chore: 处理成功的日志往外提一级

.github/workflows/docker-image-debug.yml

@@ -0,0 +1,29 @@
+name: Docker Image CI (DEBUG)
+
+on:
+    push:
+        branches:
+            - main
+
+jobs:
+    build:
+        runs-on: ubuntu-latest
+        steps:
+            -
+                name: Checkout
+                uses: actions/checkout@v3
+            -
+                name: Login to DockerHub
+                uses: docker/login-action@v3
+                with:
+                    username: ${{ secrets.DOCKERHUB_USERNAME }}
+                    password: ${{ secrets.DOCKERHUB_TOKEN }}
+            -
+                name: Build and push images
+                uses: docker/build-push-action@v5
+                with:
+                    context: .
+                    file: ./Dockerfile
+                    push: true
+                    tags: |
+                        ${{ secrets.DOCKERHUB_USERNAME }}/bili-sync:debug

.gitignore

@@ -4,7 +4,7 @@ debug.py
 videos
 config.test.json
 database.test.db*
-example.json
+example*.json
 thumbs.test
 config
 data

README.md

@@ -1,4 +1,6 @@
-# bili-sync
+![bili-sync](https://socialify.git.ci/amtoaer/bili-sync/image?description=1&font=KoHo&issues=1&language=1&logo=https%3A%2F%2Fs2.loli.net%2F2023%2F12%2F02%2F9EwT2yInOu1d3zm.png&name=1&owner=1&pattern=Signal&pulls=1&stargazers=1&theme=Light)
+
+## 简介
 
 为 NAS 用户编写的 BILIBILI 收藏夹同步工具，可方便导入 EMBY 等媒体库工具浏览。
 
@@ -37,15 +39,6 @@ class Config(DataClassJsonMixin):
 
 即：我们可以通过运行一次程序，等程序写入初始配置并提示配置错误终止后编辑 `config.json` 文件，编辑后即可重新运行。
 
-## 关于 UP 头像
-
-目前开放全局的环境变量 `THUMB_PATH` 作为 up 主头像的存储位置。
-
-在下载某条视频时，如果 UP 的头像还不存在，就会将 UP 的头像下载至 `THUMB_PATH`，同时在视频的 NFO 文件中写入 UP 头像的绝对路径。
-
-但实际测试下来，EMBY 似乎无法正常读取 NFO 文件中的本地头像路径，待找到处理办法后再修复。
-
-> 虽然但是，一个基本的逻辑是，如果期望 `bili-sync` 在 NFO 中写入的头像绝对路径能够被 EMBY 读取到，那么两个容器中头像的绝对路径必须完全相同。因此虽然头像还没办法正常加载，但为后续考虑，还是推荐将 THUMB_PATH 填写上，并确保该路径在 `bili-sync` 和 `emby` 两个容器中指向的是相同的文件夹（也就是把一个文件夹同时挂载到 `bili-sync` 和 `emby` 的 THUMB_PATH 下）。
 
 ## Docker 运行示例
 
@@ -60,17 +53,16 @@ services:
       - /home/amtoaer/Videos/Bilibilis/:/Videos/Bilibilis/  # 视频文件
       - /home/amtoaer/.config/nas/bili-sync/config/:/app/config/  # 配置文件
       - /home/amtoaer/.config/nas/bili-sync/data/:/app/data/  # 数据库
+      # 注：如需在 emby 内查看 up 主头像，需要将 emby 的 metadata/people/ 配置目录挂载至容器的 /app/thumb/
+      - /home/amtoaer/.config/nas/emby/metadata/people/:/app/thumb/
     environment:
-      - THUMB_PATH=/Videos/Bilibilis/thumb/  # 将头像放到视频文件的 thumb 文件夹下
+      - TZ=Asia/Shanghai
     restart: always
     network_mode: bridge
     hostname: bili-sync
     container_name: bili-sync
     logging:
-      driver: "json-file"
-      options:
-        max-size: "30m"
-
+      driver: "local"
 ```
 
 对应的配置文件：
@@ -92,9 +84,19 @@ services:
 }
 ```
 
-## 目前的问题
+## 支持的额外命令
 
-- [ ] 研究一下 NFO，看看怎么正常读取本地的演员头像
+为满足需要，该应用包含几个单独的命令，可在程序目录下使用 `python entry.py ${command name}` 运行。
+
+1. `once` 
+
+    处理收藏夹，和一般定时任务触发时执行的操作完全相同，但仅运行一次。
+2. `recheck` 
+
+    将本地不存在的视频文件标记成未下载，下次定时任务触发时将一并下载。
+3. `upper_thumb`
+
+    手动触发全量下载 up 主头像，为使用老版本时下载的没有 up 头像的视频添加头像。
 
 ## 路线图
 

commands.py

@@ -1,10 +1,11 @@
 import asyncio
 
-from aiofiles.os import path
 from loguru import logger
 
 from constants import MediaStatus, MediaType
-from models import FavoriteItem
+from models import FavoriteItem, Upper
+from processor import download_content, process_video
+from utils import aexists, amakedirs, aremove
 
 
 async def recheck():
@@ -14,9 +15,7 @@ async def recheck():
         status=MediaStatus.NORMAL,
         downloaded=True,
     )
-    exists = await asyncio.gather(
-        *[path.exists(item.video_path) for item in items]
-    )
+    exists = await asyncio.gather(*[aexists(item.video_path) for item in items])
     for item, exist in zip(items, exists):
         if isinstance(exist, Exception):
             logger.error(
@@ -36,3 +35,54 @@ async def recheck():
     logger.info("Updating database...")
     await FavoriteItem.bulk_update(items, fields=["downloaded"])
     logger.info("Database updated.")
+
+
+async def upper_thumb():
+    """将up主的头像批量写入数据库，从不支持up主头像的版本升级上来后需要手动调用一次"""
+    makedir_tasks = []
+    other_tasks = []
+    for upper in await Upper.all():
+        if all(
+            await asyncio.gather(
+                aexists(upper.thumb_path), aexists(upper.meta_path)
+            )
+        ):
+            logger.info(
+                "Upper {} {} already exists, skipped.", upper.mid, upper.name
+            )
+        makedir_tasks.append(amakedirs(upper.thumb_path.parent, exist_ok=True))
+        logger.info("Saving metadata for upper {} {}...", upper.mid, upper.name)
+        other_tasks.extend(
+            [
+                upper.save_metadata(),
+                download_content(upper.thumb, upper.thumb_path),
+            ]
+        )
+    await asyncio.gather(*makedir_tasks)
+    await asyncio.gather(*other_tasks)
+    logger.info("All done.")
+
+
+async def refresh_tags():
+    """刷新已存在的视频的标签，从不支持标签的版本升级上来后需要手动调用一次"""
+    items = await FavoriteItem.filter(
+        downloaded=True,
+        tags=None,
+    ).prefetch_related("upper")
+    await asyncio.gather(
+        *[aremove(item.nfo_path) for item in items],
+        return_exceptions=True,
+    )
+    await asyncio.gather(
+        *[
+            process_video(
+                item,
+                process_poster=False,
+                process_video=False,
+                process_nfo=True,
+                process_upper=False,
+            )
+            for item in items
+        ],
+        return_exceptions=True,
+    )

entry.py

@@ -4,7 +4,7 @@ import sys
 import uvloop
 from loguru import logger
 
-from commands import recheck
+from commands import recheck, refresh_tags, upper_thumb
 from models import init_model
 from processor import cleanup, process
 from settings import settings
@@ -14,16 +14,16 @@ asyncio.set_event_loop_policy(uvloop.EventLoopPolicy())
 
 async def entry() -> None:
     await init_model()
-    if any("once" in _ for _ in sys.argv):
-        # 单次运行
-        logger.info("Running once...")
-        await process()
-        return
-    if any("recheck" in _ for _ in sys.argv):
-        # 重新检查
-        logger.info("Rechecking...")
-        await recheck()
-        return
+    for command, func in [
+        ("once", process),
+        ("recheck", recheck),
+        ("upper_thumb", upper_thumb),
+        ("refresh_tags", refresh_tags),
+    ]:
+        if any(command in _ for _ in sys.argv):
+            logger.info("Running {}...", command)
+            await func()
+            return
     logger.info("Running daemon...")
     while True:
         await process()

migrations/models/2_20231204003326_update.py

@@ -0,0 +1,11 @@
+from tortoise import BaseDBAsyncClient
+
+
+async def upgrade(db: BaseDBAsyncClient) -> str:
+    return """
+        ALTER TABLE "favoriteitem" ADD "tags" JSON;"""
+
+
+async def downgrade(db: BaseDBAsyncClient) -> str:
+    return """
+        ALTER TABLE "favoriteitem" DROP COLUMN "tags";"""

models.py

@@ -13,6 +13,7 @@ from constants import (
     MediaType,
 )
 from settings import settings
+from utils import aopen
 
 
 class FavoriteList(Model):
@@ -39,7 +40,31 @@ class Upper(Model):
 
     @property
     def thumb_path(self) -> Path:
-        return DEFAULT_THUMB_PATH / f"{self.mid}.jpg"
+        return (
+            DEFAULT_THUMB_PATH / str(self.mid)[0] / f"{self.mid}" / "folder.jpg"
+        )
+
+    @property
+    def meta_path(self) -> Path:
+        return (
+            DEFAULT_THUMB_PATH / str(self.mid)[0] / f"{self.mid}" / "person.nfo"
+        )
+
+    async def save_metadata(self):
+        async with aopen(self.meta_path, "w") as f:
+            await f.write(
+                f"""
+<?xml version="1.0" encoding="utf-8" standalone="yes"?>
+<person>
+  <plot />
+  <outline />
+  <lockdata>false</lockdata>
+  <dateadded>{self.created_at.strftime("%Y-%m-%d %H:%M:%S")}</dateadded>
+  <title>{self.mid}</title>
+  <sorttitle>{self.mid}</sorttitle>
+</person>
+""".strip()
+            )
 
 
 class FavoriteItem(Model):
@@ -54,6 +79,7 @@ class FavoriteItem(Model):
     bvid = fields.CharField(max_length=255)
     desc = fields.TextField()
     cover = fields.TextField()
+    tags = fields.JSONField(null=True)
     favorite_list = fields.ForeignKeyField(
         "models.FavoriteList", related_name="items"
     )

nfo.py

@@ -2,19 +2,19 @@ import datetime
 from dataclasses import dataclass
 from pathlib import Path
 
+from utils import aopen
+
 
 @dataclass
 class Actor:
     name: str
     role: str
-    thumb: Path
 
     def to_xml(self) -> str:
         return f"""
     <actor>
         <name>{self.name}</name>
         <role>{self.role}</role>
-        <thumb>{self.thumb.resolve()}</thumb>
     </actor>
 """.strip(
             "\n"
@@ -25,16 +25,22 @@ class Actor:
 class EpisodeInfo:
     title: str
     plot: str
+    tags: list[str]
     actor: list[Actor]
     bvid: str
     aired: datetime.datetime
 
-    def write_nfo(self, path: Path) -> None:
-        with path.open("w", encoding="utf-8") as f:
-            f.write(self.to_xml())
+    async def write_nfo(self, path: Path) -> None:
+        async with aopen(path, "w", encoding="utf-8") as f:
+            await f.write(self.to_xml())
 
     def to_xml(self) -> str:
         actor = "\n".join(_.to_xml() for _ in self.actor)
+        tags = (
+            "\n".join(f"    <genre>{_}</genre>" for _ in self.tags)
+            if isinstance(self.tags, list)
+            else ""
+        )
         return f"""
 <?xml version="1.0" encoding="utf-8" standalone="yes"?>
 <episodedetails>
@@ -43,6 +49,7 @@ class EpisodeInfo:
     <title>{self.title}</title>
 {actor}
     <year>{self.aired.year}</year>
+{tags}
     <uniqueid type="bilibili">{self.bvid}</uniqueid>
     <aired>{self.aired.strftime("%Y-%m-%d")}</aired>
 </episodedetails>

processor.py

@@ -2,11 +2,8 @@ import asyncio
 import datetime
 from asyncio import Semaphore, create_subprocess_exec
 from asyncio.subprocess import DEVNULL
-from pathlib import Path
 
-import aiofiles
-import httpx
-from bilibili_api import HEADERS, favorite_list, video
+from bilibili_api import favorite_list, video
 from bilibili_api.exceptions import ResponseCodeException
 from loguru import logger
 from tortoise import Tortoise
@@ -16,8 +13,9 @@ from credential import credential
 from models import FavoriteItem, FavoriteList, Upper
 from nfo import Actor, EpisodeInfo
 from settings import settings
+from utils import aexists, amakedirs, client, download_content
 
-client = httpx.AsyncClient(headers=HEADERS)
+anchor = datetime.date.today()
 
 
 async def cleanup() -> None:
@@ -38,16 +36,6 @@ def concurrent_decorator(concurrency: int) -> callable:
     return decorator
 
 
-async def download_content(url: str, path: Path) -> None:
-    async with client.stream("GET", url) as resp, aiofiles.open(
-        path, "wb"
-    ) as f:
-        async for chunk in resp.aiter_bytes(40960):
-            if not chunk:
-                return
-            await f.write(chunk)
-
-
 async def manage_model(medias: list[dict], fav_list: FavoriteList) -> None:
     uppers = [
         Upper(
@@ -93,16 +81,16 @@ async def manage_model(medias: list[dict], fav_list: FavoriteList) -> None:
 
 async def process() -> None:
     global anchor
-    if not await credential.check_valid():
-        logger.error("Credential is invalid, skipped.")
-        return
-    if await credential.check_refresh():
-        try:
-            await credential.refresh()
-            logger.info("Credential refreshed.")
-        except Exception:
-            logger.exception("Failed to refresh credential.")
-            return
+    if (today := datetime.date.today()) > anchor:
+        anchor = today
+        logger.info("Check credential.")
+        if await credential.check_refresh():
+            try:
+                await credential.refresh()
+                logger.info("Credential refreshed.")
+            except Exception:
+                logger.exception("Failed to refresh credential.")
+                return
     for favorite_id in settings.favorite_ids:
         if favorite_id not in settings.path_mapper:
             logger.warning(
@@ -168,82 +156,141 @@ async def process_favorite(favorite_id: int) -> None:
 
 
 @concurrent_decorator(4)
-async def process_video(fav_item: FavoriteItem) -> None:
+async def process_video(
+    fav_item: FavoriteItem,
+    process_poster=True,
+    process_video=True,
+    process_nfo=True,
+    process_upper=True,
+) -> None:
     logger.info("Start to process video {} {}", fav_item.bvid, fav_item.name)
     if fav_item.type != MediaType.VIDEO:
         logger.warning("Media {} is not a video, skipped.", fav_item.name)
         return
+    v = video.Video(fav_item.bvid, credential=credential)
     try:
-        if fav_item.video_path.exists():
-            fav_item.downloaded = True
-            await fav_item.save()
-            logger.info(
-                "{} {} already exists, skipped.", fav_item.bvid, fav_item.name
-            )
-            return
-        # 写入 up 主头像
-        if not fav_item.upper.thumb_path.exists():
-            await download_content(
-                fav_item.upper.thumb, fav_item.upper.thumb_path
-            )
-        # 写入 nfo
-        EpisodeInfo(
-            title=fav_item.name,
-            plot=fav_item.desc,
-            actor=[
-                Actor(
-                    name=fav_item.upper.mid,
-                    role=fav_item.upper.name,
-                    thumb=fav_item.upper.thumb_path,
+        if process_upper:
+            # 写入 up 主头像
+            if not all(
+                await asyncio.gather(
+                    aexists(fav_item.upper.thumb_path),
+                    aexists(fav_item.upper.meta_path),
                 )
-            ],
-            bvid=fav_item.bvid,
-            aired=fav_item.ctime,
-        ).write_nfo(fav_item.nfo_path)
-        # 写入 poster
-        await download_content(fav_item.cover, fav_item.poster_path)
-        # 开始处理视频内容
-        v = video.Video(fav_item.bvid, credential=credential)
-        detector = video.VideoDownloadURLDataDetecter(
-            await v.get_download_url(page_index=0)
-        )
-        streams = detector.detect_best_streams()
-        if detector.check_flv_stream():
-            await download_content(streams[0].url, fav_item.tmp_video_path)
-            process = await create_subprocess_exec(
-                FFMPEG_COMMAND,
-                "-i",
-                str(fav_item.tmp_video_path),
-                str(fav_item.video_path),
-                stdout=DEVNULL,
-                stderr=DEVNULL,
-            )
-            await process.communicate()
-            fav_item.tmp_video_path.unlink()
-        else:
-            await asyncio.gather(
-                download_content(streams[0].url, fav_item.tmp_video_path),
-                download_content(streams[1].url, fav_item.tmp_audio_path),
-            )
-            process = await create_subprocess_exec(
-                FFMPEG_COMMAND,
-                "-i",
-                str(fav_item.tmp_video_path),
-                "-i",
-                str(fav_item.tmp_audio_path),
-                "-c",
-                "copy",
-                str(fav_item.video_path),
-                stdout=DEVNULL,
-                stderr=DEVNULL,
-            )
-            await process.communicate()
-            fav_item.tmp_video_path.unlink()
-            fav_item.tmp_audio_path.unlink()
-        fav_item.downloaded = True
-        await fav_item.save()
+            ):
+                await amakedirs(fav_item.upper.thumb_path.parent, exist_ok=True)
+                await asyncio.gather(
+                    fav_item.upper.save_metadata(),
+                    download_content(
+                        fav_item.upper.thumb, fav_item.upper.thumb_path
+                    ),
+                    return_exceptions=True,
+                )
+            else:
+                logger.info(
+                    "Upper {} {} already exists, skipped.",
+                    fav_item.upper.mid,
+                    fav_item.upper.name,
+                )
+        if process_nfo:
+            if not await aexists(fav_item.nfo_path):
+                if fav_item.tags is None:
+                    try:
+                        fav_item.tags = [
+                            _["tag_name"] for _ in await v.get_tags()
+                        ]
+                    except Exception:
+                        logger.exception(
+                            "Failed to get tags of video {} {}",
+                            fav_item.bvid,
+                            fav_item.name,
+                        )
+                # 写入 nfo
+                await EpisodeInfo(
+                    title=fav_item.name,
+                    plot=fav_item.desc,
+                    actor=[
+                        Actor(
+                            name=fav_item.upper.mid,
+                            role=fav_item.upper.name,
+                        )
+                    ],
+                    tags=fav_item.tags,
+                    bvid=fav_item.bvid,
+                    aired=fav_item.ctime,
+                ).write_nfo(fav_item.nfo_path)
+            else:
+                logger.info(
+                    "NFO of {} {} already exists, skipped.",
+                    fav_item.bvid,
+                    fav_item.name,
+                )
+        if process_poster:
+            # 写入 poster
+            if not await aexists(fav_item.poster_path):
+                await download_content(fav_item.cover, fav_item.poster_path)
+            else:
+                logger.info(
+                    "Poster of {} {} already exists, skipped.",
+                    fav_item.bvid,
+                    fav_item.name,
+                )
+        if process_video:
+            if await aexists(fav_item.video_path):
+                fav_item.downloaded = True
+                logger.info(
+                    "Video {} {} already exists, skipped.",
+                    fav_item.bvid,
+                    fav_item.name,
+                )
+            else:
+                # 开始处理视频内容
+                detector = video.VideoDownloadURLDataDetecter(
+                    await v.get_download_url(page_index=0)
+                )
+                streams = detector.detect_best_streams()
+                if detector.check_flv_stream():
+                    await download_content(
+                        streams[0].url, fav_item.tmp_video_path
+                    )
+                    process = await create_subprocess_exec(
+                        FFMPEG_COMMAND,
+                        "-i",
+                        str(fav_item.tmp_video_path),
+                        str(fav_item.video_path),
+                        stdout=DEVNULL,
+                        stderr=DEVNULL,
+                    )
+                    await process.communicate()
+                    fav_item.tmp_video_path.unlink()
+                else:
+                    await asyncio.gather(
+                        download_content(
+                            streams[0].url, fav_item.tmp_video_path
+                        ),
+                        download_content(
+                            streams[1].url, fav_item.tmp_audio_path
+                        ),
+                    )
+                    process = await create_subprocess_exec(
+                        FFMPEG_COMMAND,
+                        "-i",
+                        str(fav_item.tmp_video_path),
+                        "-i",
+                        str(fav_item.tmp_audio_path),
+                        "-c",
+                        "copy",
+                        str(fav_item.video_path),
+                        stdout=DEVNULL,
+                        stderr=DEVNULL,
+                    )
+                    await process.communicate()
+                    fav_item.tmp_video_path.unlink()
+                    fav_item.tmp_audio_path.unlink()
+                fav_item.downloaded = True
         logger.info(
-            "{} {} processed successfully.", fav_item.bvid, fav_item.name
+            "{} {} processed successfully.",
+            fav_item.bvid,
+            fav_item.name,
         )
     except ResponseCodeException as e:
         match e.code:
@@ -259,7 +306,6 @@ async def process_video(fav_item: FavoriteItem) -> None:
                     e.code,
                 )
                 return
-        await fav_item.save()
         logger.error(
             "Video {} {} is not available, marked as {}",
             fav_item.bvid,
@@ -270,3 +316,5 @@ async def process_video(fav_item: FavoriteItem) -> None:
         logger.exception(
             "Failed to process video {} {}", fav_item.bvid, fav_item.name
         )
+    finally:
+        await fav_item.save()

utils.py

@@ -0,0 +1,37 @@
+from pathlib import Path
+
+import aiofiles
+import httpx
+from aiofiles.base import AiofilesContextManager
+from aiofiles.os import makedirs, remove
+from aiofiles.ospath import exists
+from aiofiles.threadpool.text import AsyncTextIOWrapper
+from bilibili_api import HEADERS
+
+client = httpx.AsyncClient(headers=HEADERS)
+
+
+async def download_content(url: str, path: Path) -> None:
+    async with client.stream("GET", url) as resp, aopen(path, "wb") as f:
+        async for chunk in resp.aiter_bytes(40960):
+            if not chunk:
+                return
+            await f.write(chunk)
+
+
+async def aexists(path: Path) -> bool:
+    return await exists(path)
+
+
+async def amakedirs(path: Path, exist_ok=False) -> None:
+    await makedirs(path, exist_ok=exist_ok)
+
+
+def aopen(
+    path: Path, mode: str = "r", **kwargs
+) -> AiofilesContextManager[None, None, AsyncTextIOWrapper]:
+    return aiofiles.open(path, mode, **kwargs)
+
+
+async def aremove(path: Path) -> None:
+    await remove(path)