From 8b27681a37c052efa33ae04118a5e086c558cc6d Mon Sep 17 00:00:00 2001
From: Florent Thiery <florent.thiery@ubicast.eu>
Date: Fri, 13 Dec 2024 17:42:50 +0100
Subject: [PATCH] add script which produces stats about source file resolutions
 (#40)

---
 examples/report_video_resolutions.py | 80 ++++++++++++++++++++++++++++
 1 file changed, 80 insertions(+)
 create mode 100755 examples/report_video_resolutions.py

diff --git a/examples/report_video_resolutions.py b/examples/report_video_resolutions.py
new file mode 100755
index 0000000..1f43dd6
--- /dev/null
+++ b/examples/report_video_resolutions.py
@@ -0,0 +1,80 @@
+#!/usr/bin/env python3
+'''
+Script which will produce stats about the video files on the platform
+'''
+import os
+import sys
+
+
+def format_seconds(seconds):
+    m, s = divmod(seconds, 60)
+    h, m = divmod(m, 60)
+    timecode = '%d:%02d:%02d' % (h, m, s)
+    return timecode
+
+
+def format_bytes(size):
+    power = 1000
+    n = 0
+    power_labels = {0: '', 1: 'kilo', 2: 'mega', 3: 'giga', 4: 'tera'}
+    while size > power:
+        size /= power
+        n += 1
+    return f'{round(size, 1)} {power_labels[n]}bytes'
+
+
+if __name__ == '__main__':
+    sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+    from ms_client.client import MediaServerClient
+
+    local_conf = sys.argv[1] if len(sys.argv) > 1 else None
+    msc = MediaServerClient(local_conf)
+    # ping
+    print('Dumping catalog')
+    videos = msc.get_catalog(fmt='json')['videos']
+
+    all_resources_duration = dict()
+    all_resources_count = dict()
+    all_resources_size = dict()
+
+    for index, video in enumerate(videos):
+        print(f'{index + 1}/{len(videos)}', end='\r')
+        oid = video['oid']
+        duration = int(video['duration_s'])  # in seconds
+        storage = int(video['storage_used'])  # in bytes
+        resources = msc.api('/medias/resources-list/', params={'oid': oid})["resources"]
+        resources_sorted = sorted(resources, key=lambda d: d["height"], reverse=True)
+        if len(resources_sorted) > 0 and duration:
+            source_resolution = resources_sorted[0]['height']  # first one should be the largest
+            all_resources_duration.setdefault(source_resolution, 0)
+            all_resources_duration[source_resolution] += duration
+            all_resources_count.setdefault(source_resolution, 0)
+            all_resources_count[source_resolution] += 1
+            all_resources_size.setdefault(source_resolution, 0)
+            all_resources_size[source_resolution] += storage
+
+    print()
+    print('Source resolutions by duration:')
+    all_resources_duration = dict(
+        sorted(all_resources_duration.items(), key=lambda item: item[1], reverse=True)
+    )
+    for mode, duration in all_resources_duration.items():
+        mode_size = all_resources_size[mode]
+        size_per_hour = int(mode_size / (duration / 3600))
+        print(f'{mode}: {format_seconds(duration)}, average size: {format_bytes(size_per_hour)} per hour')
+
+    print()
+    print('Source resolutions by count:')
+    all_resources_count = dict(
+        sorted(all_resources_count.items(), key=lambda item: item[1], reverse=True)
+    )
+    for mode, count in all_resources_count.items():
+        print(f'{mode}: {count}')
+
+    print()
+    print('Source resolutions by size:')
+    all_resources_size = dict(
+        sorted(all_resources_size.items(), key=lambda item: item[1], reverse=True)
+    )
+    for mode, size in all_resources_size.items():
+        print(f'{mode}: {format_bytes(size)}')