feat(python): 完成 Python 全量迁移——剩余平台、登录通道与公共层修复

- 新增 5sing/apple/xm/me/qt/lr 平台包与取流实现（含 xm-sign、MissEvan 签名、蜻蜓 HMAC、Apple developer token）
- 新增 wy/kg/tx/lr 登录：扫码、手机号、账密（kg 20028 风控指纹）、cookie 绑定、kg/tx 续期
- request 层：保留多值 Set-Cookie；JSON body 紧凑化（修复 kg 网关验签）；直连不经系统代理（对齐 Node）
- 登录路由统一 Body(embed=True)，修复扫码轮询 422；lifespan 接入登录续期调度
- ruff/format/pyright 全绿；14 平台真实联网与 HTTP 路由验证通过

app/sdk/platforms/kw/comment.py

@@ -0,0 +1,89 @@
+"""酷我评论（对应 sdk/kw/comment.ts）。
+
+`ncomment.kuwo.cn/com.s` 只认安卓 Dalvik UA；`type=get_comment` 取最新评论、
+`type=get_rec_comment` 取热评，两者响应字段名不同（comments/comments_counts
+vs hot_comments/hot_comments_counts）。
+"""
+
+from __future__ import annotations
+
+import math
+from typing import Any
+from urllib.parse import unquote
+
+from app.sdk.common.request import http_fetch
+from app.sdk.common.utils import date_format2
+
+COMMENT_API = (
+    "http://ncomment.kuwo.cn/com.s?f=web&type={type}&aapiver=1&prod=kwplayer_ar_10.5.2.0"
+    "&digest=15&sid={songmid}&start={start}&msgflag=1&count={count}&newver=3&uid=0"
+)
+
+HEADERS = {"User-Agent": "Dalvik/2.1.0 (Linux; U; Android 9;)"}
+
+
+def _to_int(value: Any) -> int:
+    try:
+        return int(float(value))
+    except (TypeError, ValueError):
+        return 0
+
+
+def _reply_comment(item: dict[str, Any]) -> dict[str, Any]:
+    return {
+        "id": item.get("id"),
+        "text": item.get("msg"),
+        "time": item.get("time"),
+        "timeStr": date_format2(_to_int(item.get("time")) * 1000),
+        "userName": item.get("u_name"),
+        "avatar": item.get("u_pic"),
+        "userId": item.get("u_id"),
+        "likedCount": item.get("like_num"),
+        "images": [item["mpic"]] if item.get("mpic") else [],
+    }
+
+
+def _filter_comment(raw_list: list[dict[str, Any]] | None) -> list[dict[str, Any]]:
+    if not raw_list:
+        return []
+    result: list[dict[str, Any]] = []
+    for item in raw_list:
+        data = _reply_comment(item)
+        data["images"] = [unquote(str(item["mpic"]))] if item.get("mpic") else []
+        # 顶层评论多一个楼中楼数组；原实现对父级 mpic 做 decodeURIComponent，子级不做
+        data["reply"] = [_reply_comment(child) for child in item.get("child_comments") or []]
+        result.append(data)
+    return result
+
+
+async def _get_comment(songmid: str, page: int, limit: int, *, hot: bool) -> dict[str, Any]:
+    comment_type = "get_rec_comment" if hot else "get_comment"
+    url = COMMENT_API.format(type=comment_type, songmid=songmid, start=limit * (page - 1), count=limit)
+    resp = await http_fetch(url, headers=HEADERS)
+    body = resp.body if isinstance(resp.body, dict) else {}
+    if resp.status_code != 200 or str(body.get("code")) != "200":
+        raise RuntimeError("获取热门评论失败" if hot else "获取评论失败")
+
+    if hot:
+        total = _to_int(body.get("hot_comments_counts"))
+        comments = _filter_comment(body.get("hot_comments"))
+    else:
+        total = _to_int(body.get("comments_counts"))
+        comments = _filter_comment(body.get("comments"))
+    return {
+        "source": "kw",
+        "comments": comments,
+        "total": total,
+        "page": page,
+        "limit": limit,
+        # 原实现用 `Math.ceil(total/limit) || 1` 兜住 0 条
+        "maxPage": math.ceil(total / limit) if limit and total else 1,
+    }
+
+
+async def get_comment(song_info: dict[str, Any], page: int = 1, limit: int = 20) -> dict[str, Any]:
+    return await _get_comment(str(song_info.get("songmid") or ""), page, limit, hot=False)
+
+
+async def get_hot_comment(song_info: dict[str, Any], page: int = 1, limit: int = 100) -> dict[str, Any]:
+    return await _get_comment(str(song_info.get("songmid") or ""), page, limit, hot=True)