%titlecase_proper{} - a better %title{} template function with configurability

[trapd00r]

So it all started here.
I wasn't happy with the way the %title{} template function worked and tried various things until I came up with a simple plugin that simply used the titlecase library.

It works great with titlecasing everything that should be titlecased (New York Times Manual of Style) while leaving short words alone (a, and, for...).

However, I still wasn't completely satisfied though; at times there are words that should neither be titlecased, uppercased or lowercased but rather have a specific casing: for example the acronym 'PhD'. So I whipped up some more code and now the plugin can read ~/.beets_titlecase, where words are defined one per line, and their case will never be changed. Let's check it out:

$ cd "Dj Tcp the Reckless │2021│ So Much Drama in the Phd [CDR, MP3]"                                                                                        
$ for w in TCP DJ PhD; do echo $w; done >> ~/.beets_titlecase                                                                                                
$ beet import .                                                                                                                                              
$ beet ls -a 'so much drama' -f \$path                                                                                                                       
DJ TCP the Reckless │2021│ So Much Drama in the PhD [CDR, MP3]                                                                                               

The plugin is available here.
I'm still new to the world of python, so on the todo list is to replace the configuration file with ability to read a list of words to not be touched from config.yaml: https://beets.readthedocs.io/en/stable/dev/plugins.html#read-configuration-options

[sampsyo]

Super awesome! Let us know if you need help figuring out the configuration system to realize that last to-do item.

[trapd00r]

5 hours, 12 minutes and 11 seconds spent on titlecase_proper.py

... And all I got to show for all that time is those few lines of code below. Well, this took waay longer than expected and at one point I was on my way to a 100 sloc before I realized I could simplify things a lot (and then 12 other things broke in the process, repeat while true!). I'm having more fun than I expected learning python this way though.

beets/config.yaml:

titlecase_proper:
  asis: [
          'PhD',
          'DJ',
          'TCP',
          'YodA',
  ]

Renders:

• DJ Madfist the Invincible TCP Guru │2024│ Got Myself an PhD in the 313 (Deluxe YodA Edition) [MP3]

It seems to work like it should, but I am unsure about the error handling - is there perhaps a better way to do it in case config.yaml is missing the relevant plugin section?
Other thoughts or input? Much appreciated!

from beets         import config
from beets.plugins import BeetsPlugin
from titlecase     import titlecase
import os

# error handling, in case there's no config section for
# titlecase_proper, we should just titlecase as normal
words_to_use_asis = None
try:
  words_to_use_asis = config['titlecase_proper']['asis'].get()
except:
  words_to_use_asis = None

class titlecase_proper(BeetsPlugin):
    def __init__(self):
      super(titlecase_proper, self).__init__()
      self.template_funcs['titlecase_proper'] = _titlecase_proper


def custom_casing(word, **kwargs):
    for w in words_to_use_asis:
      # compare words uppercased, since the casing can be whatever in
      # the config.
        if word.upper() == w.upper():
            return w


def _titlecase_proper(text):
    if words_to_use_asis is not None:
        real_words = titlecase(text, callback=custom_casing)
    else:
        real_words = titlecase(text)

    return ''.join(real_words)