lortordermur / wapamux-tools

Script toolchain for mass naming and remuxing of anime and TV episodes

wapamux-tools – Script toolchain for mass naming and remuxing of anime and TV episodes

Summary

wapamux-tools provides a set of handy shell scripts for making the mass naming and tagging of anime and TV episodes a lot easier.

• wapaname: A semi-automated frontend for tvnamer.
• wapamux: Converts or remuxes (based on an mkvmerge option file) video files into a subdirectory.
• wapasplit: Splits a set of video files into subdirectories by metadata CRC.

Output is always in Matroska (.mkv) format.

Note that the scripts are deliberately intended for per-directory semi-automation only, without a true batch mode. The motivation behind this is the sensible handling of anime seasons put together from different encodes.

Prerequisites

wapamux-tools should be able to run fine on most POSIX-compatible shells, like bash, zsh and dash, as long as the dependencies are met. These are:

• tvnamer[1]
• mkvmerge (part of the mkvtoolnix[2] package)
• rsync[3]—preinstalled on just about anything non-embedded

wapamux-tools has only been tested on Linux so far; BSD/OSX/Android testers are highly welcome.

Installation

First install the dependencies, if required. Then copy the script files to somewhere in your path (like ~/bin) and chmod +x if necessary. Symlinking them will work too.

Usage

For more detailed information, run the the scripts with the -h or --help switch.

wapaname [series_name] [tvdb_series_id]

Auto-renames all TV episodes in the current directory using tvnamer. File names have to already be in a Kodi-compatible sNNeNN, NNxNN or equivalent format; just the season and episode number is already sufficient. Should autodetection by means of the directory name fail, a series name can be enforced. Should that fail too, a TVDB ID can additionally be given.

wapamux [-s | --scan-headers] [fileext]

Remuxes all files ending in .fileext found in the current directory into a single subdirectory of a hardcoded name. If an mkvmerge option file named muxopt.json is present in the directory, it will be used for metadata tagging. fileext defaults to mkv. -s only does a scan for metadata consistency, just as it would before a remux with an option file.

wapasplit [-s | --scanonly] [fileext]

Splits all video files ending in .fileext found in the current directory into subdirectories based on their metadata checksum. This is handy with source files from different encodes. Files can either be rsync-copied (more secure, plus resumeable) or moved. fileext defaults to mkv. -s only performs the metadata CRCs and displays which directories would be created.

Example workflow

In a common scenario wapaname would be called after naming the video files in a way that tvnamer can parse (one series/season per directory, which is named after the series, sNNeNN or NNxNN file numbering). The result would be numbered video files with episode names.

Next a metadata consistency check can be done using wapasplit -s or wapamux -s. If the metadata turns out inconsistent across files, one can either manually perform multiple wapamux passes or let wapasplit take care of creating a directory subtree and copying/moving the files there.

Finally one would edit the header metadata as needed for each subset (usually by means of mkvtoolnix-gui), export the changes as option files, then remux using wapamux. There is nothing to worry about filenames stored in the option file(s), by the way—wapamux will substitute them with the currently processed file during remuxing.

FAQ

“wapamux” (「ワパムケス」)?

It is an allusion to “wapanese”, a somewhat derogatory term for a japanophile person with limited knowledge and a warped notion of Japanese culture and language. Just in line with that, I was utterly unable to come up with a more creative pun.

When remuxing, can I also move tracks around or remove some of them in addition to setting header data?

wapamux uses mkvmerge as its backend so this is perfectly fine. The option file (muxopt.json) that keeps track of the changes is not touched except for filename sanitation required for batch mode.

wapaname errors out with “tvnamer: error: No valid files were supplied”

Either you tried running the script on a movie, or tvnamer is unable to match the episode numbers in the file names. This can sometimes happen with unicode characters around or next to episode numbers, for example if a (typographically correct) dash is used in place of a minus sign. Filename matching issues should be reported directly to tvnamer's issue tracker[4].

wapaname defaults to English episode names when I force a series name or ID

tvnamer behaves like this when no language is set in its configuration. You can create a default configuration file using

mkdir -p ~/.config/tvnamer && tvnamer -s ~/.config/tvnamer/tvnamerrc

wapaname will always use that file as a tvnamer configuration if it exists. Minimal example file:

{
    "language": "de", 
    "search_all_languages": true, 
    "order": "dvd" 
}

Here, German is defined as the default language and episode order is set to DVD order instead of the default “first aired” order.

wapaname automatically selects the English dub of a series even when the language code is set in the configuration file

Try passing the series name and ID directly to wapaname, as in:

wapaname "Some anime series" 123456

This should make tvnamer auto-select the correct dub in most cases.

My desktop tends to lag a lot while wapamux processes very large files

Since 0.3.5, wapamux renices itself to 19 (lowest process priority) by default, as its primary purpose is processing large file batches in the background. Depending on how your operating system's kernel was configured, you may still have a responsive user interface, or less so. If you run Linux and get stalls while processing batches of very large video files, you might try trading some throughput for interactivity by temporarily switching to the deadline I/O scheduler.

First note down the currently set scheduler. Replace sda with the disk you are running the batch on.

cat /sys/block/sda/queue/scheduler

You will get something like

noop deadline [cfq]

This means you are running the CFQ scheduler which is very likely also the default. Switch to the deadline scheduler using

sudo sh -c "echo deadline > /sys/block/sda/queue/scheduler"

You may want to double-check using the aforementioned command:

noop [deadline] cfq

Yup. Note that the block driver may have to finish any outstanding requests first before you see an improvement.

wapaname crashes when the directory name contains Unicode characters

This seems to be an issue with tvnamer that may or may not pertain to certain system configurations. You can try enforcing an ASCII-only representation of the series title by passing it as the first parameter to wapaname. You may get different or no matches though, or in the worst case every dub except the one you want if the Unicode title is the original title and the ASCII string is a redirect.