Esse é o arquivo legenda.srt: ``` 00:00,770#oh 00...

python
original = """00:00,770#oh
00:01,270#oh
00:01,770#oh
00:02,270#oh
00:02,770#oh
00:05,000#ooh
00:05,700#ooh
00:09,581#i
00:09,831#feel
00:10,361#so
00:11,271#unsure
00:15,401#as
00:15,621#i
00:15,771#take
00:16,321#your
00:16,581#hand
00:17,281#and
00:17,621#lead
00:18,021#you
00:18,321#to
00:19,421#the
00:20,001#dance
00:20,871#floor
00:22,101#as
00:22,381#the
00:22,511#music
00:23,131#dies
00:25,491#something
00:25,971#in
00:26,191#your
00:26,491#eyes
00:28,101#calls
00:28,501#to
00:28,701#mind
00:29,291#a
00:29,501#silver
00:30,231#screen
00:31,291#and
00:31,581#all
00:32,181#its
00:32,681#sad
00:33,431#goodbyes
00:34,981#i'm
00:35,281#never
00:35,581#gonna
00:35,981#dance
00:36,381#again
00:37,531#guilty
00:38,131#feet
00:38,431#have
00:38,631#got
00:39,181#no
00:39,831#rhythm
00:40,931#though
00:41,231#it's
00:41,431#easy
00:42,131#to
00:42,531#pretend
00:43,631#i
00:43,781#know
00:44,081#you're
00:44,381#not
00:44,781#a
00:44,981#fool
00:46,781#should've
00:47,181#known
00:47,481#better
00:47,931#than
00:48,131#to
00:48,281#cheat
00:48,731#a
00:48,931#friend
00:49,831#and
00:50,031#waste
00:50,481#the
00:50,731#chance
00:51,231#that
00:51,431#i've
00:51,831#been
00:52,431#given
00:53,631#so
00:53,831#i'm
00:53,931#never
00:54,281#gonna
00:54,681#dance
00:55,181#again
00:56,231#the
00:56,431#way
00:56,731#i
00:56,931#danced
00:57,381#with
00:58,031#you
00:59,031#ooh
01:00,031#ooh
01:02,300#oh
01:02,800#ho
01:07,131#time
01:07,831#can
01:08,131#never
01:09,131#mend
01:13,231#the
01:13,531#careless
01:14,531#whispers
01:16,331#of
01:16,931#a
01:17,731#good
01:18,431#friend
01:19,931#to
01:20,231#the
01:20,431#heart
01:20,831#and
01:21,031#mind
01:23,131#ignorance
01:23,931#is
01:24,131#kind
01:25,731#there's
01:26,231#no
01:26,531#comfort
01:27,331#in
01:27,631#the
01:27,831#truth
01:29,131#pain
01:29,831#is
01:30,231#all
01:30,831#you'll
01:31,831#find
01:32,981#i'm
01:33,281#never
01:33,581#gonna
01:33,981#dance
01:34,381#again
01:35,231#guilty
01:35,831#feet
01:36,131#have
01:36,331#got
01:36,931#no
01:37,531#rhythm
01:38,631#though
01:38,931#it's
01:39,131#easy
01:39,831#to
01:40,231#pretend
01:41,331#i
01:41,431#know
01:41,731#you're
01:42,031#not
01:42,431#a
01:42,631#fool
01:44,431#should've
01:44,831#known
01:45,131#better
01:45,631#than
01:45,831#to
01:45,931#cheat
01:46,431#a
01:46,631#friend
01:47,531#and
01:47,731#waste
01:48,031#this
01:48,331#chance
01:48,831#that
01:49,031#i've
01:49,431#been
01:50,031#given
01:51,231#so
01:51,431#i'm
01:51,531#never
01:51,831#gonna
01:52,231#dance
01:52,731#again
01:53,931#the
01:54,131#way
01:54,431#i
01:54,631#danced
01:55,031#with
01:55,431#you
01:56,331#oh
02:01,131#never
02:01,731#without
02:02,531#your
02:02,931#love
02:06,131#tonight
02:06,731#the
02:06,931#music
02:07,531#seems
02:08,031#so
02:08,331#loud
02:09,231#i
02:09,431#wish
02:09,731#that
02:10,131#we
02:10,431#could
02:10,731#lose
02:11,331#this
02:11,731#crowd
02:12,631#maybe
02:13,431#it's
02:13,931#better
02:14,531#this
02:14,931#way
02:15,631#we'd
02:15,831#hurt
02:16,131#each
02:16,331#other
02:16,631#with
02:16,831#the
02:17,031#things
02:17,231#we'd
02:17,431#want
02:17,731#to
02:17,931#say
""".strip().splitlines()

len_orig = len(original)
len_orig

211

python
lyrics = """Oh, oh, oh
Oh

I feel so unsure
As I take your hand
And lead you to the dance floor
As the music dies
Something in your eyes
Calls to mind a silver screen
And all its sad goodbyes

I'm never gonna dance again
Guilty feet have got no rhythm
Though it's easy to pretend
I know you're not a fool

Should've known better
Than to cheat a friend
And waste the chance
That I've been given
So I'm never gonna dance again
The way I danced with you, oh, oh

Time can never mend
The careless whispers
Of a good friend
To the heart and mind
Ignorance is kind
There's no comfort in the truth
Pain is all you'll find

I'm never gonna dance again
Guilty feet have got no rhythm
Though it's easy to pretend
I know you're not a fool

Should've known better
Than to cheat a friend
And waste this chance
That I've been given
So I'm never gonna dance again
The way I danced with you, oh

Never without your love

Tonight the music seems so loud
I wish that we could lose this crowd
Maybe it's better this way
We'd hurt each other
With the things we'd want to say
""".strip()

# tokenize lyrics removing punctuation except apostrophes in contractions and preserving words
import re, unicodedata, string, itertools, textwrap, collections, json

def normalize_word(w):
    # Remove punctuation at ends , . ? ! ; : etc
    w = w.strip()
    if not w:
        return ''
    # Keep apostrophes and contractions
    # Remove any leading/trailing punctuation excluding apostrophes
    w = re.sub(r'^[^\w\']+|[^\w\']+$', '', w)
    return w

tokens=[]
for line in lyrics.splitlines():
    for word in line.split():
        t = normalize_word(word)
        if t:
            tokens.append(t)

len_tokens = len(tokens)
len_tokens

206

python
tokens[:50]

['Oh',
 'oh',
 'oh',
 'Oh',
 'I',
 'feel',
 'so',
 'unsure',
 'As',
 'I',
 'take',
 'your',
 'hand',
 'And',
 'lead',
 'you',
 'to',
 'the',
 'dance',
 'floor',
 'As',
 'the',
 'music',
 'dies',
 'Something',
 'in',
 'your',
 'eyes',
 'Calls',
 'to',
 'mind',
 'a',
 'silver',
 'screen',
 'And',
 'all',
 'its',
 'sad',
 'goodbyes',
 "I'm",
 'never',
 'gonna',
 'dance',
 'again',
 'Guilty',
 'feet',
 'have',
 'got',
 'no',
 'rhythm']

python
orig_words = [line.split('#')[1] for line in original]
len_orig, len_orig == len(orig_words)

(211, True)

python
[o for o in tokens if o.lower()=='ooh']

[]

python
updated_lines=[]
j=0
for i,line in enumerate(original):
    time, word = line.split('#')
    lw = word.lower()
    replacement = None
    while j < len_tokens:
        tk = tokens[j]
        j += 1
        if lw == tk.lower():
            replacement = tk
            break
    if replacement is None:
        # keep capitalizing first letter if it's alphabetic and not contraction
        if word.lower() == 'i' or word.lower() == "i'm":
            replacement = word.capitalize()
        elif "'" in word:
            # For contractions, capitalize first letter if 'i' or at sentence start maybe; fallback
            replacement = word  # will handle later if need
            if word.startswith("i'" ):
                replacement = "I" + word[1:]
        elif word:
            replacement = word.capitalize()
        else: replacement = word
    updated_lines.append(f"{time}#{replacement}")

len(updated_lines)


211

python
"\n".join(updated_lines[:50])

"00:00,770#Oh\n00:01,270#oh\n00:01,770#oh\n00:02,270#Oh\n00:02,770#oh\n00:05,000#Ooh\n00:05,700#Ooh\n00:09,581#I\n00:09,831#Feel\n00:10,361#So\n00:11,271#Unsure\n00:15,401#As\n00:15,621#I\n00:15,771#Take\n00:16,321#Your\n00:16,581#Hand\n00:17,281#And\n00:17,621#Lead\n00:18,021#You\n00:18,321#To\n00:19,421#The\n00:20,001#Dance\n00:20,871#Floor\n00:22,101#As\n00:22,381#The\n00:22,511#Music\n00:23,131#Dies\n00:25,491#Something\n00:25,971#In\n00:26,191#Your\n00:26,491#Eyes\n00:28,101#Calls\n00:28,501#To\n00:28,701#Mind\n00:29,291#A\n00:29,501#Silver\n00:30,231#Screen\n00:31,291#And\n00:31,581#All\n00:32,181#Its\n00:32,681#Sad\n00:33,431#Goodbyes\n00:34,981#I'm\n00:35,281#Never\n00:35,581#Gonna\n00:35,981#Dance\n00:36,381#Again\n00:37,531#Guilty\n00:38,131#Feet\n00:38,431#Have"

python
for line in updated_lines[:20]:
    print(line)

python
tokens[:15]

['Oh',
 'oh',
 'oh',
 'Oh',
 'I',
 'feel',
 'so',
 'unsure',
 'As',
 'I',
 'take',
 'your',
 'hand',
 'And',
 'lead']

python
def transform(original_lines, tokens):
    updated=[]
    j=0
    n=len(tokens)
    for line in original_lines:
        time, word = line.split('#')
        lw=word.lower()
        replacement=None
        # look ahead up to 15 tokens
        for k in range(j, min(j+15, n)):
            if tokens[k].lower()==lw:
                replacement=tokens[k]
                j=k+1
                break
        if replacement is None:
            # if doesn't match, keep word capitalized maybe per rules
            if lw=="i":
                replacement="I"
            elif lw=="i'm":
                replacement="I'm"
            elif "'" in word:
                # preserve apostrophe case as is but maybe capital if it starts with i?
                if word.startswith("i'"):
                    replacement="I"+word[1:]
                else:
                    replacement=word
            elif word:
                # default keep original as is (lower-case)
                replacement=word.capitalize()  # maybe capitalizing first letter but may not be correct
            else:
                replacement=word
        updated.append(f"{time}#{replacement}")
    return updated

updated2=transform(original, tokens)
updated2[:30]


['00:00,770#Oh',
 '00:01,270#oh',
 '00:01,770#oh',
 '00:02,270#Oh',
 '00:02,770#Oh',
 '00:05,000#Ooh',
 '00:05,700#Ooh',
 '00:09,581#I',
 '00:09,831#feel',
 '00:10,361#so',
 '00:11,271#unsure',
 '00:15,401#As',
 '00:15,621#I',
 '00:15,771#take',
 '00:16,321#your',
 '00:16,581#hand',
 '00:17,281#And',
 '00:17,621#lead',
 '00:18,021#you',
 '00:18,321#to',
 '00:19,421#the',
 '00:20,001#dance',
 '00:20,871#floor',
 '00:22,101#As',
 '00:22,381#the',
 '00:22,511#music',
 '00:23,131#dies',
 '00:25,491#Something',
 '00:25,971#in',
 '00:26,191#your']

python
# find some lines
[l for l in updated2 if "#i'm" in l.lower()][:10]


["00:34,981#I'm", "00:53,831#I'm", "01:32,981#I'm", "01:51,431#I'm"]

python
[l for l in updated2 if "#you're" in l.lower()][:10]


["00:44,081#you're", "01:41,731#you're"]

python
[l for l in updated2 if "should've" in l.lower()][:5]


["00:46,781#Should've", "01:44,431#Should've"]

python
for line in updated2[-20:]:
    print(line)

python
updated_text = "\n".join(updated2)
len(updated_text.splitlines())

211