A collection of translations and scripts I use in my typesetting workflow for Lord of the Mysteries content like chibi minis

For running on Windows, the applications may need to have the paths to their executables added to the system's PATH variable

Python is required to run the scripts, and the required packages can be installed using your system's package manager

# Arch Linux
$: sudo pacman -S python

Video downloading uses yt-dlp

# Arch Linux
$: sudo pacman -S yt-dlp

Subtitle timings are generated using videosubfinder

# Arch Linux, using an AUR helper like yay or paru
$: yay -S videosubfinder

OCRed subtitle files are generated using RapidVideOCR

# Requires Python to be installed
$: pip install rapid_videocr

Subtitles are embedded using ffmpeg

# Arch Linux
$: sudo pacman -S ffmpeg

For each script, they optionally take in arguments for an entry to run on

Entries are defined in urls.json as a string key

{
  "example_entry": "https://www.url-to-some-video-here.com"
}

# Downloading the example_entry
$: ./download.py example_entry

If no arguments are given, the script will batch run on every entry in the urls.json

Files and output are placed in a directory with the respective entry name

Runs the main pipeline for an entry; first downloads the video file, generates the subtitle timings using the video file, creates the OCRed subtitle file, and finally applies the custom style

Requires dimensions.json to be set, since the ocr.py script uses its values

Uses yt-dlp to download a video file to the entry directory, generated video files are given the entry name as the filename and with the .mp4 file format

Uses videosubfinder to generate subtitle timings to the entry directory, which are saved to the RGBImages directory as .jpeg files whose timings are encoded in the filename

Uses RapidVideOCR to create an OCRed subtitle file to the entry directory using the RGBImages directory in the entry directory, and saves the subtitle file with the entry name as the filename and with the Advanced Substation Alpha .ass file format

Requires dimensions.json to set the subtitle crop boundary edges for the OCR search area

{
  "te": 0.20,
  "be": 0.14,
  "le": 0.25,
  "re": 0.75
}

te is the top edge, be is the bottom edge, le is the left edge, and re is the bottom edge; the values are defined as a float where 1.0 is the top-most edge of the image and 0.0 is the bottom-most edge of the image for the y-axis, and 1.0 is the right-most edge of the image and 0.0 is the left-most edge of the image for the x-axis

Adds and applies custom style defined in style.ass to all dialogue lines in the existing subtitle file with the entry name in the entry directory

Uses ffmpeg to encode a video file with the subtitles embedded (hardsubbed) in, and saves the hardsubbed video file in the entry directory with the filename [entry_name]_subbed.mp4