import re
import csv

# 定义输入与输出路径
input_path = r"E:\HDC\test0829a_merged.txt"
output_path = r"E:\HDC\test0829a_blocks_analysis3.csv"


def get_quote_pairs(line):
    """获取所有引号对的位置，严格区分中英文引号"""
    pairs = []
    stack = []  # 存放 (quote_char, index)
    for i, ch in enumerate(line):
        if ch == '“':
            stack.append(('“', i))
        elif ch == '”':
            # 找到最近的中文开引号
            for j in range(len(stack) - 1, -1, -1):
                if stack[j][0] == '“':
                    start = stack.pop(j)[1]
                    pairs.append((start, i))
                    break
        elif ch == '"':
            # ASCII 双引号，使用栈的 LIFO 行为
            if stack and stack[-1][0] == '"':
                start = stack.pop()[1]
                pairs.append((start, i))
            else:
                stack.append(('"', i))
    return pairs


def is_inside_quotes(start, end, quote_pairs):
    """检查位置是否在引号对内"""
    for q_start, q_end in quote_pairs:
        if start >= q_start and end <= q_end:
            return True
    return False


def extract_blocks_from_line(line):
    """从单行词条中提取所有节块"""
    # 初始化各类节块列表
    blocks = {
        '词头节块': [],
        '例证节块': [],
        '异写标示节块': [],
        '主条标示节块': [],
        '缩写标示节块': [],
        '异称标示节块': [],
        '简称标示节块': [],
        '同源标示节块': [],
        '详条标示节块': [],
        '他条标示节块': [],
        '异文标示节块': [],
        '义序标示节块': [],
        '读音标示节块': [],
        '畸零节块': [],
        '释文节块': []
    }

    # 存储所有已识别节块的精确位置
    all_spans = []

    # 获取所有引号对位置
    quote_pairs = get_quote_pairs(line)

    # 1. 提取词头节块（【...】）并记录位置
    citou_matches = list(re.finditer(r'【[^】]+】', line))
    for match in citou_matches:
        blocks['词头节块'].append(match.group())
        all_spans.append(match.span())

    # 2. 提取义序标示节块（三级，不在引号内）并记录位置
    yixu_blocks = []  # 存储(位置, 内容)的元组
    # 一级义序
    yixu1_pattern = r'[㊀㊁㊂㊃㊄㊅㊆㊇㊈㊉]'
    yixu1_matches = list(re.finditer(yixu1_pattern, line))
    for match in yixu1_matches:
        if not is_inside_quotes(match.start(), match.end(), quote_pairs):
            yixu_blocks.append((match.start(), match.group()))
            all_spans.append(match.span())

    # 二级义序
    yixu2_pattern = r'[❶❷❸❹❺❻❼❽❾❿⓫⓬⓭⓮⓯⓰⓱⓲⓳⓴①②③④⑤⑥⑦⑧⑨⑩⑪⑫⑬⑭⑮⑯⑰⑱⑲⑳]'
    yixu2_matches = list(re.finditer(yixu2_pattern, line))
    for match in yixu2_matches:
        if not is_inside_quotes(match.start(), match.end(), quote_pairs):
            yixu_blocks.append((match.start(), match.group()))
            all_spans.append(match.span())

    # 三级义序
    yixu3_pattern = r'[(（][1-9][0-9]*[)）]'
    yixu3_matches = list(re.finditer(yixu3_pattern, line))
    for match in yixu3_matches:
        if not is_inside_quotes(match.start(), match.end(), quote_pairs):
            yixu_blocks.append((match.start(), match.group()))
            all_spans.append(match.span())

    # 按位置排序义序标示节块
    yixu_blocks.sort(key=lambda x: x[0])
    blocks['义序标示节块'] = [block[1] for block in yixu_blocks]

    # 3. 提取读音标示节块（包含拼音字母，不在引号内）并记录位置
    # 精确匹配拼音中实际使用的带声调字母
    duinyin_pattern = r'[(（][^)）]*[a-zA-Z\u00E0\u00E1\u01CE\u0101\u00E8\u00E9\u011B\u0113\u00EC\u00ED\u01D0\u012B\u00F2\u00F3\u01D2\u014D\u00F9\u00FA\u01D4\u016B\u01D6\u01D8\u01DA\u01DC\u00FC]+[^)）]*[)）]'
    duinyin_matches = list(re.finditer(duinyin_pattern, line))
    for match in duinyin_matches:
        if not is_inside_quotes(match.start(), match.end(), quote_pairs):
            blocks['读音标示节块'].append(match.group())
            all_spans.append(match.span())

    # 4. 合并重叠的span
    all_spans.sort()
    merged_spans = []
    for start, end in all_spans:
        if merged_spans and start <= merged_spans[-1][1]:
            merged_spans[-1] = (merged_spans[-1][0], max(merged_spans[-1][1], end))
        else:
            merged_spans.append((start, end))

    # 5. 使用独特分隔符替换已识别的节块
    sep = '<<<BLOCK>>>'
    remaining_line = line
    for start, end in reversed(merged_spans):
        remaining_line = remaining_line[:start] + sep + remaining_line[end:]

    # 按独特分隔符分割成独立片段
    segments = re.split(r'(?:' + re.escape(sep) + r')+', remaining_line)
    segments = [s.strip() for s in segments if s.strip()]

    # 6. 处理每个独立片段
    for segment in segments:
        # 6.1 提取例证节块
        example_pattern = r'[^。]*?：["“][^"”]+["”]'
        example_matches = list(re.finditer(example_pattern, segment))
        for match in example_matches:
            block = match.group().strip()
            if block:
                blocks['例证节块'].append(block)

        # 从片段中移除例证节块
        temp_segment = segment
        for match in reversed(example_matches):
            start, end = match.span()
            temp_segment = temp_segment[:start] + ' ' * (end - start) + temp_segment[end:]

        # 6.2 按句号分割剩余部分
        sub_segments = re.split(r'。', temp_segment)

        # 6.3 处理最后一个非空段（畸零节块）
        if sub_segments and sub_segments[-1].strip():
            blocks['畸零节块'].append(sub_segments[-1].strip())
            sub_segments = sub_segments[:-1]

        # 6.4 处理每个普通节块
        for sub_segment in sub_segments:
            sub_segment = sub_segment.strip()
            if not sub_segment:
                continue

            # 检查各类标示节块（必须以指定关键词开头）
            if re.fullmatch(r'亦作["“][^"”]+["”](?:、["“][^"”]+["”])*', sub_segment):
                blocks['异写标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'见["“][^"”]+["”]', sub_segment):
                blocks['主条标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'亦省作["“][^"”]+["”](?:、["“][^"”]+["”])*', sub_segment):
                blocks['缩写标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'(?:亦称|也称)["“][^"”]+["”](?:、["“][^"”]+["”])*', sub_segment):
                blocks['异称标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'亦省称["“][^"”]+["”](?:、["“][^"”]+["”])*', sub_segment):
                blocks['简称标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'同["“][^"”]+["”]', sub_segment):
                blocks['同源标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'详["“][^"”]+["”](?:、["“][^"”]+["”])*', sub_segment):
                blocks['详条标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'参见["“][^"”]+["”](?:、["“][^"”]+["”])*', sub_segment):
                blocks['他条标示节块'].append(sub_segment + '。')
            elif re.fullmatch(r'.*一本作["“][^"”]+["”]', sub_segment):
                blocks['异文标示节块'].append(sub_segment + '。')
            else:
                # 普通解释节块（释文节块）
                blocks['释文节块'].append(sub_segment + '。')

    # 7. 合并同类节块
    for block_type in blocks:
        blocks[block_type] = ' || '.join(blocks[block_type]) if blocks[block_type] else ''

    return blocks


# 处理文件并生成CSV
with open(input_path, "r", encoding="utf-8") as infile, \
        open(output_path, "w", encoding="utf-8", newline="", errors='replace') as outfile:
    # 创建CSV写入器
    writer = csv.writer(outfile)
    # 写入CSV头部
    headers = ['原词条', '词头节块', '例证节块', '异写标示节块', '主条标示节块',
               '缩写标示节块', '异称标示节块', '简称标示节块', '同源标示节块',
               '详条标示节块', '他条标示节块', '异文标示节块', '义序标示节块',
               '读音标示节块', '畸零节块', '释文节块']
    writer.writerow(headers)

    # 处理每一行
    for line_num, line in enumerate(infile, 1):
        line = line.strip()
        if not line:
            continue
        try:
            # 提取所有节块
            blocks = extract_blocks_from_line(line)
            # 准备CSV行数据
            row_data = [line] + [blocks[header] for header in headers[1:]]
            # 写入CSV行
            writer.writerow(row_data)
        except Exception as e:
            print(f"处理第 {line_num} 行时出错: {e}")
            # 写入错误行
            error_row = [line] + [''] * (len(headers) - 1)
            writer.writerow(error_row)

print("✅ 改进版节块提取完成，结果已保存至：", output_path)
print("📋 提示：用WPS表格打开时请选择'UTF-8'编码")