parsehub 2.1.7__tar.gz → 2.1.9__tar.gz

{parsehub-2.1.7/src/parsehub.egg-info → parsehub-2.1.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsehub
-Version: 2.1.7
+Version: 2.1.9
 Summary: 轻量、异步、开箱即用的社交媒体聚合解析库
 Author-email: 梓澪 <zilingmio@gmail.com>
 License: MIT
@@ -72,27 +72,27 @@ Dynamic: license-file
 
 ## 🌐 支持平台
 
-| 平台              | 视频 | 图文 | 其他            |
-|-----------------|:--:|:--:|---------------|
-| **Twitter / X** | ✅  | ✅  | 📝 文章         |
-| **Instagram**   | ✅  | ✅  |               |
-| **YouTube**     | ✅  |    | 🎵 音乐         |
-| **Facebook**    | ✅  |    |               |
-| **Threads**     | ✅  | ✅  |               |
-| **Bilibili**    | ✅  |    | 📝 动态         |
-| **抖音**          | ✅  | ✅  | ☀️日常          |
-| **TikTok**      | ✅  | ✅  |               |
-| **微博**          | ✅  | ✅  |               |
-| **小红书**         | ✅  | ✅  |               |
-| **贴吧**          | ✅  | ✅  |               |
-| **微信公众号**       |    | ✅  |               |
-| **快手**          | ✅  |    |               |
-| **酷安**          |    | ✅  |               |
-| **皮皮虾**         | ✅  | ✅  |               |
-| **最右**          | ✅  | ✅  |               |
-| **小黑盒**         | ✅  | ✅  |               |
-| **Snapchat**    | ✅  |    |               |
-| **知乎**          | ✅  | ✅  | 🐶 问答, 专栏, 圈子 |
+| 平台            | 视频 | 图文 | 其他                |
+|-----------------|:----:|:----:|---------------------|
+| **Twitter / X** |  ✅  |  ✅  | 📝 文章             |
+| **Instagram**   |  ✅  |  ✅  |                     |
+| **YouTube**     |  ✅  |      | 🎵 音乐             |
+| **Facebook**    |  ✅  |      |                     |
+| **Threads**     |  ✅  |  ✅  |                     |
+| **Bilibili**    |  ✅  |      | 📝 动态             |
+| **抖音**        |  ✅  |  ✅  | ☀️日常              |
+| **TikTok**      |  ✅  |  ✅  |                     |
+| **微博**        |  ✅  |  ✅  |                     |
+| **小红书**      |  ✅  |  ✅  |                     |
+| **贴吧**        |  ✅  |  ✅  |                     |
+| **微信公众号**  |      |  ✅  |                     |
+| **快手**        |  ✅  |  ✅  |                     |
+| **酷安**        |      |  ✅  |                     |
+| **皮皮虾**      |  ✅  |  ✅  |                     |
+| **最右**        |  ✅  |  ✅  |                     |
+| **小黑盒**      |  ✅  |  ✅  |                     |
+| **Snapchat**    |  ✅  |      |                     |
+| **知乎**        |  ✅  |  ✅  | 🐶 问答, 专栏, 圈子 |
 
 ## 📦 安装
 
@@ -134,14 +134,14 @@ ph d "https://example.com/post/1"
 
 #### 常用命令
 
-| 命令                                               | 说明          |
-|--------------------------------------------------|-------------|
-| `ph ls`                                          | 查看支持的平台     |
+| 命令                                             | 说明                   |
+|--------------------------------------------------|------------------------|
+| `ph ls`                                          | 查看支持的平台         |
 | `ph set proxy <platform> <proxy>`                | 设置解析代理和下载代理 |
-| `ph set proxy <platform> <proxy> --for download` | 只设置下载代理     |
-| `ph set cookie <platform>`                       | 保存平台 Cookie |
-| `ph set list`                                    | 查看配置列表      |
-| `ph set show <platform>`                         | 查看平台配置      |
+| `ph set proxy <platform> <proxy> --for download` | 只设置下载代理         |
+| `ph set cookie <platform>`                       | 保存平台 Cookie        |
+| `ph set list`                                    | 查看配置列表           |
+| `ph set show <platform>`                         | 查看平台配置           |
 
 配置会自动按平台应用到后续解析和下载; 临时覆盖时仍可直接传参数:
 
@@ -336,6 +336,7 @@ uv run pytest
 - [SocialSisterYi/bilibili-API-collect](https://github.com/SocialSisterYi/bilibili-API-collect)
 - [Nemo2011/bilibili-api](https://github.com/Nemo2011/bilibili-api)
 - [cv-cat/ZhihuApis](https://github.com/cv-cat/ZhihuApis)
+- [ucmao/media-parser](https://github.com/ucmao/media-parser)
 
 ## 📜 开源协议
 

{parsehub-2.1.7 → parsehub-2.1.9}/README.md

@@ -30,27 +30,27 @@
 
 ## 🌐 支持平台
 
-| 平台              | 视频 | 图文 | 其他            |
-|-----------------|:--:|:--:|---------------|
-| **Twitter / X** | ✅  | ✅  | 📝 文章         |
-| **Instagram**   | ✅  | ✅  |               |
-| **YouTube**     | ✅  |    | 🎵 音乐         |
-| **Facebook**    | ✅  |    |               |
-| **Threads**     | ✅  | ✅  |               |
-| **Bilibili**    | ✅  |    | 📝 动态         |
-| **抖音**          | ✅  | ✅  | ☀️日常          |
-| **TikTok**      | ✅  | ✅  |               |
-| **微博**          | ✅  | ✅  |               |
-| **小红书**         | ✅  | ✅  |               |
-| **贴吧**          | ✅  | ✅  |               |
-| **微信公众号**       |    | ✅  |               |
-| **快手**          | ✅  |    |               |
-| **酷安**          |    | ✅  |               |
-| **皮皮虾**         | ✅  | ✅  |               |
-| **最右**          | ✅  | ✅  |               |
-| **小黑盒**         | ✅  | ✅  |               |
-| **Snapchat**    | ✅  |    |               |
-| **知乎**          | ✅  | ✅  | 🐶 问答, 专栏, 圈子 |
+| 平台            | 视频 | 图文 | 其他                |
+|-----------------|:----:|:----:|---------------------|
+| **Twitter / X** |  ✅  |  ✅  | 📝 文章             |
+| **Instagram**   |  ✅  |  ✅  |                     |
+| **YouTube**     |  ✅  |      | 🎵 音乐             |
+| **Facebook**    |  ✅  |      |                     |
+| **Threads**     |  ✅  |  ✅  |                     |
+| **Bilibili**    |  ✅  |      | 📝 动态             |
+| **抖音**        |  ✅  |  ✅  | ☀️日常              |
+| **TikTok**      |  ✅  |  ✅  |                     |
+| **微博**        |  ✅  |  ✅  |                     |
+| **小红书**      |  ✅  |  ✅  |                     |
+| **贴吧**        |  ✅  |  ✅  |                     |
+| **微信公众号**  |      |  ✅  |                     |
+| **快手**        |  ✅  |  ✅  |                     |
+| **酷安**        |      |  ✅  |                     |
+| **皮皮虾**      |  ✅  |  ✅  |                     |
+| **最右**        |  ✅  |  ✅  |                     |
+| **小黑盒**      |  ✅  |  ✅  |                     |
+| **Snapchat**    |  ✅  |      |                     |
+| **知乎**        |  ✅  |  ✅  | 🐶 问答, 专栏, 圈子 |
 
 ## 📦 安装
 
@@ -92,14 +92,14 @@ ph d "https://example.com/post/1"
 
 #### 常用命令
 
-| 命令                                               | 说明          |
-|--------------------------------------------------|-------------|
-| `ph ls`                                          | 查看支持的平台     |
+| 命令                                             | 说明                   |
+|--------------------------------------------------|------------------------|
+| `ph ls`                                          | 查看支持的平台         |
 | `ph set proxy <platform> <proxy>`                | 设置解析代理和下载代理 |
-| `ph set proxy <platform> <proxy> --for download` | 只设置下载代理     |
-| `ph set cookie <platform>`                       | 保存平台 Cookie |
-| `ph set list`                                    | 查看配置列表      |
-| `ph set show <platform>`                         | 查看平台配置      |
+| `ph set proxy <platform> <proxy> --for download` | 只设置下载代理         |
+| `ph set cookie <platform>`                       | 保存平台 Cookie        |
+| `ph set list`                                    | 查看配置列表           |
+| `ph set show <platform>`                         | 查看平台配置           |
 
 配置会自动按平台应用到后续解析和下载; 临时覆盖时仍可直接传参数:
 
@@ -294,6 +294,7 @@ uv run pytest
 - [SocialSisterYi/bilibili-API-collect](https://github.com/SocialSisterYi/bilibili-API-collect)
 - [Nemo2011/bilibili-api](https://github.com/Nemo2011/bilibili-api)
 - [cv-cat/ZhihuApis](https://github.com/cv-cat/ZhihuApis)
+- [ucmao/media-parser](https://github.com/ucmao/media-parser)
 
 ## 📜 开源协议
 

{parsehub-2.1.7 → parsehub-2.1.9}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "parsehub"
-version = "2.1.7"
+version = "2.1.9"
 description = "轻量、异步、开箱即用的社交媒体聚合解析库"
 readme = "README.md"
 requires-python = ">=3.12.0"

{parsehub-2.1.7 → parsehub-2.1.9}/src/parsehub/parsers/base/ytdlp.py

@@ -395,6 +395,13 @@ class YtParser(BaseParser, register=False):
         except Exception as e:
             raise ParseError(f"解析视频信息失败: {str(e)}") from e
 
+        if dl.get("live_status") in (
+            "is_live",
+            "is_upcoming",
+            "was_live",
+        ):
+            raise ParseError("不支持直播")
+
         if dl.get("_type") == "playlist":
             entries = dl.get("entries") or []
             if not entries:

{parsehub-2.1.7 → parsehub-2.1.9}/src/parsehub/parsers/parser/bilibili.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 import re
 from pathlib import Path
 from typing import Any, cast
-from urllib.parse import parse_qs, urlparse
+from urllib.parse import parse_qs, parse_qsl, urlencode, urlparse
 
 from loguru import logger
 
@@ -76,7 +76,10 @@ class BiliParse(BaseParser):
         if self._is_bvid(url):
             return f"https://www.bilibili.com/video/{url}"
         else:
-            return await super().get_raw_url(url, **kwargs)
+            raw_url = await super().get_raw_url(url, **kwargs)
+            u = urlparse(raw_url)
+            q = urlencode([(k, v) for k, v in parse_qsl(u.query) if (k, v) != ("p", "1")])
+            return u._replace(query=q).geturl()
 
     @staticmethod
     async def is_dynamic(url: str) -> str | None:

parsehub-2.1.9/src/parsehub/parsers/parser/kuaishou.py

@@ -0,0 +1,89 @@
+from loguru import logger
+
+from ...provider_api.kuaishou import KuaiShouAPI, KuaishouParser
+from ...types import ImageParseResult, ImageRef, ParseError, Platform, VideoParseResult, VideoRef
+from ...utils.helpers import SecretCookie
+from ..base.base import BaseParser
+
+COOKIE = SecretCookie(
+    "kpf=PC_WEB; clientid=3; did=web_bfbcdb2f5b3dc663a745deabafcf61e6; kwpsecproductname=kuaishou-"
+    "vision; didv=1773330035000; kwpsecproductname=kuaishou-vision; userId=446442483; kuaishou.serv"
+    "er.webday7_st=ChprdWFpc2hvdS5zZXJ2ZXIud2ViZGF5Ny5zdBKwAeuBbGjVcz39sj4G7d7P54r9C1etC_QftYb2I1X"
+    "Mg01WSbw9NefL7E6EmwkYxHf70B9BM3Oyk20kFv1Y0xnRcfHtGNHYUHkmKguP6cvFeACofr2zPAZYRchRkndIBk5qExOl"
+    "kr4FSoGpY-WqXeibapHNEbfZTLZl_QkQA4aGWotSZpBMv6wR3RxZWiMv60xc-CIndGICJbbRAaRGZNxz7QBj2Mr-SeU2o"
+    "0bVi7esnD1AGhKquV16S9dezebl5ZuYo_R_JKgiIAidQF8n526Yos_GTgm3KrGknnEbkK-NMiNvTw3YBehZKAUwAQ; kua"
+    "ishou.server.webday7_ph=f3720606882f1d7a76ab1ab52a489c4d44a1; bUserId=1000583835422; ktrace-c"
+    "ontext=1|MS44Nzg0NzI0NTc4Nzk2ODY5Ljg3MTE4OTQ4LjE3NzM1NzExNTEyMjQuNDQ0OTc1MTI=|MS44Nzg0NzI0NTc4"
+    "Nzk2ODY5LjUxNTU3MjM4LjE3NzM1NzExNTEyMjQuNDQ0OTc1MTM=|0|webservice-user-growth-node|webservice|"
+    "true|src-Js; kwssectoken=BIjmefxxiTpXOdz9/RQ6Gl7cR5/0J7xaPzJ18udJgBSLTrJy4O7LhrYtbeeHGW+AOJrI6"
+    "P8LQnioDWSuuQxV8Q==; kwscode=75d440673de879734b8700f363119968b4fabb4eb0369b1607e206d8e8c1ac9d;"
+    " kpn=KUAISHOU_VISION; kwfv1=PnGU+9+Y8008S+nH0U+0mjPf8fP08f+98f+nLlwnrIP9P9G98YPf8jPBQSweS0+nr9"
+    "G0mD8B+fP/L98/qlPe4f8eDI8f8jwBGh8BPAPfLEGALhGf+f+AYj+e4jPfLl+AY0G/cI+/Q0G0DEPfc98/mjw/pSPBbjGA"
+    "rh8erl+ezfG/HlP0zf+0b0+n+DGnpj+0HI+9Qj+0p0PeDF+ADIPeL7+W==; kwssectoken=IMLS/eg005i6IUbIoIB/7W"
+    "Byh8ciKMPUXULQ3a5/m3dK5D9ez8He/oMP2QLhil52v7Bk3O0CO2g6t5R/5XjSCw==; kwscode=75d440673de879734b"
+    "8700f363119968b4fabb4eb0369b1607e206d8e8c1ac9d"
+)
+
+
+class KuaiShouParser(BaseParser):
+    __platform__ = Platform.KUAISHOU
+    __supported_type__ = ["视频", "图文"]
+    __match__ = r"^(http(s)?://)?(www|v|live|v\.m)\.(kuaishou|chenzhongtech).com/.+"
+    __redirect_keywords__ = ["v.kuaishou", "/f/"]
+
+    async def _do_parse(self, raw_url: str) -> VideoParseResult | ImageParseResult:
+        ksp = await KuaishouParser.create(
+            raw_url, proxy=self.proxy, cookie=self.cookie.get_value() or COOKIE.get_value()
+        )
+        cover = ksp.get_cover_photo_url()
+        if ksp.page_type == "VIDEO":
+            try:
+                content = ksp.get_title_content()
+                video = ksp.get_real_video_url()
+            except Exception as e:
+                logger.error(f"快手 HTML 解析失败, 尝试 API 解析: {e}")
+            else:
+                if not video:
+                    raise ParseError("-1 快手解析失败")
+
+                return VideoParseResult(
+                    title=content,
+                    video=VideoRef(
+                        url=video.url,
+                        thumb_url=cover,
+                        duration=video.d,
+                        height=video.h,
+                        width=video.w,
+                    ),
+                )
+
+            ks = KuaiShouAPI(self.cookie.get_value() or COOKIE.get_value(), self.proxy)
+            try:
+                result = await ks.get_video_info(raw_url)
+            except Exception as e:
+                raise ParseError(f"-2 快手解析失败: {e}") from e
+            else:
+                return VideoParseResult(
+                    title=result.title,
+                    video=VideoRef(
+                        url=result.video_url,
+                        thumb_url=result.thumb_url,
+                        duration=result.duration,
+                        height=result.height,
+                        width=result.width,
+                    ),
+                )
+        else:
+            content = ksp.get_title_content()
+            img = ksp.get_image_list()
+            if img:
+                return ImageParseResult(
+                    title=content, photo=[ImageRef(url=i.url, height=i.h, width=i.w, ext=i.e) for i in img]
+                )
+            if cover:
+                return ImageParseResult(title=content, photo=[ImageRef(url=cover)])
+
+            raise ParseError("快手解析失败 -3")
+
+
+__all__ = ["KuaiShouParser"]

{parsehub-2.1.7 → parsehub-2.1.9}/src/parsehub/parsers/parser/twitter.py

@@ -39,13 +39,13 @@ class TwitterParser(BaseParser):
         try:
             tweet = await x.fetch_tweet(url)
         except Exception as e:
-            if any(s in str(e) for s in ("error -2",)):
+            if any(s in str(e) for s in ("error -2", "error -3")):
                 if cookie := self.cookie.get_value():
                     x2 = Twitter(self.proxy, cookie=cookie)
                     try:
                         tweet = await x2.fetch_tweet(url)
                     except Exception as e2:
-                        raise ParseError(f"Twitter 账号可能已被封禁\n\n使用的Cookie: {self.cookie}") from e2
+                        raise ParseError(f"Twitter 账号无权限或已被封禁\n\n使用的 Cookie: {self.cookie}") from e2
                 else:
                     raise ParseError(str(e)) from e
             else: