video.py

from dotenv import load_dotenv
import assemblyai as aai
from moviepy.video.fx.crop import crop
from termcolor import colored
from moviepy.editor import *
import moviepy.video.fx.all as vfx
import random


load_dotenv('.env')
aai.settings.api_key = os.getenv('ASSEMBLYAI_API_KEY')

FILE_URL = './speech/speech.mp3'


def create_video(title, i):
    words = transcribe()
    gameplay = cut_video(words[-1].end)
    gameplay = add_audio(gameplay)
    gameplay = subtitle(gameplay, words)
    gameplay = gameplay.fx(vfx.speedx, 1.25)
    gameplay.write_videofile(f'D:\\videos\\{i}{title}.mp4', codec='libx264', audio_codec='aac', bitrate="5000k")


def transcribe():
    transcriber = aai.Transcriber()
    transcript = transcriber.transcribe(FILE_URL)
    if transcript.status == aai.TranscriptStatus.error:
        print(colored(transcript.error))
    else:
        print(colored("transcribed", 'green'))
        return transcript.words


def cut_video(length: int):
    options = ["gta", "satisfying", "minecraft"]
    choice = options[random.randint(0, 2)]
    print(colored(choice, 'green'))
    gameplay = VideoFileClip(f"D:\\BACKGROUNDVIDEOS\\{choice}.mp4")
    gameplay = gameplay.without_audio()

    (w, h) = gameplay.size
    cropped_clip = crop(gameplay, width=600, height=5000, x_center=w / 2, y_center=h / 2)

    start_time = random.uniform(0, gameplay.duration-(length / 1000)-2)
    end_time = start_time + (length / 1000) + 2

    return cropped_clip.subclip(start_time, end_time)


def add_audio(gameplay: VideoFileClip):
    audio_clip = AudioFileClip(FILE_URL)
    new_audio_clip = CompositeAudioClip([audio_clip])
    gameplay.audio = new_audio_clip
    return gameplay


def subtitle(gameplay: VideoFileClip, words):
    clip_list = [gameplay]

    objects = three_per_line(words)

    for word_group in objects:

        text = ' '.join(word.text for word in word_group)
        print(colored(text, "blue"))

        duration = word_group[-1].end - word_group[0].start

        txt_clip = (TextClip(text, fontsize=50, color='white', font='TheBoldFont',
                             size=(gameplay.w, gameplay.h), method="caption",
                             stroke_width=2, stroke_color="black")
                    .set_position(('center', 'center'))
                    .set_duration(duration / 1000)
                    .set_start(t=word_group[0].start / 1000))

        clip_list.append(txt_clip)

    final_clip = CompositeVideoClip(clip_list)
    return final_clip


def three_per_line(words):
    for i in range(0, len(words), 3):
        yield words[i:i + 3]