From c6d55654b87e4f2880d671821336d6e4637d4fb8 Mon Sep 17 00:00:00 2001
From: Wuchen-4810 <Wuchen-4810@users.noreply.github.com>
Date: Tue, 12 May 2026 00:01:49 +0800
Subject: [PATCH] =?UTF-8?q?feat:=20add=20merged=20forward=20message=20(?=
 =?UTF-8?q?=E5=90=88=E5=B9=B6=E8=BD=AC=E5=8F=91)=20parsing=20support?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- Add _format_merged_forward_message to parse app_type 17/19 XML
- Extract individual messages from <recorditem> <dataitem> elements
- Show sender, timestamp, and content for each forwarded message
- Handle nested forwarded messages and file references
- Fix XML parsing failure when <?xml?> declaration is embedded inside <msg>
- Increase _XML_PARSE_MAX_LEN from 20000 to 200000 for long chat records

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
---
 wechat_cli/core/messages.py | 129 +++++++++++++++++++++++++++++++++++-
 1 file changed, 128 insertions(+), 1 deletion(-)

diff --git a/wechat_cli/core/messages.py b/wechat_cli/core/messages.py
index d62ef33..6dadbfc 100644
--- a/wechat_cli/core/messages.py
+++ b/wechat_cli/core/messages.py
@@ -14,7 +14,7 @@
 
 _zstd_dctx = zstd.ZstdDecompressor()
 _XML_UNSAFE_RE = re.compile(r'<!DOCTYPE|<!ENTITY', re.IGNORECASE)
-_XML_PARSE_MAX_LEN = 20000
+_XML_PARSE_MAX_LEN = 200000
 _QUERY_LIMIT_MAX = 500
 _HISTORY_QUERY_BATCH_SIZE = 500
 
@@ -149,9 +149,130 @@ def _parse_int(value, fallback=0):
         return fallback
 
 
+def _format_merged_forward_message(content, appmsg):
+    """解析合并转发消息 (app_type=17/19)，提取其中嵌套的逐条消息。
+
+    返回格式:
+        [合并转发] 标题
+          ├ 发送者: 消息内容
+          ├ 发送者: [文件] 文件名
+          └ ...
+    如果无法解析则返回 None。
+    """
+    records = appmsg.findall('recorditem')
+    if not records:
+        return None
+
+    title = _collapse_text(appmsg.findtext('title') or '') or '聊天记录'
+    # 从 appmsg/des 获取摘要文本（也有可能是完整描述）
+    summary_text = _collapse_text(appmsg.findtext('des') or '')
+
+    parts = [f"[合并转发] {title}"]
+
+    for record_elem in records:
+        cdata = record_elem.text or ''
+        if not cdata.strip():
+            continue
+        # 去掉 CDATA 包装标记，提取纯 XML
+        inner_xml = cdata.strip()
+        if inner_xml.startswith('<![CDATA['):
+            inner_xml = inner_xml[9:]
+        if inner_xml.endswith(']]>'):
+            inner_xml = inner_xml[:-3] + inner_xml[-3:].replace(']]>', '')
+        # 清理
+        inner_xml = inner_xml.replace('<![CDATA[', '').replace(']]>', '').strip()
+        if not inner_xml:
+            continue
+
+        record_root = _parse_xml_root(inner_xml)
+        if record_root is None:
+            continue
+
+        # 尝试解析 <datalist> 中的 <dataitem> 元素（最精确）
+        datalist = record_root.find('.//datalist')
+        parsed_items = False
+        if datalist is not None:
+            dataitems = datalist.findall('dataitem')
+            if dataitems:
+                parsed_items = True
+                for item in dataitems:
+                    sender = _collapse_text(item.findtext('sourcename') or '')
+                    msg_time = _collapse_text(item.findtext('sourcetime') or '')
+                    datatype = _parse_int(item.findtext('datatype') or '0', 0)
+                    datadesc = _collapse_text(item.findtext('datadesc') or '')
+                    datatitle = _collapse_text(item.findtext('datatitle') or '')
+                    datafmt = _collapse_text(item.findtext('datafmt') or '')
+
+                    # 根据数据类型格式化
+                    type_labels = {1: '', 2: '[图片]', 8: '[文件]', 17: '[合并转发]'}
+                    type_label = type_labels.get(datatype, f'[type={datatype}]')
+
+                    if datatype == 0 and not datadesc and not datatitle:
+                        continue  # 空分隔符/占位项，跳过
+                    if datatype == 8:  # 文件
+                        fname = datatitle or datadesc or '未知文件'
+                        ext = f' .{datafmt}' if datafmt else ''
+                        body = f'[文件] {fname}{ext}'
+                    elif datatype == 2:  # 图片
+                        body = '[图片]'
+                    elif datatype == 17:  # 嵌套合并转发
+                        body = f'[合并转发] {datatitle}'
+                    else:
+                        body = datadesc or datatitle or type_label
+
+                    # 截断过长消息
+                    if len(body) > 200:
+                        body = body[:200] + '...'
+
+                    prefix = f'  ├ {sender}' if sender else '  ├'
+                    if msg_time:
+                        prefix += f' [{msg_time}]'
+                    parts.append(f'{prefix}: {body}')
+                continue  # 已处理此 recorditem
+
+        # 回退：解析 <desc> 字段（摘要模式）
+        desc_text = _collapse_text(record_root.findtext('desc') or '')
+        if desc_text:
+            # <desc> 的格式: "发送者: 内容\n发送者: 内容\n..."
+            # HTML 实体解码
+            import html
+            desc_text = html.unescape(desc_text)
+            for line in desc_text.split('\n'):
+                line = line.strip()
+                if not line:
+                    continue
+                if ':' in line:
+                    colon_pos = line.index(':')
+                    sender_part = line[:colon_pos].strip()
+                    msg_part = line[colon_pos + 1:].strip()
+                    if len(msg_part) > 200:
+                        msg_part = msg_part[:200] + '...'
+                    parts.append(f'  ├ {sender_part}: {msg_part}')
+                else:
+                    # 没有冒号 = 可能是纯内容
+                    if len(line) > 200:
+                        line = line[:200] + '...'
+                    parts.append(f'  ├ {line}')
+
+    # 如果只解析出标题（空转发），回退 summary_text
+    if len(parts) == 1 and summary_text:
+        for line in summary_text.split('\n'):
+            line = line.strip()
+            if not line:
+                continue
+            if len(line) > 200:
+                line = line[:200] + '...'
+            parts.append(f'  ├ {line}')
+
+    return '\n'.join(parts)
+
+
 def _format_app_message_text(content, local_type, is_group, chat_username, chat_display_name, names, _display_name_fn, resolve_media=False, db_dir=None, create_time_ts=0):
     if not content or '<appmsg' not in content:
         return None
+    # 部分消息在 <msg> 内部嵌入了 <?xml version="1.0"?> 声明，导致 XML 解析失败
+    # 将其移除后再解析
+    content = re.sub(r'<\?xml\b[^?]*\?>', '', content)
     _, sub_type = _split_msg_type(local_type)
     root = _parse_xml_root(content)
     if root is None:
@@ -198,6 +319,12 @@ def _format_app_message_text(content, local_type, is_group, chat_username, chat_
         return f"[链接] {title}" if title else "[链接]"
     if app_type in (33, 36, 44):
         return f"[小程序] {title}" if title else "[小程序]"
+    if app_type in (17, 19):
+        merged = _format_merged_forward_message(content, appmsg)
+        if merged is not None:
+            return merged
+        # 回退：至少显示标题
+        return f"[合并转发] {title}" if title else "[合并转发]"
     if title:
         return f"[链接/文件] {title}"
     return "[链接/文件]"