File size: 4,172 Bytes
eff85b8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
!pip install -U openai-whisper
!pip install -U git+https://github.com/linto-ai/whisper-timestamped
!pip install gradio moviepy whisper-timestamped

import os
import datetime as dt
import json
import whisper_timestamped as whisper
from moviepy.video.io.VideoFileClip import VideoFileClip
import gradio as gr

# Helper functions and global variables
outdir = dt.datetime.now().strftime("%Y%m%d%H%M")
if os.path.exists(outdir):
    random_digits = str(random.randint(1000, 9999))
    new_outdir = outdir + random_digits
    os.mkdir(new_outdir)
    outdir = new_outdir
    print("Created new output directory:", new_outdir)
else:
    os.system(f"mkdir {outdir}")
print("date time now:" + outdir)

model = whisper.load_model("base")

def generate_timestamps(vidname):
    audio = whisper.load_audio(vidname)
    result = whisper.transcribe(model, audio, language="en")
    return result

def get_segment_info(data):
    new_list = []
    for segment in data.get("segments", []):
        if "id" in segment and "start" in segment and "end" in segment and "text" in segment:
            new_item = {
                "id": segment["id"],
                "start": segment["start"],
                "end": segment["end"],
                "text": segment["text"]
            }
            new_list.append(new_item)
    return new_list

def combine_entries(entries):
    combined_entries = []
    current_entry = None
    total_duration = 0

    for entry in entries:
        entry_duration = entry["end"] - entry["start"]

        if total_duration + entry_duration > 30:
            if current_entry:
                current_entry["end"] = entry["end"]
                combined_entries.append(current_entry)

            current_entry = {
                "start": entry["start"],
                "end": entry["end"],
                "text": entry["text"]
            }
            total_duration = entry_duration
        else:
            if current_entry:
                current_entry["end"] = entry["end"]
                current_entry["text"] += " " + entry["text"]
                total_duration += entry_duration
            else:
                current_entry = {
                    "start": entry["start"],
                    "end": entry["end"],
                    "text": entry["text"]
                }
                total_duration = entry_duration

    if current_entry:
        combined_entries.append(current_entry)

    return combined_entries

def extract_video_segment(input_video, output_video, start_time, end_time):
    video_clip = VideoFileClip(input_video).subclip(start_time, end_time)
    video_clip.write_videofile(output_video, codec="libx264", audio_codec="aac")
    video_clip.close()

def save_segments(outdir, name, combined_entries):
    segments = combined_entries
    input_video = name
    for i, segment in enumerate(segments):
        start_time = segment['start']
        end_time = segment['end']
        output_video_file = f'{outdir}/output_segment_{i + 1}.mp4'
        extract_video_segment(input_video, output_video_file, start_time, end_time)

def split_up_video(video_path, output_dir):
    result = generate_timestamps(video_path)
    combined_entries = combine_entries(get_segment_info(result))

    scribeout = open(f"{output_dir}/transcript.txt", "w")
    scribeout.write(json.dumps(combined_entries, indent=2, ensure_ascii=False))
    scribeout.close()

    save_segments(output_dir, video_path, combined_entries)

    filename, extension = os.path.splitext(video_path)
    os.system(f"zip -r {filename}.zip {output_dir}")

    return f"{filename}.zip"

# Gradio interface
def process_video(video):
    output_dir = dt.datetime.now().strftime("%Y%m%d%H%M")
    os.mkdir(output_dir)
    video_path = video.name
    output_zip = split_up_video(video_path, output_dir)
    return output_zip

iface = gr.Interface(
    fn=process_video,
    inputs=gr.File(file_count="single", type="filepath", label="Upload a Video"),
    outputs=gr.File(label="Download Zipped Segments and Transcript"),
    title="Video Splitter",
    description="Upload a video and get a zipped file with segmented videos and a transcript."
)

iface.launch()