misc_scripts

Miscellaneous scripts to automate common tasks.

srt_to_transcript.py

Saves contents of the specified .srt files to a plain text transcripts.

Requires

• srt

transcribe_to_srt.py

Transcribes the specified media files such as .mkv to .srt subtitles. Defaults to model WhisperX and language English ("en") for transcription. The model should automatically download and install when the script is run.

Requires

• SubsAI (https://github.com/abdeladim-s/subsai)
  • Model: m-bain/whisperX

Recommended

Torch with CUDA support is highly recommended if you have a CUDA capable machine. For SubsAI with torch-2.0.1 requirement, install torch-2.0.1+cu118 per instruction https://pytorch.org/get-started/previous-versions/#v201 instead of default one in SubsAI "requirements.txt" file.

insanely-fast-whisper.py

Minimalistic script to generate transcription using Whisper.

Requires

• torch
• transformers

mp4-to-mp3-converter-with-origin.py

Converts mp4 files to mp3 files. I use this to easily convert my Udio songs to .mp3s for my iPhone.

The converted MP3-files include:

• the audio from the video file
• a thumbnail based on the first frame of the video
• the following metadata:
  • Title & Artist (based on the filename, "Artist - Title.mp4")
    • Defaults to "Udio" if nothing else is specified
  • Comments: Refferer and HostUrl based on the Windows 10/11 metadata stored with the file when downloaded

Requires

Windows 10 or Windows 11.

• moviepy
• eyed3

clipboard-monitor.py

Monitors the clipboard for changes and appends the contents to "clipboard.csv" file. It plays a sound when a change is detected and saved to the file.

Requires

Windows 10 or Windows 11.

• winsound
• win32clipboard

file-renamer-script.py

Takes a "clipboard.csv" file as input and uses the first column as a file_id that it tries to find in the files in the same folder as the script. If it finds the file, it is added to the list, with the proposed filename in the second column.

It then outputs the full list of proposed changes as file "rename_mappings.csv", so the user can verify the changes.

If the user approves them, the user simply responds "y", or "yes" or presses enter to confirm that they want to rename the files in the folder with the proposed filenames.

Requires

• csv
• logging
• traceback

m3u8-to-mp4-flask-webservice.py

A flask web service that takes a m3u8 file as input and converts it into an MP4 file.

The webservice exposes the following interfaces:

Interface	Methods	Functions	Parameters
stream_info	GET	stream_info	url
convert	POST & GET	convert_m3u8_to_mp4	url, alt_url, title, video_id, description

stream_info

Returns a JSON structure describing the overall metadata of the base stream, and the contained audio and video streams.

Parameter	Description
url	The URL of the m3u8 file to be queried.

convert

Converts an input m3u8 file into a MP4 file. The input is sent as a multipart/form-data request, with the key "file" and the value being the m3u8 file to be converted.

Parameter	Description
url	The URL of the m3u8 file to be converted.
alt_url	Can be used to provide the web page where the video was located for example.
title	The title of the video.
video_id	A unique ID that is used to identify the video.
description	A description of the video.

Requires

• uuid
• flask
• requests
• m3u8
• urllib
• ollama

m3u8-to-mp4-flask-webservice-simple.py

A flask web service that takes a m3u8 file as input and converts it into an MP4 file.

The webservice exposes the following interfaces:

Interface	Methods	Functions	Parameters
get	POST & GET	convert_m3u8_to_mp4	url, filename

get

Streams a m3u8 to save it as a mp4, real-time saving only, so will take as long as the stream itself is.

Parameter	Description
url	The URL of the m3u8 file to be converted.
filename	Target filename (optional).

Requires

• uuid
• flask
• requests
• m3u8
• urllib

merge-audio-files-to-one-output.py

Simple merge a bunch of audio files into one single output file. Just drag all the input files onto the script and it will be output in the same folder as the first file with the name "combined_output.<ext>". The script will ask what format, bitrate etc the output shall get.

Requires

• pydub
• inquirer
• tqdm

udio-flask-webservice.py (udio-download_ext-button.user.js)

A flask web service that adds metadata including cover art to your song files downloaded from Udio. It comes with a user script (e.g. Tampermonkey) that simplifies this process by adding a new button to the song pages "Download with metadata" that calls the webservice.

The webservice exposes the following interfaces:

Interface	Methods	Functions	Parameters
/api/download_ext	POST & GET	download_ext	mp3_url, image_url, title, artist, album, genre, year, cannonical, lyrics

/api/download_ext

Downloads the specified .mp3 file and adds the provided metadata to it.

Parameter	Tag	Description
mp3_url	Not used	The URL of the .mp3 file to be converted.
image_url	Images (Cover)	The URL of the cover art in .jpg format to use.
title	Title	The title of the track.
artist	Artist	Artist name(s) and/or alias(es).
album	Album	The title of the album.
genre	Genre	The genre of the track.
year	Year	Year of release.
cannonical	WWWAUDIOFILE	The source url of the track where it can be found permanently.
lyrics	UNSYNCEDLYRICS	Lyrics to add to the track.

Requires

• flask
• eyed3
• magic (python-magic-bin on windows)

video-optimizer.py

A script that allows for quick and eazy optimization of videos. Just supply a list of videos on the command line or drag and drop them onto the script. You get a list of choices based on the contents of the videos such as which subtitles to make default, and which audio to make default along with target quality and resolution.

It is made specifically for transcoding for example tv-shows from your legacy media in a quick and simple way. Jsut drag a whole season onto the script and easily convert it for use on your phone.

Requires

• ffmpeg (in the system path)
• ffmpeg-python
• inquirer
• tqdm

lyrics-timing-generator.py

A script that generates timed lyric for music files containing non-timed lyrics or no pre-prepared lyrics.

Requires

• pydub
• mutagen
• numpy
• librosa
• tensorflow
• spleeter
• soundfile
• whisper
• tqdm
• dataclasses
• typing

Experimental

mini-dlna-server.py

A script that runs a local dlna server instance on the computer, taking as input a command line argument pointing out the folder to serve to clients.

It is intended if I get time to do it, to stabilize it to properly handle conenctions, work better with Windows 11, and transcode media to the client using ffmpeg.

Note! Currently it is extremely unstable and mostly doesn't work. If anyone wants to refactor it and fix some of the remaining issues that would be cool. :)

Requires

• mutagen

lyrics-timing-generator.py

Intended to generate timed lyrics for audio files (.lrc). Uses whisper library to generate timed lyrics, and ollama and an llm to structure them.

But it is not good. Really not good. It's a start, but not quite there yet. Need to restart from a known base to generate the timed subtitles which is a known working thing, and then convert that to lyrics, using an llm to format them.

• ollama
• whisper
• mutagen
• numpy
• typing
• soundfile etc...