From 33ecd73137c9ecaaa91cae7a108bc542068e6cdc Mon Sep 17 00:00:00 2001
From: Simon <simobilleter@gmail.com>
Date: Mon, 22 Apr 2024 17:49:26 +0200
Subject: [PATCH] ffmpeg download script, separate build step

---
 Dockerfile                       | 32 +++++---------
 docker_assets/ffmpeg_download.py | 71 ++++++++++++++++++++++++++++++++
 2 files changed, 82 insertions(+), 21 deletions(-)
 create mode 100644 docker_assets/ffmpeg_download.py

diff --git a/Dockerfile b/Dockerfile
index f60e9e4..1f7f505 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,5 +1,5 @@
 # multi stage to build tube archivist
-# first stage to build python wheel, copy into final image
+# build python wheel, download and extract ffmpeg, copy into final image
 
 
 # First stage to build python wheel
@@ -13,6 +13,11 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
 COPY ./tubearchivist/requirements.txt /requirements.txt
 RUN pip install --user -r requirements.txt
 
+# build ffmpeg
+FROM python:3.11.3-slim-bullseye as ffmpeg-builder
+COPY docker_assets/ffmpeg_download.py ffmpeg_download.py
+RUN python ffmpeg_download.py $TARGETPLATFORM
+
 # build final image
 FROM python:3.11.3-slim-bullseye as tubearchivist
 
@@ -25,30 +30,15 @@ ENV PYTHONUNBUFFERED 1
 COPY --from=builder /root/.local /root/.local
 ENV PATH=/root/.local/bin:$PATH
 
+# copy ffmpeg
+COPY --from=ffmpeg-builder ./ffmpeg/ffmpeg /usr/bin/ffmpeg
+COPY --from=ffmpeg-builder ./ffprobe/ffprobe /usr/bin/ffprobe
+
 # install distro packages needed
 RUN apt-get clean && apt-get -y update && apt-get -y install --no-install-recommends \
     nginx \
     atomicparsley \
-    curl \
-    xz-utils && rm -rf /var/lib/apt/lists/*
-
-# install patched ffmpeg build, default to linux64
-RUN if [ "$TARGETPLATFORM" = "linux/arm64" ] ; then \
-    curl -s https://api.github.com/repos/yt-dlp/FFmpeg-Builds/releases/latest \
-        | grep browser_download_url \
-        | grep ".*master.*linuxarm64.*tar.xz" \
-        | cut -d '"' -f 4 \
-        | xargs curl -L --output ffmpeg.tar.xz ; \
-    else \
-    curl -s https://api.github.com/repos/yt-dlp/FFmpeg-Builds/releases/latest \
-        | grep browser_download_url \
-        | grep ".*master.*linux64.*tar.xz" \
-        | cut -d '"' -f 4 \
-        | xargs curl -L --output ffmpeg.tar.xz ; \
-    fi && \
-    tar -xf ffmpeg.tar.xz --strip-components=2 --no-anchored -C /usr/bin/ "ffmpeg" && \
-    tar -xf ffmpeg.tar.xz --strip-components=2 --no-anchored -C /usr/bin/ "ffprobe" && \
-    rm ffmpeg.tar.xz
+    curl && rm -rf /var/lib/apt/lists/*
 
 # install debug tools for testing environment
 RUN if [ "$INSTALL_DEBUG" ] ; then \
diff --git a/docker_assets/ffmpeg_download.py b/docker_assets/ffmpeg_download.py
new file mode 100644
index 0000000..9ab7f82
--- /dev/null
+++ b/docker_assets/ffmpeg_download.py
@@ -0,0 +1,71 @@
+"""
+ffmpeg link builder
+copied as into build step in Dockerfile
+"""
+
+import json
+import os
+import sys
+import tarfile
+import urllib.request
+from enum import Enum
+
+API_URL = "https://api.github.com/repos/yt-dlp/FFmpeg-Builds/releases/latest"
+BINARIES = ["ffmpeg", "ffprobe"]
+
+
+class PlatformFilter(Enum):
+    """options"""
+
+    ARM64 = "linuxarm64"
+    AMD64 = "linux64"
+
+
+def get_assets():
+    """get all available assets from latest build"""
+    with urllib.request.urlopen(API_URL) as f:
+        all_links = json.loads(f.read().decode("utf-8"))
+
+    return all_links
+
+
+def pick_url(all_links, platform):
+    """pick url for platform"""
+    filter_by = PlatformFilter[platform.split("/")[1].upper()].value
+    options = [i for i in all_links["assets"] if filter_by in i["name"]]
+    if not options:
+        raise ValueError(f"no valid asset found for filter {filter_by}")
+
+    url_pick = options[0]["browser_download_url"]
+
+    return url_pick
+
+
+def download_extract(url):
+    """download and extract binaries"""
+    print("download file")
+    filename, _ = urllib.request.urlretrieve(url)
+    print("extract file")
+    with tarfile.open(filename, "r:xz") as tar:
+        for member in tar.getmembers():
+            member.name = os.path.basename(member.name)
+            if member.name in BINARIES:
+                print(f"extract {member.name}")
+                tar.extract(member, member.name)
+
+
+def main():
+    """entry point"""
+    args = sys.argv
+    if len(args) == 1:
+        platform = "linux/amd64"
+    else:
+        platform = args[1]
+
+    all_links = get_assets()
+    url = pick_url(all_links, platform)
+    download_extract(url)
+
+
+if __name__ == "__main__":
+    main()