使用来自 jellyfin 的 django 进行视频流传输

Question

我用 jellyfin 服务器上的 django 制作了一个 streaming 应用程序，我使用视频 js 来播放我的视频媒体。问题是，当我尝试播放超过 1 小时且更长的视频时，会花费疯狂的时间，似乎视频是先下载然后播放的，而且我无法推进媒体并将其放在我的位置想，我首先怀疑有一个名为 preload 的 video js 选项，所以我尝试了值 'none'，'metadata' 和 'auto'，但没有任何结果实际上，一位同事给了我一个猜测也许 django 不支持这种功能！谁能证实这一点吗？ django 的解决方案是什么？

例如，这是我获取剧集类型电影的功能：

views.py

def episode_detail(request, episode_id):
    user_id = request.session.get('jellyfin_user_id')
    access_token = request.session.get('jellyfin_access_token')

    if not user_id or not access_token:
        return redirect('jellyfin_login')

    # try:
    #     user = JellyfinUser.objects.get(jellyfin_user_id=user_id)
    # except JellyfinUser.DoesNotExist:
    #     return redirect('jellyfin_login')

    headers = {
        'X-Emby-Token': settings.JELLYFIN_API_KEY
    }

    episode_url = f"{settings.JELLYFIN_URL}/Users/{user_id}/Items/{episode_id}"
    episode_response = requests.get(episode_url, headers=headers)

    if episode_response.status_code != 200:
        return render(request, 'error.html', context={'error': 'Unable to fetch episode details'})

    episode = episode_response.json()

    if episode['Type'] == 'Movie':

        people = episode.get('People', [])
        if people is None:
            people = []

        genres_url = f"{settings.JELLYFIN_URL}/Genres?ParentId={episode_id}"
        genres_response = requests.get(genres_url, headers=headers)
        genres = genres_response.json().get('Items', []) if genres_response.status_code == 200 else []
        episode['Genres'] = [genre['Name'] for genre in genres]

        studio_url = f"{settings.JELLYFIN_URL}/Studios?ParentId={episode_id}"
        studio_response = requests.get(studio_url, headers=headers)
        studios = studio_response.json().get('Items', []) if studio_response.status_code == 200 else []
        episode['Studios'] = [studio['Name'] for studio in studios]

        directors = [person['Name'] for person in people if person['Type'] == 'Director']
        episode['Directors'] = directors

        actors = []
        for person in people:
            if person['Type'] == 'Actor':

                actor_image_url = f"{settings.JELLYFIN_URL}/Persons/{person['Name']}/Images/Primary"
                actor_image_url_response = requests.get(actor_image_url, headers=headers)

                if actor_image_url_response.status_code == 200:
                    person['ImageURL'] = actor_image_url
                else:
                    person['ImageURL'] = static("library/images/default/default-user.jpg")

                actor_info = {
                    'name': person['Name'],
                    'role': person['Role'],
                    'image': person['ImageURL'],
                    'id': person['Id'],
                }
                actors.append(actor_info)

        episode['Actors'] = actors

        writers = [person['Name'] for person in people if person['Type'] == 'Writer']
        episode['Writers'] = writers

        producers = [person['Name'] for person in people if person['Type'] == 'Producer']
        episode['Producers'] = producers

        similar_url = f"{settings.JELLYFIN_URL}/Items/{episode_id}/Similar"
        similar_response = requests.get(similar_url, headers=headers)
        similar_episodes = similar_response.json().get('Items', []) if similar_response.status_code == 200 else []
        for item in similar_episodes:
            item['ImageURL'] = f"{settings.JELLYFIN_URL}/Items/{item['Id']}/Images/Primary"
        episode['SimilarEpisodes'] = similar_episodes

        episode['VideoURL'] = f"{settings.JELLYFIN_URL}/Videos/{episode['Id']}/stream"
        episode['ImageURL'] = f"{settings.JELLYFIN_URL}/Items/{episode['Id']}/Images/Primary"
        episode['BackdropURL'] = f"{settings.JELLYFIN_URL}/Items/{episode['Id']}/Images/Backdrop"
        episode['CommunityRating'] = episode.get('CommunityRating', 'N/A')
        episode['Overview'] = episode.get('Overview', 'Aucune description disponible')

        episode['Duration'] = format_duration(episode['RunTimeTicks'])
            print("NO FILE")

    user_data = episode.get('UserData', {})
    if user_data is None:
        user_data = {}

    episode['IsFavorite'] = user_data.get('IsFavorite', False)

    print(episode['Type'])

    return render(request, 'episode_detail.html', context={'episode': episode})

这是我的 Episode_detail.html：

episode_detail.html

{% extends 'base.html' %}

{% block title %} {{ episode.Name }} {% endblock %}

{% block style %}

        <!-- My Styles -->

{% endblock %}

{% block content %}

<div class="container mb-5">

    {% if episode.Type == 'Movie' %}

    <div class="row">

        <!-- AFFICHE DU FILM -->
        <div class="col-md-4">
            <img src="{{ episode.ImageURL }}" class="img-fluid movie-poster shadow-lg" alt="{{ episode.Name }}">
        </div>

        <!-- INFOS DU FILM -->
        <div class="col-md-8">

            <h1 class="text-white">{{ episode.Name }}</h1>

            <div class="d-flex align-items-center">
                <span class="production-year">{{ episode.ProductionYear }}</span>
                <span class="duration">{{ episode.Duration }}</span>
                <span class="note-commu">
                    <span>{{ episode.CommunityRating|floatformat:1 }}</span>
                    <span class="blue-star"><i class="fa-solid fa-star"></i></span>
                </span>
            </div>

            <!-- SYNOPSIS DU FILM -->
            {% if episode.Overview %}
            <div class="mt-3">
                <h4 class="section-title"><b>Synopsis</b></h4>
                <p>{{ episode.Overview }}</p>
            </div>
            {% endif %}

            <!-- DETAILS DU FILM -->
            <div class="mt-3">

                {% if episode.Genres or episode.Studios or episode.Directors or episode.Actors %}
                    <h4 class="section-title"><b>Details</b></h4>
                {% endif %}

                {% if episode.Genres %}<p><strong>Genre:</strong> {{ episode.Genres|join:", " }}</p>{% endif %}
                {% if episode.Studios %}<p><strong>Studios:</strong> {{ episode.Studios|join:", " }}</p>{% endif %}
                {% if episode.Directors %}<p><strong>Réalisateurs :</strong> {{episode.Directors|join:", " }}</p>{% endif %}
                {% if episode.Writers %}<p><strong>Scénaristes :</strong> {{episode.Writers|join:", " }}</p>{% endif %}
                {% if episode.Producers %}<p><strong>Producteurs:</strong> {{ episode.Producers|join:", " }}</p>{% endif %}
                {% if episode.Actors_name %}<p><strong>Acteurs:</strong> {{ episode.Actors_name|join:", " }}</p>{% endif %}
            </div>


            <!-- SECTION VIDEO DU FILM -->
            <div class="mt-5">
                <video
                        id="my-movie-video"
                        class="video-js vjs-big-play-centered vjs-fluid vjs-fill"
                        controls
                        preload="none"
                        poster="{{ episode.BackdropURL }}"
                        data-setup="{}"
                >
                    <source src="{{ episode.VideoURL }}" type="video/mp4" />
                    <source src="{{ episode.VideoURL }}" type="video/webm" />
                    <source src="{{ episode.VideoURL }}" type="video/avi" />
                    <source src="{{ episode.VideoURL }}" type="video/mkv" />
                    <p class="vjs-no-js">
                        To view this video please enable JavaScript, and consider upgrading to a
                        web browser that
                        <a href="https://videojs.com/html5-video-support/" target="_blank">supports HTML5 video</a>
                    </p>
                </video>
            </div>

        </div>
    </div>
</div>

{% endblock %}

Answer 1

“似乎视频先下载然后播放需要疯狂的时间”

检查您的 MP4 的元数据是否位于前面（而不是文件的后面）。

您可以使用
此站点 检查 MP4（向下滚动到 MP4 原子位置 部分）。 如果它显示
“未优化”，则元数据位于后面（并且还需要完整下载）。

您可以使用 FFmpeg 等工具进行修复，甚至可以在 Google 中搜索：

MP4 moov 快速启动修复建议。

“我无法将介质推进并放到我想要的位置，””

当您自己提供字节（通过流式传输）时，您必须自己处理查找。

这意味着监听
seek 事件（由 HTML5 视频标签触发）并为该查找点提供新字节。您的 MP4 必须是碎片 MP4才能通过字节流轻松查找。当您的应用程序获得一个搜索点（以秒为单位）时，您将计算需要读取哪个文件（如果按时间分隔）或哪个字节范围（如果所有视频持续时间都在一个文件内）。您的流媒体现在改为发送这些字节。这些字节将是一个 MP4 片段，表示请求的“查找”时间的数据。