From 814abba2f99d3764f3de5ab61855330610904cb8 Mon Sep 17 00:00:00 2001
From: 2977094657 <2977094657@qq.com>
Date: Mon, 9 Feb 2026 18:31:00 +0800
Subject: [PATCH] =?UTF-8?q?feat(chat):=20=E7=BE=A4=E8=81=8A=E9=A2=84?=
 =?UTF-8?q?=E8=A7=88=E8=A1=A5=E9=BD=90=E7=BE=A4=E5=90=8D=E7=89=87=E5=B9=B6?=
 =?UTF-8?q?=E5=AE=8C=E5=96=84=E7=B3=BB=E7=BB=9F=E6=B6=88=E6=81=AF=E8=A7=A3?=
 =?UTF-8?q?=E6=9E=90?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- 新增系统撤回消息解析：优先提取 replacemsg，并统一清洗文本

- 群聊会话预览文本规范化（[表情] -> [动画表情]），并支持发送者前缀展示名替换

- 群名片解析来源扩展：contact.db ext_buffer + WCDB realtime（可选新 DLL 接口）

- 图片接口增强：支持 server_id + username 反查消息提取 md5，提升引用图片命中
---
 .../chat_export_service.py                    |   9 +-
 src/wechat_decrypt_tool/chat_helpers.py       | 348 ++++++++++++++-
 src/wechat_decrypt_tool/media_helpers.py      |  11 +-
 src/wechat_decrypt_tool/routers/chat.py       | 398 +++++++++++++++---
 src/wechat_decrypt_tool/routers/chat_media.py |  83 ++++
 src/wechat_decrypt_tool/wcdb_realtime.py      |  46 ++
 6 files changed, 815 insertions(+), 80 deletions(-)

diff --git a/src/wechat_decrypt_tool/chat_export_service.py b/src/wechat_decrypt_tool/chat_export_service.py
index d4435fa..eead512 100644
--- a/src/wechat_decrypt_tool/chat_export_service.py
+++ b/src/wechat_decrypt_tool/chat_export_service.py
@@ -28,6 +28,7 @@ from .chat_helpers import (
     _load_contact_rows,
     _lookup_resource_md5,
     _parse_app_message,
+    _parse_system_message_content,
     _parse_pat_message,
     _pick_display_name,
     _quote_ident,
@@ -954,13 +955,7 @@ def _parse_message_for_export(
 
     if local_type == 10000:
         render_type = "system"
-        if "revokemsg" in raw_text:
-            content_text = "撤回了一条消息"
-        else:
-            import re as _re
-
-            content_text = _re.sub(r"</?[_a-zA-Z0-9]+[^>]*>", "", raw_text)
-            content_text = _re.sub(r"\\s+", " ", content_text).strip() or "[系统消息]"
+        content_text = _parse_system_message_content(raw_text)
     elif local_type == 49:
         parsed = _parse_app_message(raw_text)
         render_type = str(parsed.get("renderType") or "text")
diff --git a/src/wechat_decrypt_tool/chat_helpers.py b/src/wechat_decrypt_tool/chat_helpers.py
index 57f44c2..ca56065 100644
--- a/src/wechat_decrypt_tool/chat_helpers.py
+++ b/src/wechat_decrypt_tool/chat_helpers.py
@@ -645,6 +645,43 @@ def _extract_xml_tag_or_attr(xml_text: str, name: str) -> str:
     return _extract_xml_attr(xml_text, name)
 
 
+def _parse_system_message_content(raw_text: str) -> str:
+    text = str(raw_text or "").strip()
+    if not text:
+        return "[系统消息]"
+
+    def _clean_system_text(value: str) -> str:
+        candidate = str(value or "").strip()
+        if not candidate:
+            return ""
+
+        nested_content = _extract_xml_tag_text(candidate, "content")
+        if nested_content:
+            candidate = nested_content
+
+        candidate = re.sub(r"<!\[CDATA\[", "", candidate, flags=re.IGNORECASE)
+        candidate = re.sub(r"\]\]>", "", candidate)
+        candidate = re.sub(r"</?[_a-zA-Z0-9]+[^>]*>", "", candidate)
+        candidate = re.sub(r"\s+", " ", candidate).strip()
+        return candidate
+
+    if "revokemsg" in text.lower():
+        replace_msg = _extract_xml_tag_text(text, "replacemsg")
+        cleaned_replace_msg = _clean_system_text(replace_msg)
+        if cleaned_replace_msg:
+            return cleaned_replace_msg
+
+        revoke_msg = _extract_xml_tag_text(text, "revokemsg")
+        cleaned_revoke_msg = _clean_system_text(revoke_msg)
+        if cleaned_revoke_msg:
+            return cleaned_revoke_msg
+
+        return "撤回了一条消息"
+
+    content_text = _clean_system_text(text)
+    return content_text or "[系统消息]"
+
+
 def _extract_refermsg_block(xml_text: str) -> str:
     if not xml_text:
         return ""
@@ -1053,11 +1090,7 @@ def _build_latest_message_preview(
 
     content_text = ""
     if local_type == 10000:
-        if "revokemsg" in raw_text:
-            content_text = "撤回了一条消息"
-        else:
-            content_text = re.sub(r"</?[_a-zA-Z0-9]+[^>]*>", "", raw_text)
-            content_text = re.sub(r"\s+", " ", content_text).strip() or "[系统消息]"
+        content_text = _parse_system_message_content(raw_text)
     elif local_type == 244813135921:
         parsed = _parse_app_message(raw_text)
         qt = str(parsed.get("quoteTitle") or "").strip()
@@ -1093,7 +1126,7 @@ def _build_latest_message_preview(
     elif local_type == 43 or local_type == 62:
         content_text = "[视频]"
     elif local_type == 47:
-        content_text = "[表情]"
+        content_text = "[动画表情]"
     else:
         if raw_text and (not raw_text.startswith("<")) and (not raw_text.startswith('"<')):
             content_text = raw_text
@@ -1107,6 +1140,101 @@ def _build_latest_message_preview(
     return content_text
 
 
+def _extract_group_preview_sender_username(preview_text: str) -> str:
+    text = str(preview_text or "").strip()
+    if not text:
+        return ""
+
+    match = re.match(r"^([^:\s]{1,128}):\s*.+$", text)
+    if not match:
+        return ""
+
+    sender = str(match.group(1) or "").strip()
+    if not sender:
+        return ""
+
+    if sender.startswith("wxid_") or sender.endswith("@chatroom") or ("@" in sender):
+        return sender
+    if re.fullmatch(r"[A-Za-z][A-Za-z0-9_-]{1,127}", sender):
+        return sender
+    return ""
+
+
+def _normalize_session_preview_text(
+    preview_text: str,
+    *,
+    is_group: bool,
+    sender_display_names: Optional[dict[str, str]] = None,
+) -> str:
+    text = re.sub(r"\s+", " ", str(preview_text or "").strip()).strip()
+    if not text:
+        return ""
+
+    text = text.replace("[表情]", "[动画表情]")
+    if (not is_group) or text.startswith("[草稿]"):
+        return text
+
+    match = re.match(r"^([^:\s]{1,128}):\s*(.+)$", text)
+    if not match:
+        return text
+
+    sender_username = str(match.group(1) or "").strip()
+    body = str(match.group(2) or "").strip()
+    if (not sender_username) or (not body):
+        return text
+
+    display_name = str((sender_display_names or {}).get(sender_username) or "").strip()
+    if display_name and display_name != sender_username:
+        return f"{display_name}: {body}"
+    return text
+
+
+def _replace_preview_sender_prefix(preview_text: str, sender_display_name: str) -> str:
+    text = re.sub(r"\s+", " ", str(preview_text or "").strip()).strip()
+    if not text:
+        return ""
+
+    display_name = str(sender_display_name or "").strip()
+    if (not display_name) or text.startswith("[草稿]"):
+        return text
+
+    match = re.match(r"^([^:\n]{1,128}):\s*(.+)$", text)
+    if not match:
+        return text
+
+    body = re.sub(r"\s+", " ", str(match.group(2) or "").strip()).strip()
+    if not body:
+        return text
+    return f"{display_name}: {body}"
+
+
+def _build_group_sender_display_name_map(
+    contact_db_path: Path,
+    previews: dict[str, str],
+) -> dict[str, str]:
+    group_sender_usernames: set[str] = set()
+    for conv_username, preview_text in previews.items():
+        if not str(conv_username or "").endswith("@chatroom"):
+            continue
+        sender_username = _extract_group_preview_sender_username(preview_text)
+        if sender_username:
+            group_sender_usernames.add(sender_username)
+
+    if not group_sender_usernames:
+        return {}
+
+    display_names: dict[str, str] = {}
+    sender_contact_rows = _load_contact_rows(contact_db_path, list(group_sender_usernames))
+    for sender_username in group_sender_usernames:
+        row = sender_contact_rows.get(sender_username)
+        if row is None:
+            continue
+        display_name = _pick_display_name(row, sender_username)
+        if display_name and display_name != sender_username:
+            display_names[sender_username] = display_name
+    return display_names
+
+
 def _load_latest_message_previews(account_dir: Path, usernames: list[str]) -> dict[str, str]:
     if not usernames:
         return {}
@@ -1338,6 +1466,208 @@ def _load_contact_rows(contact_db_path: Path, usernames: list[str]) -> dict[str,
         conn.close()
 
 
+def _load_group_nickname_map_from_contact_db(
+    contact_db_path: Path,
+    chatroom_id: str,
+    sender_usernames: list[str],
+) -> dict[str, str]:
+    """Best-effort mapping for group member nickname (aka group card) from contact.db.
+
+    WeChat stores per-chatroom member nicknames in `contact.db.chat_room.ext_buffer` as a protobuf-like blob.
+    This helper parses that blob and returns { sender_username -> group_nickname } for the requested senders.
+
+    Notes:
+    - Best-effort: never raises; returns {} on any failure.
+    - Only resolves usernames included in `sender_usernames` to keep parsing cheap.
+    """
+
+    chatroom = str(chatroom_id or "").strip()
+    if not chatroom.endswith("@chatroom"):
+        return {}
+
+    targets = list(dict.fromkeys([str(x or "").strip() for x in sender_usernames if str(x or "").strip()]))
+    if not targets:
+        return {}
+    target_set = set(targets)
+
+    def decode_varint(raw: bytes, offset: int) -> tuple[Optional[int], int]:
+        value = 0
+        shift = 0
+        pos = int(offset)
+        n = len(raw)
+        while pos < n:
+            byte = raw[pos]
+            pos += 1
+            value |= (byte & 0x7F) << shift
+            if (byte & 0x80) == 0:
+                return value, pos
+            shift += 7
+            if shift > 63:
+                return None, n
+        return None, n
+
+    def iter_fields(raw: bytes):
+        idx = 0
+        n = len(raw)
+        while idx < n:
+            tag, idx_next = decode_varint(raw, idx)
+            if tag is None or idx_next <= idx:
+                break
+            idx = idx_next
+            field_no = int(tag) >> 3
+            wire_type = int(tag) & 0x7
+
+            if wire_type == 0:
+                _, idx_next = decode_varint(raw, idx)
+                if idx_next <= idx:
+                    break
+                idx = idx_next
+                continue
+
+            if wire_type == 2:
+                size, idx_next = decode_varint(raw, idx)
+                if size is None or idx_next <= idx:
+                    break
+                idx = idx_next
+                end = idx + int(size)
+                if end > n:
+                    break
+                chunk = raw[idx:end]
+                idx = end
+                yield field_no, wire_type, chunk
+                continue
+
+            if wire_type == 1:
+                idx += 8
+                continue
+            if wire_type == 5:
+                idx += 4
+                continue
+            break
+
+    def is_strong_username_hint(s: str) -> bool:
+        v = str(s or "").strip()
+        return v.startswith("wxid_") or v.endswith("@chatroom") or v.startswith("gh_") or ("@" in v)
+
+    def looks_like_username(s: str) -> bool:
+        v = str(s or "").strip()
+        if not v:
+            return False
+        if is_strong_username_hint(v):
+            return True
+        # Common alias-style WeChat IDs are ASCII-ish and do not contain whitespace.
+        if len(v) < 6 or len(v) > 32:
+            return False
+        if re.search(r"\s", v):
+            return False
+        if not re.match(r"^[A-Za-z][A-Za-z0-9_-]+$", v):
+            return False
+        if v.isdigit():
+            return False
+        return True
+
+    def pick_display(strings: list[tuple[int, str]], target: str) -> str:
+        best_score = -1
+        best = ""
+        for i, (fno, value) in enumerate(strings):
+            v = str(value or "").strip()
+            if (not v) or v == target:
+                continue
+            if is_strong_username_hint(v):
+                continue
+            if "\n" in v or "\r" in v:
+                continue
+            if len(v) > 64:
+                continue
+
+            score = 0
+            if int(fno) == 2:
+                score += 100
+            if not looks_like_username(v):
+                score += 20
+            score += max(0, 32 - len(v))
+            # Stable tie-breaker: prefer earlier appearance.
+            score = score * 1000 - i
+            if score > best_score:
+                best_score = score
+                best = v
+        return best
+
+    try:
+        conn = sqlite3.connect(str(contact_db_path))
+    except Exception:
+        return {}
+
+    try:
+        row = conn.execute(
+            "SELECT ext_buffer FROM chat_room WHERE username = ? LIMIT 1",
+            (chatroom,),
+        ).fetchone()
+        if row is None:
+            return {}
+
+        ext = row[0]
+        if ext is None:
+            return {}
+        if isinstance(ext, memoryview):
+            ext_buf = ext.tobytes()
+        elif isinstance(ext, (bytes, bytearray)):
+            ext_buf = bytes(ext)
+        else:
+            return {}
+        if not ext_buf:
+            return {}
+
+        out: dict[str, str] = {}
+        for _, wire_type, chunk in iter_fields(ext_buf):
+            if wire_type != 2 or (not chunk):
+                continue
+
+            # Parse submessage and collect UTF-8 strings.
+            strings: list[tuple[int, str]] = []
+            try:
+                for sfno, swire, sval in iter_fields(chunk):
+                    if swire != 2:
+                        continue
+                    if not sval:
+                        continue
+                    if len(sval) > 256:
+                        continue
+                    try:
+                        txt = bytes(sval).decode("utf-8", errors="strict")
+                    except Exception:
+                        continue
+                    txt = txt.strip()
+                    if not txt:
+                        continue
+                    strings.append((int(sfno), txt))
+            except Exception:
+                continue
+
+            if not strings:
+                continue
+
+            present = [v for _, v in strings if v in target_set and v not in out]
+            if not present:
+                continue
+
+            for target in present:
+                disp = pick_display(strings, target)
+                if disp:
+                    out[target] = disp
+            if len(out) >= len(target_set):
+                break
+
+        return out
+    except Exception:
+        return {}
+    finally:
+        try:
+            conn.close()
+        except Exception:
+            pass
+
+
 def _load_usernames_by_display_names(contact_db_path: Path, names: list[str]) -> dict[str, str]:
     """Best-effort mapping from display name -> username using contact.db.
 
@@ -1515,11 +1845,7 @@ def _row_to_search_hit(
 
     if local_type == 10000:
         render_type = "system"
-        if "revokemsg" in raw_text:
-            content_text = "撤回了一条消息"
-        else:
-            content_text = re.sub(r"</?[_a-zA-Z0-9]+[^>]*>", "", raw_text)
-            content_text = re.sub(r"\s+", " ", content_text).strip() or "[系统消息]"
+        content_text = _parse_system_message_content(raw_text)
     elif local_type == 49:
         parsed = _parse_app_message(raw_text)
         render_type = str(parsed.get("renderType") or "text")
diff --git a/src/wechat_decrypt_tool/media_helpers.py b/src/wechat_decrypt_tool/media_helpers.py
index 59ed9c8..92a9b20 100644
--- a/src/wechat_decrypt_tool/media_helpers.py
+++ b/src/wechat_decrypt_tool/media_helpers.py
@@ -23,17 +23,17 @@ logger = get_logger(__name__)
 
 
 # 运行时输出目录（桌面端可通过 WECHAT_TOOL_DATA_DIR 指向可写目录）
-_OUTPUT_DATABASES_DIR = get_output_databases_dir()
 _PACKAGE_ROOT = Path(__file__).resolve().parent
 
 
 def _list_decrypted_accounts() -> list[str]:
     """列出已解密输出的账号目录名（仅保留包含 session.db + contact.db 的账号）"""
-    if not _OUTPUT_DATABASES_DIR.exists():
+    output_db_dir = get_output_databases_dir()
+    if not output_db_dir.exists():
         return []
 
     accounts: list[str] = []
-    for p in _OUTPUT_DATABASES_DIR.iterdir():
+    for p in output_db_dir.iterdir():
         if not p.is_dir():
             continue
         if (p / "session.db").exists() and (p / "contact.db").exists():
@@ -45,6 +45,7 @@ def _list_decrypted_accounts() -> list[str]:
 
 def _resolve_account_dir(account: Optional[str]) -> Path:
     """解析账号目录，并进行路径安全校验（防止路径穿越）"""
+    output_db_dir = get_output_databases_dir()
     accounts = _list_decrypted_accounts()
     if not accounts:
         raise HTTPException(
@@ -53,8 +54,8 @@ def _resolve_account_dir(account: Optional[str]) -> Path:
         )
 
     selected = account or accounts[0]
-    base = _OUTPUT_DATABASES_DIR.resolve()
-    candidate = (_OUTPUT_DATABASES_DIR / selected).resolve()
+    base = output_db_dir.resolve()
+    candidate = (output_db_dir / selected).resolve()
 
     if candidate != base and base not in candidate.parents:
         raise HTTPException(status_code=400, detail="Invalid account path.")
diff --git a/src/wechat_decrypt_tool/routers/chat.py b/src/wechat_decrypt_tool/routers/chat.py
index 27a67ab..9b3de1d 100644
--- a/src/wechat_decrypt_tool/routers/chat.py
+++ b/src/wechat_decrypt_tool/routers/chat.py
@@ -39,11 +39,17 @@ from ..chat_helpers import (
     _make_snippet,
     _match_tokens,
     _load_contact_rows,
+    _load_group_nickname_map_from_contact_db,
     _load_usernames_by_display_names,
     _load_latest_message_previews,
+    _build_group_sender_display_name_map,
+    _normalize_session_preview_text,
+    _extract_group_preview_sender_username,
+    _replace_preview_sender_prefix,
     _lookup_resource_md5,
     _normalize_xml_url,
     _parse_app_message,
+    _parse_system_message_content,
     _parse_pat_message,
     _pick_display_name,
     _query_head_image_usernames,
@@ -69,6 +75,8 @@ from ..wcdb_realtime import (
     WCDB_REALTIME,
     get_avatar_urls as _wcdb_get_avatar_urls,
     get_display_names as _wcdb_get_display_names,
+    get_group_members as _wcdb_get_group_members,
+    get_group_nicknames as _wcdb_get_group_nicknames,
     get_messages as _wcdb_get_messages,
     get_sessions as _wcdb_get_sessions,
 )
@@ -97,6 +105,142 @@ def _avatar_url_unified(
     return _build_avatar_url(str(account_dir.name or ""), u)
 
 
+def _load_group_nickname_map_from_wcdb(
+    *,
+    account_dir: Path,
+    chatroom_id: str,
+    sender_usernames: list[str],
+    rt_conn=None,
+) -> dict[str, str]:
+    chatroom = str(chatroom_id or "").strip()
+    if not chatroom.endswith("@chatroom"):
+        return {}
+
+    targets = list(dict.fromkeys([str(x or "").strip() for x in sender_usernames if str(x or "").strip()]))
+    if not targets:
+        return {}
+
+    try:
+        wcdb_conn = rt_conn or WCDB_REALTIME.ensure_connected(account_dir)
+    except Exception:
+        return {}
+
+    target_set = set(targets)
+    out: dict[str, str] = {}
+
+    try:
+        with wcdb_conn.lock:
+            nickname_map = _wcdb_get_group_nicknames(wcdb_conn.handle, chatroom)
+        for username, nickname in (nickname_map or {}).items():
+            su = str(username or "").strip()
+            nn = str(nickname or "").strip()
+            if su and nn and su in target_set:
+                out[su] = nn
+    except Exception:
+        pass
+
+    unresolved = [u for u in targets if u not in out]
+    if not unresolved:
+        return out
+
+    try:
+        with wcdb_conn.lock:
+            members = _wcdb_get_group_members(wcdb_conn.handle, chatroom)
+    except Exception:
+        return out
+
+    if not members:
+        return out
+
+    unresolved_set = set(unresolved)
+    for member in members:
+        try:
+            username = str(member.get("username") or "").strip()
+        except Exception:
+            username = ""
+        if (not username) or (username not in unresolved_set):
+            continue
+
+        nickname = ""
+        for key in ("nickname", "displayName", "remark", "originalName"):
+            try:
+                candidate = str(member.get(key) or "").strip()
+            except Exception:
+                candidate = ""
+            if candidate:
+                nickname = candidate
+                break
+        if nickname:
+            out[username] = nickname
+
+    return out
+
+
+def _load_group_nickname_map(
+    *,
+    account_dir: Path,
+    contact_db_path: Path,
+    chatroom_id: str,
+    sender_usernames: list[str],
+    rt_conn=None,
+) -> dict[str, str]:
+    """Resolve group member nickname (group card) via WCDB and contact.db ext_buffer (best-effort)."""
+
+    contact_map: dict[str, str] = {}
+    try:
+        contact_map = _load_group_nickname_map_from_contact_db(
+            contact_db_path,
+            chatroom_id,
+            sender_usernames,
+        )
+    except Exception:
+        contact_map = {}
+
+    wcdb_map: dict[str, str] = {}
+    try:
+        wcdb_map = _load_group_nickname_map_from_wcdb(
+            account_dir=account_dir,
+            chatroom_id=chatroom_id,
+            sender_usernames=sender_usernames,
+            rt_conn=rt_conn,
+        )
+    except Exception:
+        wcdb_map = {}
+
+    if not contact_map and not wcdb_map:
+        return {}
+
+    # Merge: WCDB wins (newer DLLs may provide higher-quality group nicknames).
+    merged: dict[str, str] = {}
+    merged.update(contact_map)
+    merged.update(wcdb_map)
+    return merged
+
+
+def _resolve_sender_display_name(
+    *,
+    sender_username: str,
+    sender_contact_rows: dict[str, sqlite3.Row],
+    wcdb_display_names: dict[str, str],
+    group_nicknames: Optional[dict[str, str]] = None,
+) -> str:
+    su = str(sender_username or "").strip()
+    if not su:
+        return ""
+
+    gn = str((group_nicknames or {}).get(su) or "").strip()
+    if gn:
+        return gn
+
+    row = sender_contact_rows.get(su)
+    display_name = _pick_display_name(row, su)
+    if display_name == su:
+        wd = str(wcdb_display_names.get(su) or "").strip()
+        if wd and wd != su:
+            display_name = wd
+    return display_name
+
+
 def _realtime_sync_lock(account: str, username: str) -> threading.Lock:
     key = (str(account or "").strip(), str(username or "").strip())
     with _REALTIME_SYNC_MU:
@@ -557,8 +701,11 @@ def _upsert_session_table_rows(conn: sqlite3.Connection, rows: list[dict[str, An
         "draft",
         "last_timestamp",
         "sort_timestamp",
+        "last_msg_locald_id",
         "last_msg_type",
         "last_msg_sub_type",
+        "last_msg_sender",
+        "last_sender_display_name",
     ]
     update_cols = [c for c in desired_cols if c in cols]
     if not update_cols:
@@ -583,7 +730,15 @@ def _upsert_session_table_rows(conn: sqlite3.Connection, rows: list[dict[str, An
             continue
         values: list[Any] = []
         for c in update_cols:
-            if c in {"unread_count", "is_hidden", "last_timestamp", "sort_timestamp", "last_msg_type", "last_msg_sub_type"}:
+            if c in {
+                "unread_count",
+                "is_hidden",
+                "last_timestamp",
+                "sort_timestamp",
+                "last_msg_locald_id",
+                "last_msg_type",
+                "last_msg_sub_type",
+            }:
                 values.append(_int((r or {}).get(c)))
             else:
                 values.append(_text((r or {}).get(c)))
@@ -1510,8 +1665,17 @@ def sync_chat_realtime_messages_all(
                     "sort_timestamp",
                     item.get("sortTimestamp", item.get("last_timestamp", item.get("lastTimestamp", 0))),
                 ),
+                "last_msg_locald_id": item.get(
+                    "last_msg_locald_id",
+                    item.get("lastMsgLocaldId", item.get("lastMsgLocalId", 0)),
+                ),
                 "last_msg_type": item.get("last_msg_type", item.get("lastMsgType", 0)),
                 "last_msg_sub_type": item.get("last_msg_sub_type", item.get("lastMsgSubType", 0)),
+                "last_msg_sender": item.get("last_msg_sender", item.get("lastMsgSender", "")),
+                "last_sender_display_name": item.get(
+                    "last_sender_display_name",
+                    item.get("lastSenderDisplayName", ""),
+                ),
             }
             # Prefer the row with the newer sort timestamp for the same username.
             prev = realtime_rows_by_user.get(uname)
@@ -2137,11 +2301,7 @@ def _append_full_messages_from_rows(
 
         if local_type == 10000:
             render_type = "system"
-            if "revokemsg" in raw_text:
-                content_text = "撤回了一条消息"
-            else:
-                content_text = re.sub(r"</?[_a-zA-Z0-9]+[^>]*>", "", raw_text)
-                content_text = re.sub(r"\s+", " ", content_text).strip() or "[系统消息]"
+            content_text = _parse_system_message_content(raw_text)
         elif local_type == 49:
             parsed = _parse_app_message(raw_text)
             render_type = str(parsed.get("renderType") or "text")
@@ -2598,6 +2758,13 @@ def _postprocess_full_messages(
         wcdb_display_names = {}
         wcdb_avatar_urls = {}
 
+    group_nicknames = _load_group_nickname_map(
+        account_dir=account_dir,
+        contact_db_path=contact_db_path,
+        chatroom_id=username,
+        sender_usernames=uniq_senders,
+    )
+
     for m in merged:
         # If appmsg doesn't provide sourcedisplayname, try mapping sourceusername to display name.
         if (not str(m.get("from") or "").strip()) and str(m.get("fromUsername") or "").strip():
@@ -2613,13 +2780,12 @@ def _postprocess_full_messages(
         su = str(m.get("senderUsername") or "")
         if not su:
             continue
-        row = sender_contact_rows.get(su)
-        display_name = _pick_display_name(row, su)
-        if display_name == su:
-            wd = str(wcdb_display_names.get(su) or "").strip()
-            if wd and wd != su:
-                display_name = wd
-        m["senderDisplayName"] = display_name
+        m["senderDisplayName"] = _resolve_sender_display_name(
+            sender_username=su,
+            sender_contact_rows=sender_contact_rows,
+            wcdb_display_names=wcdb_display_names,
+            group_nicknames=group_nicknames,
+        )
         avatar_url = base_url + _avatar_url_unified(
             account_dir=account_dir,
             username=su,
@@ -2836,6 +3002,17 @@ def list_chat_sessions(
                     "sort_timestamp": item.get("sort_timestamp", item.get("sortTimestamp", item.get("last_timestamp", 0))),
                     "last_msg_type": item.get("last_msg_type", item.get("lastMsgType", 0)),
                     "last_msg_sub_type": item.get("last_msg_sub_type", item.get("lastMsgSubType", 0)),
+                    # Keep these fields so group session previews can render "sender: content" without
+                    # crashing (realtime rows are dicts, not sqlite Rows).
+                    "last_msg_sender": item.get("last_msg_sender", item.get("lastMsgSender", "")),
+                    "last_sender_display_name": item.get(
+                        "last_sender_display_name",
+                        item.get("lastSenderDisplayName", ""),
+                    ),
+                    "last_msg_locald_id": item.get(
+                        "last_msg_locald_id",
+                        item.get("lastMsgLocaldId", item.get("lastMsgLocalId", 0)),
+                    ),
                 }
             )
 
@@ -2923,12 +3100,16 @@ def list_chat_sessions(
     try:
         need_display: list[str] = []
         need_avatar: list[str] = []
+        if source_norm == "realtime":
+            # In realtime mode, always ask WCDB for display names: decrypted contact.db can be stale.
+            need_display = [str(u or "").strip() for u in usernames if str(u or "").strip()]
         for u in usernames:
             if not u:
                 continue
-            row = contact_rows.get(u)
-            if _pick_display_name(row, u) == u:
-                need_display.append(u)
+            if source_norm != "realtime":
+                row = contact_rows.get(u)
+                if _pick_display_name(row, u) == u:
+                    need_display.append(u)
             if source_norm == "realtime":
                 # In realtime mode, prefer WCDB-resolved avatar URLs (contact.db can be stale).
                 if u not in local_avatar_usernames:
@@ -2983,14 +3164,40 @@ def list_chat_sessions(
                 if v:
                     last_previews[u] = v
 
+    group_sender_display_names: dict[str, str] = _build_group_sender_display_name_map(
+        contact_db_path,
+        last_previews,
+    )
+    unresolved = []
+    for conv_username, preview_text in last_previews.items():
+        if not str(conv_username or "").endswith("@chatroom"):
+            continue
+        sender_username = _extract_group_preview_sender_username(preview_text)
+        if sender_username and sender_username not in group_sender_display_names:
+            unresolved.append(sender_username)
+    unresolved = list(dict.fromkeys(unresolved))
+    if unresolved:
+        try:
+            wcdb_conn = rt_conn or WCDB_REALTIME.ensure_connected(account_dir)
+            with wcdb_conn.lock:
+                wcdb_names = _wcdb_get_display_names(wcdb_conn.handle, unresolved)
+            for sender_username in unresolved:
+                wcdb_name = str(wcdb_names.get(sender_username) or "").strip()
+                if wcdb_name and wcdb_name != sender_username:
+                    group_sender_display_names[sender_username] = wcdb_name
+        except Exception:
+            pass
+
     sessions: list[dict[str, Any]] = []
     for r in filtered:
         username = r["username"]
         c_row = contact_rows.get(username)
 
         display_name = _pick_display_name(c_row, username)
-        if display_name == username:
-            wd = str(wcdb_display_names.get(username) or "").strip()
+        wd = str(wcdb_display_names.get(username) or "").strip()
+        if source_norm == "realtime" and wd and wd != username:
+            display_name = wd
+        elif display_name == username:
             if wd and wd != username:
                 display_name = wd
 
@@ -3046,6 +3253,37 @@ def list_chat_sessions(
             if last_msg_type == 81604378673 or (last_msg_type == 49 and last_msg_sub_type == 19):
                 last_message = "[聊天记录]"
 
+        last_message = _normalize_session_preview_text(
+            last_message,
+            is_group=bool(str(username or "").endswith("@chatroom")),
+            sender_display_names=group_sender_display_names,
+        )
+        if str(username or "").endswith("@chatroom") and str(last_message or "") and not str(last_message).startswith("[草稿]"):
+            # Prefer group card nickname when available. In realtime mode, WCDB session rows can provide
+            # `last_sender_display_name`, but we may still get a summary that doesn't include "sender:".
+            # Also guard against URL schemes like "https://..." being mis-parsed as "https: //...".
+            raw_sender_display = ""
+            try:
+                raw_sender_display = r["last_sender_display_name"]
+            except Exception:
+                try:
+                    raw_sender_display = r.get("last_sender_display_name", "")
+                except Exception:
+                    raw_sender_display = ""
+            sender_display = _decode_sqlite_text(raw_sender_display).strip()
+            if sender_display:
+                text = re.sub(r"\s+", " ", str(last_message or "").strip()).strip()
+                match = re.match(r"^([^:\n]{1,128}):\s*(.+)$", text)
+                if match:
+                    prefix = str(match.group(1) or "").strip()
+                    body = re.sub(r"\s+", " ", str(match.group(2) or "").strip()).strip()
+                    if prefix.lower() in {"http", "https"} and body.startswith("//"):
+                        last_message = f"{sender_display}: {text}"
+                    else:
+                        last_message = f"{sender_display}: {body}"
+                else:
+                    last_message = f"{sender_display}: {text}"
+
         last_time = _format_session_time(r["sort_timestamp"] or r["last_timestamp"])
 
         sessions.append(
@@ -3248,13 +3486,7 @@ def _collect_chat_messages(
 
                 if local_type == 10000:
                     render_type = "system"
-                    if "revokemsg" in raw_text:
-                        content_text = "撤回了一条消息"
-                    else:
-                        import re
-
-                        content_text = re.sub(r"</?[_a-zA-Z0-9]+[^>]*>", "", raw_text)
-                        content_text = re.sub(r"\s+", " ", content_text).strip() or "[系统消息]"
+                    content_text = _parse_system_message_content(raw_text)
                 elif local_type == 49:
                     parsed = _parse_app_message(raw_text)
                     render_type = str(parsed.get("renderType") or "text")
@@ -3957,13 +4189,7 @@ def list_chat_messages(
 
                 if local_type == 10000:
                     render_type = "system"
-                    if "revokemsg" in raw_text:
-                        content_text = "撤回了一条消息"
-                    else:
-                        import re
-
-                        content_text = re.sub(r"</?[_a-zA-Z0-9]+[^>]*>", "", raw_text)
-                        content_text = re.sub(r"\s+", " ", content_text).strip() or "[系统消息]"
+                    content_text = _parse_system_message_content(raw_text)
                 elif local_type == 49:
                     parsed = _parse_app_message(raw_text)
                     render_type = str(parsed.get("renderType") or "text")
@@ -4412,6 +4638,13 @@ def list_chat_messages(
         wcdb_display_names = {}
         wcdb_avatar_urls = {}
 
+    group_nicknames = _load_group_nickname_map(
+        account_dir=account_dir,
+        contact_db_path=contact_db_path,
+        chatroom_id=username,
+        sender_usernames=uniq_senders,
+    )
+
     for m in merged:
         # If appmsg doesn't provide sourcedisplayname, try mapping sourceusername to display name.
         if (not str(m.get("from") or "").strip()) and str(m.get("fromUsername") or "").strip():
@@ -4427,13 +4660,12 @@ def list_chat_messages(
         su = str(m.get("senderUsername") or "")
         if not su:
             continue
-        row = sender_contact_rows.get(su)
-        display_name = _pick_display_name(row, su)
-        if display_name == su:
-            wd = str(wcdb_display_names.get(su) or "").strip()
-            if wd and wd != su:
-                display_name = wd
-        m["senderDisplayName"] = display_name
+        m["senderDisplayName"] = _resolve_sender_display_name(
+            sender_username=su,
+            sender_contact_rows=sender_contact_rows,
+            wcdb_display_names=wcdb_display_names,
+            group_nicknames=group_nicknames,
+        )
         avatar_url = base_url + _avatar_url_unified(
             account_dir=account_dir,
             username=su,
@@ -4930,19 +5162,24 @@ async def _search_chat_messages_via_fts(
             username=username,
             local_avatar_usernames=local_avatar_usernames,
         )
+        group_nicknames = _load_group_nickname_map(
+            account_dir=account_dir,
+            contact_db_path=contact_db_path,
+            chatroom_id=username,
+            sender_usernames=[str(x.get("senderUsername") or "") for x in hits],
+        )
 
         for h in hits:
             su = str(h.get("senderUsername") or "").strip()
             h["conversationName"] = conv_name
             h["conversationAvatar"] = conv_avatar
             if su:
-                row = contact_rows.get(su)
-                display_name = _pick_display_name(row, su) if row is not None else (conv_name if su == username else su)
-                if display_name == su:
-                    wd = str(wcdb_display_names.get(su) or "").strip()
-                    if wd and wd != su:
-                        display_name = wd
-                h["senderDisplayName"] = display_name
+                h["senderDisplayName"] = _resolve_sender_display_name(
+                    sender_username=su,
+                    sender_contact_rows=contact_rows,
+                    wcdb_display_names=wcdb_display_names,
+                    group_nicknames=group_nicknames,
+                )
                 avatar_url = base_url + _avatar_url_unified(
                     account_dir=account_dir,
                     username=su,
@@ -4986,6 +5223,23 @@ async def _search_chat_messages_via_fts(
             wcdb_display_names = {}
             wcdb_avatar_urls = {}
 
+        group_senders_by_room: dict[str, list[str]] = {}
+        for h in hits:
+            cu = str(h.get("username") or "").strip()
+            su = str(h.get("senderUsername") or "").strip()
+            if (not cu.endswith("@chatroom")) or (not su):
+                continue
+            group_senders_by_room.setdefault(cu, []).append(su)
+
+        group_nickname_cache: dict[str, dict[str, str]] = {}
+        for cu, senders in group_senders_by_room.items():
+            group_nickname_cache[cu] = _load_group_nickname_map(
+                account_dir=account_dir,
+                contact_db_path=contact_db_path,
+                chatroom_id=cu,
+                sender_usernames=senders,
+            )
+
         for h in hits:
             cu = str(h.get("username") or "").strip()
             su = str(h.get("senderUsername") or "").strip()
@@ -5003,13 +5257,12 @@ async def _search_chat_messages_via_fts(
             )
             h["conversationAvatar"] = conv_avatar
             if su:
-                row = contact_rows.get(su)
-                display_name = _pick_display_name(row, su) if row is not None else (conv_name if su == cu else su)
-                if display_name == su:
-                    wd = str(wcdb_display_names.get(su) or "").strip()
-                    if wd and wd != su:
-                        display_name = wd
-                h["senderDisplayName"] = display_name
+                h["senderDisplayName"] = _resolve_sender_display_name(
+                    sender_username=su,
+                    sender_contact_rows=contact_rows,
+                    wcdb_display_names=wcdb_display_names,
+                    group_nicknames=group_nickname_cache.get(cu, {}),
+                )
                 avatar_url = base_url + _avatar_url_unified(
                     account_dir=account_dir,
                     username=su,
@@ -5272,13 +5525,23 @@ async def search_chat_messages(
         contact_rows = _load_contact_rows(contact_db_path, uniq_usernames)
         conv_row = contact_rows.get(username)
         conv_name = _pick_display_name(conv_row, username)
+        group_nicknames = _load_group_nickname_map(
+            account_dir=account_dir,
+            contact_db_path=contact_db_path,
+            chatroom_id=username,
+            sender_usernames=[str(x.get("senderUsername") or "") for x in page],
+        )
 
         for h in page:
             su = str(h.get("senderUsername") or "").strip()
             h["conversationName"] = conv_name
             if su:
-                row = contact_rows.get(su)
-                h["senderDisplayName"] = _pick_display_name(row, su) if row is not None else (conv_name if su == username else su)
+                h["senderDisplayName"] = _resolve_sender_display_name(
+                    sender_username=su,
+                    sender_contact_rows=contact_rows,
+                    wcdb_display_names={},
+                    group_nicknames=group_nicknames,
+                )
 
         return {
             "status": "success",
@@ -5360,6 +5623,23 @@ async def search_chat_messages(
     )
     contact_rows = _load_contact_rows(contact_db_path, uniq_contacts)
 
+    group_senders_by_room: dict[str, list[str]] = {}
+    for h in page:
+        cu = str(h.get("username") or "").strip()
+        su = str(h.get("senderUsername") or "").strip()
+        if (not cu.endswith("@chatroom")) or (not su):
+            continue
+        group_senders_by_room.setdefault(cu, []).append(su)
+
+    group_nickname_cache: dict[str, dict[str, str]] = {}
+    for cu, senders in group_senders_by_room.items():
+        group_nickname_cache[cu] = _load_group_nickname_map(
+            account_dir=account_dir,
+            contact_db_path=contact_db_path,
+            chatroom_id=cu,
+            sender_usernames=senders,
+        )
+
     for h in page:
         cu = str(h.get("username") or "").strip()
         su = str(h.get("senderUsername") or "").strip()
@@ -5367,8 +5647,12 @@ async def search_chat_messages(
         conv_name = _pick_display_name(crow, cu) if cu else ""
         h["conversationName"] = conv_name or cu
         if su:
-            row = contact_rows.get(su)
-            h["senderDisplayName"] = _pick_display_name(row, su) if row is not None else (conv_name if su == cu else su)
+            h["senderDisplayName"] = _resolve_sender_display_name(
+                sender_username=su,
+                sender_contact_rows=contact_rows,
+                wcdb_display_names={},
+                group_nicknames=group_nickname_cache.get(cu, {}),
+            )
 
     return {
         "status": "success",
diff --git a/src/wechat_decrypt_tool/routers/chat_media.py b/src/wechat_decrypt_tool/routers/chat_media.py
index fcf4e31..7be8bb5 100644
--- a/src/wechat_decrypt_tool/routers/chat_media.py
+++ b/src/wechat_decrypt_tool/routers/chat_media.py
@@ -688,6 +688,83 @@ def _lookup_resource_md5_by_server_id(account_dir_str: str, server_id: int, want
             pass
 
 
+@lru_cache(maxsize=4096)
+def _lookup_image_md5_by_server_id_from_messages(account_dir_str: str, server_id: int, username: str) -> str:
+    account_dir_str = str(account_dir_str or "").strip()
+    username = str(username or "").strip()
+    if not account_dir_str or not username:
+        return ""
+
+    try:
+        sid = int(server_id or 0)
+    except Exception:
+        sid = 0
+    if not sid:
+        return ""
+
+    try:
+        chat_hash = hashlib.md5(username.encode()).hexdigest()
+    except Exception:
+        return ""
+    if not chat_hash:
+        return ""
+
+    table_name = f"Msg_{chat_hash}"
+    account_dir = Path(account_dir_str)
+
+    db_paths: list[Path] = []
+    try:
+        for p in account_dir.glob("message_*.db"):
+            try:
+                if p.is_file():
+                    db_paths.append(p)
+            except Exception:
+                continue
+    except Exception:
+        db_paths = []
+
+    if not db_paths:
+        return ""
+    db_paths.sort(key=lambda p: p.name)
+
+    for db_path in db_paths:
+        try:
+            conn = sqlite3.connect(str(db_path))
+        except Exception:
+            continue
+
+        try:
+            row = conn.execute(
+                f"SELECT local_type, packed_info_data FROM {table_name} "
+                "WHERE server_id = ? ORDER BY create_time DESC LIMIT 1",
+                (sid,),
+            ).fetchone()
+        except Exception:
+            row = None
+        finally:
+            try:
+                conn.close()
+            except Exception:
+                pass
+
+        if not row:
+            continue
+
+        try:
+            local_type = int(row[0] or 0)
+        except Exception:
+            local_type = 0
+        if local_type != 3:
+            continue
+
+        md5 = _extract_md5_from_packed_info(row[1])
+        md5_norm = str(md5 or "").strip().lower()
+        if _is_valid_md5(md5_norm):
+            return md5_norm
+
+    return ""
+
+
 def _is_safe_http_url(url: str) -> bool:
     u = str(url or "").strip()
     if not u:
@@ -1062,6 +1139,12 @@ async def get_chat_image(
         resource_md5 = _lookup_resource_md5_by_server_id(str(account_dir), int(server_id), want_local_type=3)
         if resource_md5:
             md5 = resource_md5
+        elif username:
+            md5_from_msg = _lookup_image_md5_by_server_id_from_messages(
+                str(account_dir), int(server_id), str(username)
+            )
+            if md5_from_msg:
+                md5 = md5_from_msg
 
     # md5 模式：优先从解密资源目录读取（更快）
     if md5:
diff --git a/src/wechat_decrypt_tool/wcdb_realtime.py b/src/wechat_decrypt_tool/wcdb_realtime.py
index ce954d6..bc02be9 100644
--- a/src/wechat_decrypt_tool/wcdb_realtime.py
+++ b/src/wechat_decrypt_tool/wcdb_realtime.py
@@ -102,6 +102,17 @@ def _load_wcdb_lib() -> ctypes.CDLL:
         lib.wcdb_get_group_members.argtypes = [ctypes.c_int64, ctypes.c_char_p, ctypes.POINTER(ctypes.c_char_p)]
         lib.wcdb_get_group_members.restype = ctypes.c_int
 
+        # Optional (newer DLLs): wcdb_get_group_nicknames(handle, chatroom_id, out_json)
+        try:
+            lib.wcdb_get_group_nicknames.argtypes = [
+                ctypes.c_int64,
+                ctypes.c_char_p,
+                ctypes.POINTER(ctypes.c_char_p),
+            ]
+            lib.wcdb_get_group_nicknames.restype = ctypes.c_int
+        except Exception:
+            pass
+
         # Optional: execute arbitrary SQL on a selected database kind/path.
         # Signature: wcdb_exec_query(handle, kind, path, sql, out_json)
         try:
@@ -355,6 +366,41 @@ def get_avatar_urls(handle: int, usernames: list[str]) -> dict[str, str]:
     return {}
 
 
+def get_group_members(handle: int, chatroom_id: str) -> list[dict[str, Any]]:
+    _ensure_initialized()
+    lib = _load_wcdb_lib()
+    cid = str(chatroom_id or "").strip()
+    if not cid:
+        return []
+    out_json = _call_out_json(lib.wcdb_get_group_members, ctypes.c_int64(int(handle)), cid.encode("utf-8"))
+    decoded = _safe_load_json(out_json)
+    if isinstance(decoded, list):
+        out: list[dict[str, Any]] = []
+        for x in decoded:
+            if isinstance(x, dict):
+                out.append(x)
+        return out
+    return []
+
+
+def get_group_nicknames(handle: int, chatroom_id: str) -> dict[str, str]:
+    _ensure_initialized()
+    lib = _load_wcdb_lib()
+    fn = getattr(lib, "wcdb_get_group_nicknames", None)
+    if not fn:
+        return {}
+
+    cid = str(chatroom_id or "").strip()
+    if not cid:
+        return {}
+
+    out_json = _call_out_json(fn, ctypes.c_int64(int(handle)), cid.encode("utf-8"))
+    decoded = _safe_load_json(out_json)
+    if isinstance(decoded, dict):
+        return {str(k): str(v) for k, v in decoded.items()}
+    return {}
+
+
 def exec_query(handle: int, *, kind: str, path: Optional[str], sql: str) -> list[dict[str, Any]]:
     """Execute raw SQL on a specific db kind/path via WCDB.