index streams and file size

tubearchivist/home/src/es/index_mapping.json

@@ -146,6 +146,9 @@
                     "type": "keyword",
                     "index": false
                 },
+                "media_size": {
+                    "type": "long"
+                },
                 "tags": {
                     "type": "text",
                     "analyzer": "english",
@@ -239,6 +242,30 @@
                         }
                     }
                 },
+                "streams": {
+                    "properties": {
+                        "type": {
+                            "type": "keyword",
+                            "index": false
+                        },
+                        "index": {
+                            "type": "short",
+                            "index": false
+                        },
+                        "codec": {
+                            "type": "text"
+                        },
+                        "width": {
+                            "type": "short"
+                        },
+                        "height": {
+                            "type": "short"
+                        },
+                        "bitrate": {
+                            "type": "integer"
+                        }
+                    }
+                },
                 "sponsorblock": {
                     "properties": {
                         "last_refresh": {

tubearchivist/home/src/index/video.py

@@ -16,7 +16,10 @@ from home.src.index import playlist as ta_playlist
 from home.src.index.generic import YouTubeItem
 from home.src.index.subtitle import YoutubeSubtitle
 from home.src.index.video_constants import VideoTypeEnum
-from home.src.index.video_streams import DurationConverter
+from home.src.index.video_streams import (
+    DurationConverter,
+    MediaStreamExtractor,
+)
 from home.src.ta.helper import clean_string, randomizor
 from home.src.ta.ta_redis import RedisArchivist
 from ryd_client import ryd_client
@@ -153,6 +156,7 @@ class YoutubeVideo(YouTubeItem, YoutubeSubtitle):
         self._add_stats()
         self.add_file_path()
         self.add_player(media_path)
+        self.add_streams(media_path)
         if self.config["downloads"]["integrate_ryd"]:
             self._get_ryd_stats()
 
@@ -254,6 +258,17 @@ class YoutubeVideo(YouTubeItem, YoutubeSubtitle):
             }
         )
 
+    def add_streams(self, media_path=False):
+        """add stream metadata"""
+        vid_path = self._get_vid_path(media_path)
+        media = MediaStreamExtractor(vid_path)
+        self.json_data.update(
+            {
+                "streams": media.extract_metadata(),
+                "media_size": media.get_file_size(),
+            }
+        )
+
     def _get_vid_path(self, media_path=False):
         """get path of media file"""
         if media_path:

tubearchivist/home/src/index/video_streams.py

@@ -1,6 +1,8 @@
 """extract metadata from video streams"""
 
+import json
 import subprocess
+from os import stat
 
 
 class DurationConverter:
@@ -52,3 +54,79 @@ class DurationConverter:
             duration_str = duration_str + "00:"
         duration_str = duration_str + str(secs).zfill(2)
         return duration_str
+
+
+class MediaStreamExtractor:
+    """extract stream metadata"""
+
+    def __init__(self, media_path):
+        self.media_path = media_path
+        self.metadata = []
+
+    def extract_metadata(self):
+        """entry point to extract metadata"""
+
+        cmd = [
+            "ffprobe",
+            "-v",
+            "quiet",
+            "-print_format",
+            "json",
+            "-show_streams",
+            "-show_format",
+            self.media_path,
+        ]
+        result = subprocess.run(
+            cmd, capture_output=True, text=True, check=False
+        )
+
+        if result.returncode != 0:
+            return self.metadata
+
+        streams = json.loads(result.stdout).get("streams")
+        for stream in streams:
+            self.process_stream(stream)
+
+        return self.metadata
+
+    def process_stream(self, stream):
+        """parse stream to metadata"""
+        codec_type = stream.get("codec_type")
+        if codec_type == "video":
+            self._extract_video_metadata(stream)
+        elif codec_type == "audio":
+            self._extract_audio_metadata(stream)
+        else:
+            return
+
+    def _extract_video_metadata(self, stream):
+        """parse video metadata"""
+        if "bit_rate" not in stream:
+            # is probably thumbnail
+            return
+
+        self.metadata.append(
+            {
+                "type": "video",
+                "index": stream["index"],
+                "codec": stream["codec_name"],
+                "width": stream["width"],
+                "height": stream["height"],
+                "bitrate": int(stream["bit_rate"]),
+            }
+        )
+
+    def _extract_audio_metadata(self, stream):
+        """extract audio metadata"""
+        self.metadata.append(
+            {
+                "type": "audio",
+                "index": stream["index"],
+                "codec": stream["codec_name"],
+                "bitrate": int(stream["bit_rate"]),
+            }
+        )
+
+    def get_file_size(self):
+        """get filesize in bytes"""
+        return stat(self.media_path).st_size

tubearchivist/home/templates/home/video.html

@@ -56,20 +56,6 @@
                 {% else %}
                     <p>Youtube: Deactivated</p>
                 {% endif %}
-                {% if reindex %}
-                    <p>Reindex scheduled</p>
-                {% else %}
-                    <div id="reindex-button" class="button-box">
-                        <button data-id="{{ video.youtube_id }}" data-type="video" onclick="reindex(this)" title="Reindex {{ video.title }}">Reindex</button>
-                    </div>
-                {% endif %}
-                <div class="button-box">
-                    <a download="" href="/media/{{ video.media_url }}"><button id="download-item">Download File</button></a>
-                    <button onclick="deleteConfirm()" id="delete-item">Delete Video</button>
-                    <div class="delete-confirm" id="delete-button">
-                        <span>Are you sure? </span><button class="danger-button" onclick="deleteVideo(this)" data-id="{{ video.youtube_id }}" data-redirect = "{{ video.channel.channel_id }}">Delete</button> <button onclick="cancelDelete()">Cancel</button>
-                    </div>
-                </div>
             </div>
         </div>
         <div class="info-box-item">
@@ -89,6 +75,34 @@
             </div>
         </div>
     </div>
+    <div class="info-box info-box-2">
+        <div class="info-box-item">
+            <div class="button-box">
+                {% if reindex %}
+                    <p>Reindex scheduled</p>
+                {% else %}
+                    <div id="reindex-button" class="button-box">
+                        <button data-id="{{ video.youtube_id }}" data-type="video" onclick="reindex(this)" title="Reindex {{ video.title }}">Reindex</button>
+                    </div>
+                {% endif %}
+                <a download="" href="/media/{{ video.media_url }}"><button id="download-item">Download File</button></a>
+                <button onclick="deleteConfirm()" id="delete-item">Delete Video</button>
+                <div class="delete-confirm" id="delete-button">
+                    <span>Are you sure? </span><button class="danger-button" onclick="deleteVideo(this)" data-id="{{ video.youtube_id }}" data-redirect = "{{ video.channel.channel_id }}">Delete</button> <button onclick="cancelDelete()">Cancel</button>
+                </div>
+            </div>
+        </div>
+        <div class="info-box-item">
+            {% if video.media_size %}
+                <p>File size: {{ video.media_size|filesizeformat }}</p>
+            {% endif %}
+            {% if video.streams %}
+                {% for stream in video.streams %}
+                    <p>{{ stream.type|title }}: {{ stream.codec }} {{ stream.bitrate|filesizeformat }}/s{% if stream.width %} <span class="space-carrot">|</span> {{ stream.width }}x{{ stream.height}}{% endif %}</p>
+                {% endfor %}
+            {% endif %}
+        </div>
+    </div>
     {% if video.tags %}
         <div class="description-box">
             <div class="video-tag-box">

tubearchivist/static/css/style.css

@@ -660,6 +660,10 @@ video:-webkit-full-screen {
     background-color: var(--highlight-bg);
 }
 
+.info-box-item p {
+    width: 100%;
+}
+
 .description-text {
     width: 100%;
 }