parsehub 2.2.2__tar.gz → 2.2.4__tar.gz

{parsehub-2.2.2/src/parsehub.egg-info → parsehub-2.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsehub
-Version: 2.2.2
+Version: 2.2.4
 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

{parsehub-2.2.2 → parsehub-2.2.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "parsehub"
-version = "2.2.2"
+version = "2.2.4"
 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-2.2.2 → parsehub-2.2.4}/src/parsehub/__init__.py

@@ -179,6 +179,8 @@ class ParseHub:
             raise UnknownPlatform(url)
         try:
             return await parser(proxy=proxy).get_raw_url(url, clean_all=clean_all)
+        except ParseError:
+            raise
         except Exception as e:
             raise ParseError from e
 

{parsehub-2.2.2 → parsehub-2.2.4}/src/parsehub/parsers/parser/xhs.py

@@ -14,6 +14,7 @@ from ...types import (
     VideoParseResult,
     VideoRef,
 )
+from ...utils.helpers import UA
 from ..base import BaseParser
 
 
@@ -24,6 +25,19 @@ class XHSParser(BaseParser):
     __redirect_keywords__ = ["xhslink"]
     __after_clean_parameters__ = ["xsec_token"]
 
+    async def get_raw_url(self, url: str, *, clean_all: bool = False, headers: dict | None = None) -> str:
+        headers = {
+            "User-Agent": UA,
+            "accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,"
+            "*/*;q=0.8,application/signed-exchange;v=b3;q=0.7",
+        }
+        url = await super().get_raw_url(url, clean_all=clean_all, headers=headers)
+        if "/login" in url:
+            raise ParseError("该帖子需要登录后查看")
+        elif "/404" in url:
+            raise ParseError("帖子不存在")
+        return url
+
     async def _do_parse(self, raw_url: str) -> Union["VideoParseResult", "ImageParseResult", "MultimediaParseResult"]:
         xhs = XHSAPI(proxy=self.proxy, cookie=self.cookie.get_value())
         result = await xhs.extract(raw_url)

{parsehub-2.2.2 → parsehub-2.2.4}/src/parsehub/provider_api/xhs.py

@@ -10,9 +10,8 @@ from typing import Any, cast
 import httpx
 from bs4 import BeautifulSoup
 
-from parsehub.utils.helpers import UA
-
 from ..errors import ParseError
+from ..utils.helpers import UA
 
 
 class XHSAPI:
@@ -24,7 +23,11 @@ class XHSAPI:
         async with httpx.AsyncClient(proxy=self.proxy, cookies=self.cookie, follow_redirects=True) as client:
             result = await client.get(
                 url,
-                headers={"User-Agent": UA},
+                headers={
+                    "User-Agent": UA,
+                    "accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp"
+                    ",image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7",
+                },
                 timeout=30,
             )
             if "/login" in str(result.url):

{parsehub-2.2.2 → parsehub-2.2.4/src/parsehub.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsehub
-Version: 2.2.2
+Version: 2.2.4
 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

{parsehub-2.2.2 → parsehub-2.2.4}/test/test_core_offline.py

@@ -660,6 +660,7 @@ class TestPlatformUrlMatching(unittest.TestCase):
                 "https://www.zhihu.com/question/2057559076813510452",
                 "https://zhuanlan.zhihu.com/p/1989096494578558904",
                 "https://www.zhihu.com/question/597674895/answer/3004370705",
+                "https://daily.zhihu.com/story/9792555",
             ],
         }