChar == 'r' || nextChar == 't')
                return true;

            // Unicode 转义检查: \uXXXX
            if (nextChar == 'u')
            {
                if (nextCharIndex + 4 < text.Length)
                {
                    for (int k = 1; k <= 4; k++)
                    {
                        char hex = text[nextCharIndex + k];
                        bool isHex = (hex >= '0' && hex <= '9') ||
                                     (hex >= 'a' && hex <= 'f') ||
                                     (hex >= 'A' && hex <= 'F');
                        if (!isHex) return false;
                    }
                    return true;
                }
                return false;
            }

            return false;
        }
        /// <summary>
        /// 提取json字符串
        /// </summary>
        /// <param name="input"></param>
        /// <returns></returns>
        public static List<string> ExtractJsonStrings(this string input)
        {
            input = input.ToSafeJsonString();
            var results = new List<string>();
            if (string.IsNullOrWhiteSpace(input)) return results;

            int braceCount = 0;
            int bracketCount = 0;
            int startIndex = -1;
            bool inString = false;
            bool isEscaped = false;

            for (int i = 0; i < input.Length; i++)
            {
                char c = input[i];

                // 1. 处理转义字符 (例如 \")
                if (isEscaped)
                {
                    isEscaped = false;
                    continue;
                }

                if (c == '\\')
                {
                    isEscaped = true;
                    continue;
                }

                // 2. 处理字符串边界
                if (c == '"')
                {
                    inString = !inString;
                    continue;
                }

                // 3. 如果在字符串内，忽略括号逻辑
                if (inString) continue;

                // 4. 处理 JSON 对象和数组的开始
                if (c == '{' || c == '[')
                {
                    if (braceCount == 0 && bracketCount == 0)
                    {
                        startIndex = i;
                    }
                    if (c == '{') braceCount++;
                    else bracketCount++;
                }
                // 5. 处理 JSON 对象和数组的结束
                else if (c == '}' || c == ']')
                {
                    if (c == '}') braceCount--;
                    else bracketCount--;

                    if (braceCount == 0 && bracketCount == 0 && startIndex != -1)
                    {
                        string potentialJson = input.Substring(startIndex, i - startIndex + 1);
                        if (IsValidJson(potentialJson))
                        {
                            results.Add(potentialJson);
                        }
                        startIndex = -1;
                    }
                }
            }
            return results;
        }

        public static bool IsValidJson(string candidate)
        {
            if (string.IsNullOrEmpty(candidate))
                return false;
            try
            {
                JsonDocument.Parse(candidate);
                return true;
            }
            catch( Exception e)
            {
                return false;
            }
        }
    }
}