File size: 6,060 Bytes
83aa67f
 
 
 
3c4bc36
83aa67f
 
 
d3f6dad
83aa67f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3c4bc36
 
 
83aa67f
 
 
3c4bc36
83aa67f
 
3c4bc36
 
 
 
 
 
 
 
83aa67f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
import os
import gradio as gr
from cerebras.cloud.sdk import Cerebras
from gtts import gTTS
from moviepy import VideoFileClip, concatenate_videoclips, AudioFileClip
import requests

# Initialize Cerebras client
Cerekey = os.getenv("CKey")
client = Cerebras(api_key= Cerekey)

# Pexels API key
pexkey = os.getenv("Pkey")
PEXELS_API_KEY = pexkey

# Modify the system prompt to include the estimated word count based on video duration
def generate_script(prompt, max_duration):
    system_message = f"You are an expert video content creator and narration writer who is proficient in generating narration from user prompts and crafting a concise and poetic narration that aligns with the prompt. Craft a concise, poetic narration for the prompt. Go straight to the narration, don't write a foreward or a description of your action. The narration should be suitable for a video that can be read in less than {max_duration} seconds."

    stream = client.chat.completions.create(
        messages=[{"role": "system", "content": system_message}, {"role": "user", "content": prompt}],
        model="llama-3.3-70b",
        stream=False,
        max_completion_tokens=1024,
        temperature=0.7,
        top_p=1
    )
    return stream.choices[0].message.content


def search_and_download_videos(query, max_duration, aspect_ratio, download_folder, max_results=6):
    url = "https://api.pexels.com/videos/search"
    headers = {"Authorization": PEXELS_API_KEY}
    params = {"query": query, "per_page": max_results}

    try:
        response = requests.get(url, headers=headers, params=params)
        response.raise_for_status()
        videos = response.json().get("videos", [])

        if not os.path.exists(download_folder):
            os.makedirs(download_folder)

        downloaded_files = []
        for video in videos:
            duration = video.get("duration")
            width = video.get("width")
            height = video.get("height")
            if width and height:
                video_aspect_ratio = "landscape" if width > height else "portrait" if height > width else "square"
                if duration <= max_duration and video_aspect_ratio == aspect_ratio:
                    video_url = video["video_files"][0]["link"]
                    video_id = video["id"]
                    video_filename = os.path.join(download_folder, f"{video_id}.mp4")
                    video_response = requests.get(video_url, stream=True)
                    with open(video_filename, "wb") as file:
                        for chunk in video_response.iter_content(chunk_size=1024):
                            file.write(chunk)

                    downloaded_files.append(video_filename)
        return downloaded_files
    except requests.exceptions.RequestException as e:
        print(f"Error: {e}")
        return []


def generate_narration(script, output_file="narration.mp3"):
    tts = gTTS(script, lang="en")
    tts.save(output_file)
    return output_file


def load_videos_from_folder(folder_path):
    if not os.path.exists(folder_path):
        print(f"Error: The folder '{folder_path}' does not exist.")
        return []

    video_files = [
        os.path.join(folder_path, file)
        for file in os.listdir(folder_path)
        if file.endswith(('.mp4', '.mov', '.avi', '.mkv'))
    ]
    return video_files


def aggregate_videos(clips):
    if not clips:
        return None
    return concatenate_videoclips(clips, method="compose")


def trim_video_to_audio_length(final_video, audio_length):
    if final_video.duration > audio_length:
        # Use subclipped method for CompositeVideoClip
        final_video = final_video.subclipped(0, audio_length)
    return final_video



# Function to add narration to the final video
def add_narration_to_video(final_video, narration_path):
    if os.path.exists(narration_path):
        narration_audio = AudioFileClip(narration_path)
        narration_audio = narration_audio.with_duration(final_video.duration)  # Adjust duration to match video
        final_video = final_video.with_audio(narration_audio)  # Use with_audio instead of set_audio
    return final_video



def save_final_video(final_video, output_path):
    final_video.write_videofile(output_path, codec="libx264", audio_codec="aac", preset="ultrafast")


def generate_video(prompt, max_duration, aspect_ratio, download_folder="downloaded_videos", max_results=6):
    script = generate_script(prompt, max_duration)
    videos = search_and_download_videos(prompt, max_duration, aspect_ratio, download_folder, max_results)
    if not videos:
        return "No videos were downloaded.", None, script

    video_clips = [VideoFileClip(video) for video in videos]
    final_video = aggregate_videos(video_clips)

    if final_video:
        narration_file = generate_narration(script)
        final_video = trim_video_to_audio_length(final_video, AudioFileClip(narration_file).duration)
        final_video = add_narration_to_video(final_video, narration_file)
        output_video_path = "final_video_with_narration.mp4"
        save_final_video(final_video, output_video_path)
        return narration_file, output_video_path, script
    return "Error generating video.", None, script


iface = gr.Interface(
    fn=generate_video,
    inputs=[
        gr.Textbox(label="Enter Text Prompt", placeholder="Enter the text to generate the video script."),
        gr.Slider(minimum=1, maximum=30, step=1, label="Video Length (seconds)", value=10),
        gr.Radio(choices=["portrait", "landscape", "square"], label="Select Aspect Ratio", value="landscape"),
    ],
    outputs=[
        gr.Audio(label="Narration Audio"),
        gr.Video(label="Generated Video"),
        gr.Textbox(label="Generated Script", interactive=False)
    ],
    title="Sepia Text-to-Video Generator",
    description="Enter a text prompt, specify the length of the video (maximum 30 seconds), select the aspect ratio, and click 'Submit' to get the narrated audio, the video and the script.",
    live=False
)

iface.launch(debug=True)