countUCLetters.py

#!/usr/bin/env python

import re

IGNORE_BEGINNING = True

allUppercase = re.compile(r'[A-ZÄÜÖ][A-ZÄÜÖ]+')

matchLetters = re.compile(r'[a-zäüößA-ZÄÜÖ]')
matchUCLetters = re.compile(r'[A-ZÄÜÖ]')

# Source https://www.biblesupersearch.com/bible-downloads/
for name in ['kjv.txt', 'luther.txt']:
    letters = 0
    ucletters = 0

    with open(name) as f:
        for line in f:
            words = line.split(' ')
            if len(words) < 4:
                continue
            first_word = 2
            if ':' in words[first_word]:
                first_word += 1
            if words[first_word] == '¶':
                first_word += 1
            if IGNORE_BEGINNING:
                first_word += 1
            sentence = ' '.join(
                filter(
                    lambda word: allUppercase.match(word) is None,
                    words[first_word:]))

            letters += len(matchLetters.findall(sentence))
            ucletters += len(matchUCLetters.findall(sentence))

    book_name = name.split(".")[0]
    percentage = 100 * ucletters / letters
    print(f'{book_name}:\t{percentage:.1f}%\t({letters}/{ucletters})')

Added IGNORE_BEGINNING if you want to ignore short and empty lines that probably have more uppercase words and the beginning of a line that usually starts with a number and an uppercase word.

Results:
IGNORE_BEGINNING = False
kjv.txt
2087073
72887
3.4923071689394667
luther.txt
3150777
221274
7.022839128253126

IGNORE_BEGINNING = True
kjv.txt
1808897
50564
2.795294591123762
luther.txt
2684794
172202
6.4139744054851136

Changed the code to only count uppercase letters if they appear a the start of a word (don't count all CAP), using IGNORE_BEGINNING = True

kjv.txt
2087073
44015
2.108934378433337

luther.txt
3150777
170745
5.41913946940707

Updated cleaning of the lines again and downloaded another version (machine readable instead of plain text which has hard line breaks):

kjv:	1.9%	(3090749/57577)
luther:	5.2%	(2993877/156041)