parsehub 2.3.8__tar.gz → 2.3.9__tar.gz

{parsehub-2.3.8/src/parsehub.egg-info → parsehub-2.3.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsehub
-Version: 2.3.8
+Version: 2.3.9
 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

{parsehub-2.3.8 → parsehub-2.3.9}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "parsehub"
-version = "2.3.8"
+version = "2.3.9"
 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-2.3.8 → parsehub-2.3.9}/src/parsehub/parsers/parser/xiaoheihe.py

@@ -54,7 +54,23 @@ class XiaoHeiHeParser(BaseParser):
         images: list[ImageRef | AniRef] = []
         for media in xhh.media or []:
             if media.type == XiaoHeiHeMediaType.IMAGE:
-                images.append(ImageRef(url=media.url, width=media.width or 0, height=media.height or 0))
+                images.append(
+                    ImageRef(
+                        url=media.url,
+                        thumb_url=media.thumb_url,
+                        ext=media.ext or "jpg",
+                        width=media.width or 0,
+                        height=media.height or 0,
+                    )
+                )
             else:
-                images.append(AniRef(url=media.url, width=media.width or 0, height=media.height or 0))
+                images.append(
+                    AniRef(
+                        url=media.url,
+                        thumb_url=media.thumb_url,
+                        ext=media.ext or "gif",
+                        width=media.width or 0,
+                        height=media.height or 0,
+                    )
+                )
         return images

{parsehub-2.3.8 → parsehub-2.3.9}/src/parsehub/provider_api/xiaoheihe.py

@@ -23,6 +23,11 @@ from cryptography.hazmat.primitives.ciphers.modes import CBC, ECB
 from markdownify import MarkdownConverter
 
 
+def original_image_url(url: str) -> str:
+    path = url.split("?", 1)[0]
+    return re.sub(r"/thumb\.(\w+)$", r".\1", path)
+
+
 class XiaoHeiHePostType(Enum):
     VIDEO = "video"
     IMAGE = "image"  # 图文, 单张或多张图片 + 文案
@@ -40,6 +45,7 @@ class XiaoHeiHeMedia:
     type: XiaoHeiHeMediaType
     url: str
     thumb_url: str | None = None
+    ext: str | None = None
     height: int | None = None
     width: int | None = None
 
@@ -89,11 +95,14 @@ class XiaoHeiHeAPI:
             for image in text_list[1:]:
                 if image["type"] == "img":
                     media_type = XiaoHeiHeMediaType.GIF if ".gif" in image["url"] else XiaoHeiHeMediaType.IMAGE
+                    url = original_image_url(image["url"])
+                    ext_match = re.search(r"\.([^./]+)$", url)
                     images.append(
                         XiaoHeiHeMedia(
                             type=media_type,
-                            url=image["url"],
+                            url=url,
                             thumb_url=image["url"],
+                            ext=ext_match.group(1) if ext_match else None,
                             height=int(float(image.get("height", 0))),
                             width=int(float(image.get("width", 0))),
                         )
@@ -134,7 +143,7 @@ class XiaoHeiHeAPI:
             "owner_only": "1",
             **sig_params,
         }
-        cookies = {"x_xhh_tokenid": await SecuritySm.get_d_id()}
+        cookies = {"x_xhh_tokenid": await SecuritySm.get_d_id(self.proxy)}
         async with httpx.AsyncClient(proxy=self.proxy) as cli:
             result = await cli.get(self.api_url + "/bbs/app/link/tree", params=params, cookies=cookies)
             result.raise_for_status()
@@ -355,6 +364,8 @@ class XHHConverter(MarkdownConverter):
     def convert_img(self, el, _, parent_tags):
         alt = el.attrs.get("alt", None) or ""
         src = el.attrs.get("data-original", None) or ""
+        if src:
+            src = original_image_url(src)
         title = el.attrs.get("title", None) or ""
         title_part = ' "{}"'.format(title.replace('"', r"\"")) if title else ""
         options = cast(dict[str, Any], getattr(self, "options"))  # noqa: B009
@@ -504,7 +515,7 @@ class SecuritySm:
         return v + smsk_web + "0"
 
     @classmethod
-    async def get_d_id(cls):
+    async def get_d_id(cls, proxy: str | None = None):
         uid = str(uuid.uuid4()).encode("utf-8")
         priId = hashlib.md5(uid).hexdigest()[0:16]
         encrypted_uid = cls.PK.encrypt(uid, padding.PKCS1v15())
@@ -531,7 +542,7 @@ class SecuritySm:
         des_target["tn"] = hashlib.md5(cls.get_tn(des_target).encode()).hexdigest()
 
         des_result = cls._AES(cls.GZIP(cls._DES(des_target)), priId.encode("utf-8"))
-        async with httpx.AsyncClient() as client:
+        async with httpx.AsyncClient(proxy=proxy) as client:
             response = await client.post(
                 cls.DEVICES_INFO_URL,
                 json={

{parsehub-2.3.8 → parsehub-2.3.9/src/parsehub.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsehub
-Version: 2.3.8
+Version: 2.3.9
 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