Carnet Video

This is a custom static site generator coded for the website: https://vincent.club1.fr/carnet-video/

It produce a movie clip notebook accessible from the Web from a folder containing videos file and JSON metadata. Management is done thanks to an interactive CLI (this get close to a TUI)

It goes allong with some scripts to capture the video clips from movies using MPV plugins. See CAPTURE.md, for more info about this part.

⚠️ limitation

Currently the HTML templates are hardcoded in french (home page is hardcoded too). It's done like this as I'm the only user of this tool. If you're interested in using it, please contact me at [email protected]. It may be motivating enough for me to find a way to include other users !

Setup

Installation process (for Debian or derivates)

Dependencies

This programm is supposed to depend only on packages that are in Debian. (that's a design strategy) It use composer to call a few PHP packages.

• packages
  • ffmpeg
  • mediainfo
  • imagemagick
  • jq
  • timg
  • php
  • composer
    • plates
• Wikidata (online service)

Command to install dependencies (on Debian)

sudo apt install ffmpeg mediainfo imagemagick jq timg php composer

Install

Go to desired folder. Then clone the git repo in current folder.

git clone https://github.com/vincent-peugnet/carnet-video/ .

Install composer dependencies

composer install

All scripts at the root of directory should have execution permission (including the PHP one).

chmod +x *.sh build-html.php

Create base directories and files for "src" and "config".

mkdir -p src/clips src/collections src/movies config
touch src/tags src/aspectRatios config/newClip config/importDirectory

Configure

• Fill config/newClip with number 1
• Fill config/importDirectory with the temporary path where new video files are captured. It's recommended to use an absolute path (starting with a /)
• If the site will be published inside subfolers, fill config/basePath with the name of the path that lead to the site. (⚠️ witout leading or trailing slashes)

Use

Architecture

📁 src
    📁 clips
        🎞️ <clipId>.mkv
        📄 <clipId>.json
        ...
    📁 collections
        📄 <collectionID>
        ...
    📁 movies
        📄 <movieId>.json
        ...
    📄 tags
    📄 aspectRatios

Movie ID used is the official Wikidata id. The src/movies folder is automatically populated during building.

Import new video clips

This programm assumes you have a folder that contain newly captured video file(see CAPTURE.md). This directory have to be declared in the file config/importDirectory. The ID of the future clip have to be a valid number in config/newClip. This number will be auto incremented during each import phase.

To import clip use the follow command:

./importClips.sh

Manage

Main command to view, edit a clip:

./display.sh <ClipID>

Will allow you to display all info about a clip. After what, pressing dedicated key will open editors for description, tags and collections.

This is equivalent to use those commands:

./clipDescription.sh <ClipID>
./clipTags.sh <ClipID>
./clipCollections.sh <ClipID>
./clipMovie.sh <ClipID>

To create/edit a collection, it's also possible to use

./collection.sh <collectionID> <ClipID>...

And to provide any number of clip to be added to the collection.

Build

The generator build from src dir to build dir. This folders are hardcoded.

To lauch the build:

./build.sh

This will first encode thumnails, previews (animated webp), and videos files. Then check for clip's movies. And finally generate HTML.

Infos

Encoding

Quality setting for video is hardcoded. in file build-thumbnail.sh Resolution output depend on aspect ratio. video with aspect ratio under 2 are encoded to 480p and above use 576p.