Create Subtitles with Speaker Labels

1
import assemblyai as aai
2
3
# SETTINGS
4
aai.settings.api_key = "YOUR-API-KEY"
5
filename = "YOUR-FILE-NAME"
6
transcriber = aai.Transcriber(config=aai.TranscriptionConfig(speaker_labels=True))
7
transcript = transcriber.transcribe(filename)
8
9
# Maximum number of words per subtitle
10
max_words_per_subtitle = 6
11
12
# Color assignments for speakers
13
speaker_colors = {
14
    "A": "red",
15
    "B": "orange",
16
    "C": "yellow",
17
    "D": "yellowgreen",
18
    "E": "green",
19
    "F": "lightskyblue",
20
    "G": "purple",
21
    "H": "mediumpurple",
22
    "I": "pink",
23
    "J": "brown",
24
}
25
26
# Process transcription segments
27
def process_segments(segments):
28
    srt_content = ""
29
    subtitle_index = 1
30
    for segment in segments:
31
        speaker = segment.speaker
32
        color = speaker_colors.get(speaker, "black")  # Default color is black
33
34
        # Split text into words and group into chunks
35
        words = segment.words
36
        for i in range(0, len(words), max_words_per_subtitle):
37
            chunk = words[i:i + max_words_per_subtitle]
38
            start_time = chunk[0].start  # -1 indicates continuation
39
            end_time = chunk[-1].end
40
            srt_content += create_subtitle(subtitle_index, start_time, end_time, chunk, color)
41
            subtitle_index += 1
42
43
    return srt_content
44
45
46
# Create a single subtitle
47
def create_subtitle(index, start_time, end_time, words, color):
48
    text = ""
49
    for word in words:
50
        text += word.text + ' '
51
    start_srt = format_time(start_time)
52
    end_srt = format_time(end_time)
53
    return f"{index}\n{start_srt} --> {end_srt}\n<font color=\"{color}\">{text}</font>\n\n"
54
55
# Format time in SRT style
56
def format_time(milliseconds):
57
    hours, remainder = divmod(milliseconds, 3600000)
58
    minutes, remainder = divmod(remainder, 60000)
59
    seconds, milliseconds = divmod(remainder, 1000)
60
    return f"{int(hours):02}:{int(minutes):02}:{int(seconds):02},{int(milliseconds):03}"
61
62
# Generate SRT content
63
sentences = transcript.get_sentences()
64
srt_content = process_segments(sentences)
65
66
# Save to SRT file
67
with open(filename + '.srt', 'w') as file:
68
    file.write(srt_content)
69
70
print(f"SRT file generated: {filename}.srt")

1
import assemblyai as aai
2
3
# SETTINGS
4
aai.settings.api_key = "YOUR-API-KEY"
5
filename = "YOUR-FILE-NAME"
6
transcriber = aai.Transcriber(config=aai.TranscriptionConfig(speaker_labels=True))
7
transcript = transcriber.transcribe(filename)
8
9
# Maximum number of words per subtitle
10
max_words_per_subtitle = 6

1
# Color assignments for speakers
2
speaker_colors = {
3
    "A": "red",
4
    "B": "orange",
5
    "C": "yellow",
6
    "D": "yellowgreen",
7
    "E": "green",
8
    "F": "lightskyblue",
9
    "G": "purple",
10
    "H": "mediumpurple",
11
    "I": "pink",
12
    "J": "brown",
13
}
14
15
# Process transcription segments
16
def process_segments(segments):
17
    srt_content = ""
18
    subtitle_index = 1
19
    for segment in segments:
20
        speaker = segment.speaker
21
        color = speaker_colors.get(speaker, "black")  # Default color is black
22
23
        # Split text into words and group into chunks
24
        words = segment.words
25
        for i in range(0, len(words), max_words_per_subtitle):
26
            chunk = words[i:i + max_words_per_subtitle]
27
            start_time = chunk[0].start  # -1 indicates continuation
28
            end_time = chunk[-1].end
29
            srt_content += create_subtitle(subtitle_index, start_time, end_time, chunk, color)
30
            subtitle_index += 1
31
32
    return srt_content
33
34
# Create a single subtitle
35
def create_subtitle(index, start_time, end_time, words, color):
36
    text = ""
37
    for word in words:
38
        text += word.text + ' '
39
    start_srt = format_time(start_time)
40
    end_srt = format_time(end_time)
41
    return f"{index}\n{start_srt} --> {end_srt}\n<font color=\"{color}\">{text}</font>\n\n"
42
43
# Format time in SRT style
44
def format_time(milliseconds):
45
    hours, remainder = divmod(milliseconds, 3600000)
46
    minutes, remainder = divmod(remainder, 60000)
47
    seconds, milliseconds = divmod(remainder, 1000)
48
    return f"{int(hours):02}:{int(minutes):02}:{int(seconds):02},{int(milliseconds):03}"

1
# Generate SRT content
2
sentences = transcript.get_sentences()
3
srt_content = process_segments(sentences)
4
5
# Save to SRT file
6
with open(filename + '.srt', 'w') as file:
7
    file.write(srt_content)
8
9
print(f"SRT file generated: {filename}.srt")