parsehub 2.0.8__tar.gz → 2.0.10__tar.gz

{parsehub-2.0.8/src/parsehub.egg-info → parsehub-2.0.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsehub
-Version: 2.0.8
+Version: 2.0.10
 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-2.0.8 → parsehub-2.0.10}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "parsehub"
-version = "2.0.8"
+version = "2.0.10"
 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-2.0.8 → parsehub-2.0.10}/src/parsehub/parsers/base/ytdlp.py

@@ -1,5 +1,4 @@
 import asyncio
-from concurrent.futures import ProcessPoolExecutor
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Union
@@ -17,8 +16,6 @@ from ...types import (
 )
 from .base import BaseParser
 
-EXC = ProcessPoolExecutor()
-
 
 def download_video(yto_params: dict, urls: list[str]) -> None:
     """在独立进程中下载视频"""
@@ -30,6 +27,15 @@ def download_video(yto_params: dict, urls: list[str]) -> None:
         raise RuntimeError(error_msg) from None
 
 
+def progress_hook(d: dict):
+    if d["status"] == "downloading":
+        downloaded = d.get("downloaded_bytes", 0)
+        total = d.get("total_bytes") or d.get("total_bytes_estimate") or 0
+        print(f"{downloaded}/{total}")
+    elif d["status"] == "finished":
+        print("下载完成，准备处理文件")
+
+
 class YtParser(BaseParser, register=False):
     """yt-dlp解析器"""
 
@@ -103,6 +109,7 @@ class YtParser(BaseParser, register=False):
             #     }
             # ],
             "playlist_items": "1",  # 分p列表默认解析第一个
+            # "progress_hooks": [progress_hook],
         }
         return params
 
@@ -143,17 +150,40 @@ class YtVideoParseResult(VideoParseResult):
         #     paramss["format"] = "worstvideo* + worstaudio / worst"
 
         if callback:
-            await callback(0, 1, "count", *callback_args, **callback_kwargs)
-
-        await self.__download(paramss)
+            loop = asyncio.get_running_loop()
+            progress_mode = "bytes"
+
+            def _progress_hook(d):
+                nonlocal progress_mode
+                if d["status"] == "downloading":
+                    # 已知问题: yt-dlp 返回的总进度不统一
+                    downloaded = int(d.get("downloaded_bytes", 0))
+                    total = int(d.get("total_bytes") or d.get("total_bytes_estimate") or 0)
+                    if total and progress_mode == "bytes":
+                        asyncio.run_coroutine_threadsafe(
+                            callback(downloaded, total, "bytes", *callback_args, **callback_kwargs),
+                            loop,
+                        )
+                    else:
+                        progress_mode = "count"
+                        asyncio.run_coroutine_threadsafe(
+                            callback(0, 1, "count", *callback_args, **callback_kwargs),
+                            loop,
+                        )
+                elif d["status"] == "finished" and progress_mode == "count":
+                    asyncio.run_coroutine_threadsafe(
+                        callback(1, 1, "count", *callback_args, **callback_kwargs),
+                        loop,
+                    )
+
+            paramss["progress_hooks"] = [_progress_hook]
+
+        await self._run_download(paramss)
 
         v = list(output_dir.glob("*.mp4")) or list(output_dir.glob("*.mkv")) or list(output_dir.glob("*.webm"))
         if not v:
             raise DownloadError("下载失败 -1")
 
-        if callback:
-            await callback(1, 1, "count", *callback_args, **callback_kwargs)
-
         video_path = v[0]
         return DownloadResult(
             VideoFile(
@@ -165,14 +195,13 @@ class YtVideoParseResult(VideoParseResult):
             output_dir,
         )
 
-    async def __download(self, paramss: dict, count: int = 0) -> None:
+    async def _run_download(self, paramss: dict, count: int = 0) -> None:
         if count > 2:
             raise DownloadError("下载失败 -2")
 
-        loop = asyncio.get_running_loop()
         try:
             await asyncio.wait_for(
-                loop.run_in_executor(EXC, download_video, paramss, [self.dl.url]),
+                asyncio.to_thread(download_video, paramss, [self.dl.url]),
                 timeout=300,
             )
         except TimeoutError as e:
@@ -187,7 +216,7 @@ class YtVideoParseResult(VideoParseResult):
                 )
             ):
                 paramss.pop("writeautomaticsub", None)
-                await self.__download(paramss, count + 1)
+                await self._run_download(paramss, count + 1)
 
         except Exception as e:
             raise DownloadError(f"下载失败: {str(e)}") from e

{parsehub-2.0.8 → parsehub-2.0.10}/src/parsehub/parsers/parser/coolapk.py

@@ -1,3 +1,4 @@
+import re
 from pathlib import Path
 from typing import Union
 
@@ -37,18 +38,26 @@ class CoolapkParser(BaseParser):
                 media=media,
                 markdown_content=coolapk.markdown_content,
             )
+        content = self.hashtag_handler(coolapk.text_content)
         if any(isinstance(m, AniRef) for m in media):
             return CoolapkMultimediaParseResult(
                 title=coolapk.title,
                 media=media,
-                content=coolapk.text_content,
+                content=content,
             )
         return CoolapkImageParseResult(
             title=coolapk.title,
             photo=media,
-            content=coolapk.text_content,
+            content=content,
         )
 
+    @staticmethod
+    def hashtag_handler(desc: str):
+        hashtags = re.findall(r" ?#[^#]+# ?", desc)
+        for hashtag in hashtags:
+            desc = desc.replace(hashtag, f" {hashtag.strip().removesuffix('#')} ")
+        return desc
+
 
 class CoolapkParseResult(ParseResult):
     async def _do_download(

{parsehub-2.0.8 → parsehub-2.0.10/src/parsehub.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsehub
-Version: 2.0.8
+Version: 2.0.10
 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