File size: 11,709 Bytes
da25681
 
 
 
 
 
a3e352f
f69554c
da25681
 
f69554c
 
 
 
da25681
 
 
 
2bce8ff
cbce214
da25681
 
cbce214
c9696b8
da25681
a23ccb7
2bce8ff
 
 
 
 
 
 
 
 
da25681
8882e69
f69554c
2bce8ff
a3e352f
da25681
2bce8ff
 
da25681
a3e352f
da25681
2bce8ff
 
 
da25681
a3e352f
f69554c
da25681
 
f69554c
da25681
a3e352f
8882e69
a3e352f
8882e69
 
f69554c
8882e69
 
 
f69554c
8882e69
 
 
 
a3e352f
c9696b8
f69554c
c9696b8
f69554c
c9696b8
 
 
f69554c
da25681
cbce214
da25681
cbce214
da25681
a3e352f
da25681
cbce214
 
 
 
 
da25681
a3e352f
f69554c
cbce214
f69554c
da25681
 
cbce214
f69554c
cbce214
a23ccb7
a3e352f
f69554c
 
cbce214
da25681
f69554c
a3e352f
cbce214
f69554c
cbce214
 
f69554c
da25681
a3e352f
f69554c
da25681
f69554c
 
da25681
 
a7c9b9d
f69554c
a7c9b9d
 
f69554c
 
 
a7c9b9d
8882e69
f69554c
a7c9b9d
8882e69
a7c9b9d
 
f69554c
a7c9b9d
8882e69
a23ccb7
a7c9b9d
 
 
 
 
f69554c
 
a7c9b9d
8882e69
f69554c
a7c9b9d
 
 
 
 
f69554c
da25681
f69554c
da25681
a3e352f
f69554c
b2b31f2
a3e352f
 
b2b31f2
f69554c
b2b31f2
a3e352f
b2b31f2
a3e352f
f69554c
a3e352f
 
b2b31f2
f69554c
b2b31f2
f69554c
b2b31f2
a3e352f
f69554c
b2b31f2
a3e352f
b2b31f2
a3e352f
f69554c
 
f959be9
8882e69
 
f69554c
f959be9
a3e352f
b2b31f2
a3e352f
da25681
f69554c
a3e352f
 
f69554c
a3e352f
 
b2b31f2
a3e352f
b2b31f2
 
a3e352f
 
f69554c
a3e352f
8882e69
f69554c
a3e352f
 
 
 
 
c9696b8
f69554c
a3e352f
cbce214
 
a3e352f
cbce214
f69554c
a3e352f
 
cbce214
a3e352f
d52bea3
f69554c
d52bea3
 
a3e352f
f69554c
d52bea3
f69554c
d52bea3
 
 
 
a3e352f
d52bea3
f69554c
d52bea3
 
b2b31f2
f69554c
b2b31f2
a3e352f
f69554c
b2b31f2
a3e352f
 
8882e69
d52bea3
 
 
f69554c
d52bea3
da25681
a23ccb7
8882e69
 
f69554c
a3e352f
 
f69554c
da25681
 
a23ccb7
cbce214
 
49e56cb
7eea0f7
 
a23ccb7
a3e352f
f959be9
7eea0f7
 
 
 
a3e352f
f69554c
a23ccb7
 
f69554c
a23ccb7
a3e352f
da25681
 
a3e352f
da25681
 
 
 
 
f69554c
da25681
d7e99cb
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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
import requests
import json
import time
import subprocess
import gradio as gr
import uuid
import os
import logging
from dotenv import load_dotenv

# Set up logging
logging.basicConfig(level=logging.DEBUG, format='%(asctime)s - %(levelname)s - %(message)s')
logger = logging.getLogger(__name__)

# Load environment variables
load_dotenv()

# API Keys
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
REPLICATE_API_TOKEN = os.getenv("REPLICATE_API_TOKEN")

# URLs
REPLICATE_API_URL = "https://api.replicate.com/v1/predictions"
UPLOAD_URL = os.getenv("UPLOAD_URL")

def get_voices():
    # OpenAI TTS voices
    return [
        ("alloy", "alloy"),
        ("echo", "echo"),
        ("fable", "fable"),
        ("onyx", "onyx"),
        ("nova", "nova"),
        ("shimmer", "shimmer")
    ]

def text_to_speech(voice, text, session_id):
    logger.info(f"Starting text-to-speech conversion for session {session_id}")
    url = "https://api.openai.com/v1/audio/speech"

    headers = {
        "Authorization": f"Bearer {OPENAI_API_KEY}",
        "Content-Type": "application/json"
    }

    data = {
        "model": "tts-1",
        "input": text,
        "voice": voice
    }

    logger.debug(f"Sending request to OpenAI TTS API for session {session_id}")
    response = requests.post(url, json=data, headers=headers)
    if response.status_code != 200:
        logger.error(f"Failed to generate speech audio for session {session_id}. Status code: {response.status_code}")
        return None

    # Save temporary audio file with session ID
    audio_file_path = f'tempvoice{session_id}.mp3'
    with open(audio_file_path, 'wb') as audio_file:
        audio_file.write(response.content)
    logger.info(f"Audio file saved: {audio_file_path}")
    return audio_file_path

def upload_file(file_path):
    logger.info(f"Uploading file: {file_path}")
    with open(file_path, 'rb') as file:
        files = {'fileToUpload': (os.path.basename(file_path), file)}
        data = {'reqtype': 'fileupload'}
        response = requests.post(UPLOAD_URL, files=files, data=data)

    if response.status_code == 200:
        logger.info(f"File uploaded successfully: {file_path}")
        return response.text.strip()
    logger.error(f"Failed to upload file: {file_path}. Status code: {response.status_code}")
    return None

def lipsync_api_call(video_url, audio_url):
    logger.info(f"Initiating lip-sync API call with video: {video_url} and audio: {audio_url}")
    headers = {
        "Authorization": f"Bearer {REPLICATE_API_TOKEN}",
        "Content-Type": "application/json",
        "Prefer": "wait"
    }

    data = {
        "version": "db5a650c807b007dc5f9e5abe27c53e1b62880d1f94d218d27ce7fa802711d67",
        "input": {
            "face": video_url,
            "input_audio": audio_url
        }
    }

    logger.debug(f"Sending request to Replicate API with data: {json.dumps(data)}")
    response = requests.post(REPLICATE_API_URL, headers=headers, json=data)
    logger.debug(f"Received response from Replicate API: {response.text}")
    return response.json()

def check_job_status(prediction_id):
    logger.info(f"Checking job status for prediction ID: {prediction_id}")
    headers = {"Authorization": f"Bearer {REPLICATE_API_TOKEN}"}
    max_attempts = 30  # Limit the number of attempts

    for attempt in range(max_attempts):
        logger.debug(f"Attempt {attempt + 1} to check job status")
        response = requests.get(f"{REPLICATE_API_URL}/{prediction_id}", headers=headers)
        data = response.json()
        logger.debug(f"Job status response: {json.dumps(data)}")

        if data["status"] == "succeeded":
            logger.info(f"Job completed successfully for prediction ID: {prediction_id}")
            return data["output"]
        elif data["status"] == "failed":
            logger.error(f"Job failed for prediction ID: {prediction_id}")
            return None

        logger.info(f"Job still in progress. Waiting for 10 seconds before next check.")
        time.sleep(10)
    
    logger.warning(f"Max attempts reached for prediction ID: {prediction_id}")
    return None

def get_media_duration(file_path):
    logger.info(f"Getting media duration for: {file_path}")
    cmd = ['ffprobe', '-v', 'error', '-show_entries', 'format=duration', '-of', 'default=noprint_wrappers=1:nokey=1', file_path]
    result = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    duration = float(result.stdout.strip())
    logger.info(f"Media duration: {duration} seconds")
    return duration

def combine_audio_video(video_path, audio_path, output_path):
    logger.info(f"Combining audio and video: video={video_path}, audio={audio_path}, output={output_path}")
    video_duration = get_media_duration(video_path)
    audio_duration = get_media_duration(audio_path)

    if video_duration > audio_duration:
        logger.info("Video longer than audio. Trimming video.")
        cmd = [
            'ffmpeg', '-i', video_path, '-i', audio_path,
            '-t', str(audio_duration),  # Trim video to audio duration
            '-map', '0:v', '-map', '1:a',
            '-c:v', 'copy', '-c:a', 'aac',
            '-y', output_path
        ]
    else:
        logger.info("Audio longer than video. Looping video.")
        loop_count = int(audio_duration // video_duration) + 1
        cmd = [
            'ffmpeg', '-stream_loop', str(loop_count), '-i', video_path, '-i', audio_path,
            '-t', str(audio_duration),
            '-map', '0:v', '-map', '1:a',
            '-c:v', 'copy', '-c:a', 'aac',
            '-shortest', '-y', output_path
        ]

    logger.debug(f"Running ffmpeg command: {' '.join(cmd)}")
    subprocess.run(cmd, check=True)
    logger.info(f"Audio and video combined successfully: {output_path}")

def create_video_from_image(image_url, session_id):
    logger.info(f"Creating video from image: {image_url}")
    response = requests.get(image_url)
    image_path = f"tempimage{session_id}.jpg"
    with open(image_path, "wb") as f:
        f.write(response.content)
    logger.info(f"Image downloaded: {image_path}")
    
    video_path = f"tempvideo{session_id}.mp4"
    cmd = [
        'ffmpeg', '-loop', '1', '-i', image_path,
        '-vf', 'scale=trunc(iw/2)*2:trunc(ih/2)*2',
        '-c:v', 'libx264', '-t', '10', '-pix_fmt', 'yuv420p',
        video_path
    ]
    logger.debug(f"Running ffmpeg command: {' '.join(cmd)}")
    subprocess.run(cmd, check=True)
    logger.info(f"Video created from image: {video_path}")
    
    os.remove(image_path)
    logger.info(f"Temporary image file removed: {image_path}")
    
    return video_path

def process_video(voice, url, text, progress=gr.Progress()):
    session_id = str(uuid.uuid4())
    logger.info(f"Starting video processing for session {session_id}")
    progress(0, desc="Generating speech...")
    audio_path = text_to_speech(voice, text, session_id)
    if not audio_path:
        logger.error(f"Failed to generate speech audio for session {session_id}")
        return None, "Failed to generate speech audio."

    progress(0.2, desc="Processing media...")

    try:
        logger.info(f"Checking content type of URL: {url}")
        response = requests.head(url)
        content_type = response.headers.get('Content-Type', '')
        logger.info(f"Content type of URL: {content_type}")
        
        if content_type.startswith('image'):
            progress(0.3, desc="Converting image to video...")
            video_path = create_video_from_image(url, session_id)
            video_url = upload_file(video_path)
        else:
            video_url = url

        logger.info(f"Video URL: {video_url}")
        progress(0.4, desc="Uploading audio...")
        audio_url = upload_file(audio_path)
        logger.info(f"Audio URL: {audio_url}")

        if not audio_url or not video_url:
            raise Exception("Failed to upload audio or video file")

        progress(0.5, desc="Initiating lipsync...")
        job_data = lipsync_api_call(video_url, audio_url)
        logger.info(f"Lipsync job data: {json.dumps(job_data)}")

        if "error" in job_data:
            raise Exception(job_data.get("error", "Unknown error"))

        prediction_id = job_data["id"]
        logger.info(f"Lipsync prediction ID: {prediction_id}")

        progress(0.6, desc="Processing lipsync...")
        result_url = check_job_status(prediction_id)

        if result_url:
            logger.info(f"Lipsync result URL: {result_url}")
            progress(0.9, desc="Downloading result...")
            response = requests.get(result_url)
            output_path = f"output{session_id}.mp4"
            with open(output_path, 'wb') as f:
                f.write(response.content)
            logger.info(f"Lipsync result saved to: {output_path}")
            progress(1.0, desc="Complete!")
            return output_path, "Lipsync completed successfully!"
        else:
            raise Exception("Lipsync processing failed or timed out")

    except Exception as e:
        logger.error(f"Error during lipsync process: {str(e)}")
        progress(0.8, desc="Falling back to simple combination...")
        try:
            if 'video_path' not in locals():
                logger.info("Downloading video from URL")
                video_response = requests.get(video_url)
                video_path = f"tempvideo{session_id}.mp4"
                with open(video_path, 'wb') as f:
                    f.write(video_response.content)

            output_path = f"output{session_id}.mp4"
            combine_audio_video(video_path, audio_path, output_path)
            progress(1.0, desc="Complete!")
            return output_path, f"Used fallback method. Original error: {str(e)}"
        except Exception as fallback_error:
            logger.error(f"Fallback method failed: {str(fallback_error)}")
            return None, f"All methods failed. Error: {str(fallback_error)}"
    finally:
        # Cleanup
        if os.path.exists(audio_path):
            os.remove(audio_path)
            logger.info(f"Removed temporary audio file: {audio_path}")
        if os.path.exists(f"tempvideo{session_id}.mp4"):
            os.remove(f"tempvideo{session_id}.mp4")
            logger.info(f"Removed temporary video file: tempvideo{session_id}.mp4")

def create_interface():
    voices = get_voices()
    
    with gr.Blocks() as app:
        gr.Markdown("# Generator")
        with gr.Row():
            with gr.Column():
                voice_dropdown = gr.Dropdown(choices=[v[0] for v in voices], label="Select Voice", value=voices[0][0] if voices else None)
                url_input = gr.Textbox(label="Enter Video or Image URL")
                text_input = gr.Textbox(label="Enter text", lines=3)
                generate_btn = gr.Button("Generate Video")
            with gr.Column():
                video_output = gr.Video(label="Generated Video")
                status_output = gr.Textbox(label="Status", interactive=False)
        def on_generate(voice_name, url, text):
            logger.info(f"Generation started with voice: {voice_name}, URL: {url}")
            voice_id = next((v[1] for v in voices if v[0] == voice_name), None)
            if not voice_id:
                logger.error(f"Invalid voice selected: {voice_name}")
                return None, "Invalid voice selected."
            return process_video(voice_id, url, text)
        generate_btn.click(
            fn=on_generate,
            inputs=[voice_dropdown, url_input, text_input],
            outputs=[video_output, status_output]
        )
    return app

if __name__ == "__main__":
    logger.info("Starting the application")
    app = create_interface()
    app.launch()