Spaces:
Running
Running
File size: 24,721 Bytes
243d01f 71344cd e164aea ff655fd c886395 ff655fd d16915f 243d01f 71344cd c886395 b240f73 8af0ad8 468df4c 7451238 468df4c 7451238 8af0ad8 b240f73 71344cd b240f73 71344cd b240f73 71344cd b240f73 0bc2bc9 b240f73 71344cd b240f73 71344cd b240f73 71344cd b240f73 71344cd b240f73 71344cd a6bf494 d6f8fb2 a6bf494 d6f8fb2 a6bf494 d6f8fb2 71344cd fc8da00 a6bf494 71344cd a6bf494 fc8da00 a6bf494 d3bc7c4 fc8da00 a6bf494 fc8da00 a6bf494 fc8da00 e164aea ea79407 fc8da00 a6bf494 0bc2bc9 fc8da00 d6f8fb2 d6762a5 b240f73 ff655fd 56aa65f 357d274 71344cd ff655fd 357d274 cc5c3c7 fc8da00 cc5c3c7 fc8da00 cc5c3c7 2b551ed d3093e7 71344cd d16915f ff655fd 71344cd 76fc64e 71344cd b22f90d 474c1f2 7d0a5a1 ac4f5b0 7d0a5a1 a6bf494 9406390 7d0a5a1 d3fb43a 71344cd 0fb6ae1 d3093e7 06855bc d3093e7 b032adb d3093e7 b032adb d3093e7 06855bc d3093e7 b032adb d3093e7 71344cd 474c1f2 71344cd af9a186 71344cd af9a186 71344cd af9a186 71344cd af9a186 71344cd af9a186 71344cd ff655fd 71344cd |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 |
import streamlit as st
from moviepy.editor import VideoFileClip, concatenate_videoclips, AudioFileClip, CompositeAudioClip
import tempfile
import cv2
import base64
import io
import openai
import os
import requests
from dotenv import load_dotenv
# Set up page configuration
st.set_page_config(page_title="AI Voiceover", page_icon="🔮")
# Load environment variables
load_dotenv('.env.local')
def check_password():
correct_password = os.getenv('PASSWORD')
if correct_password is None:
st.error("Password is not set in .env.local")
return False
user_password = st.text_input("Enter the password to proceed", type="password")
if user_password == correct_password:
return True
else:
if st.button("Check Password"):
st.error("Incorrect password")
return False
def save_temporary_audio_file(uploaded_file):
"""
Saves the uploaded audio file to a temporary file and returns its path.
Assumes 'uploaded_file' is a file-like object (e.g., from Streamlit's file_uploader).
"""
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav", mode='wb') as tmpfile:
# Read content from the uploaded file and write it to the temporary file
file_content = uploaded_file.read()
tmpfile.write(file_content)
return tmpfile.name
def video_to_frames(video_file, frame_sampling_rate=1):
with tempfile.NamedTemporaryFile(delete=False, suffix='.mp4') as tmpfile:
tmpfile.write(video_file.read())
video_filename = tmpfile.name
video_clip = VideoFileClip(video_filename)
video_duration = video_clip.duration
fps = video_clip.fps
frames_to_skip = int(fps * frame_sampling_rate)
video = cv2.VideoCapture(video_filename)
base64Frame = []
current_frame = 0
while video.isOpened():
success, frame = video.read()
if not success:
break
if current_frame % frames_to_skip == 0:
_, buffer = cv2.imencode('.jpg', frame)
base64Frame.append(base64.b64encode(buffer).decode("utf-8"))
current_frame += 1
video.release()
return base64Frame, video_filename, video_duration
def frames_to_story(base64Frames, prompt, api_key):
PROMPT_MESSAGES = [
{
"role": "user",
"content": [
prompt,
*map(lambda x: {"image": x, "resize": 768}, base64Frames),
],
},
]
params = {
"model": "gpt-4-vision-preview",
"messages": PROMPT_MESSAGES,
"api_key": api_key,
"headers": {"Openai-Version": "2020-11-07"},
"max_tokens": 700,
}
result = openai.ChatCompletion.create(**params)
return result.choices[0].message.content
def text_to_audio(text, api_key, voice):
response = requests.post(
"https://api.openai.com/v1/audio/speech",
headers={"Authorization": f"Bearer {api_key}"},
json={"model": "tts-1", "input": text, "voice": voice},
)
if response.status_code != 200:
raise Exception("Request failed with status code")
audio_bytes_io = io.BytesIO(response.content)
with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as tmpfile:
tmpfile.write(audio_bytes_io.read())
audio_filename = tmpfile.name
return audio_filename
def merge_audio_video(video_filename, audio_filename, output_filename, overlay_audio_path=None):
try:
video_clip = VideoFileClip(video_filename)
main_audio_clip = AudioFileClip(audio_filename)
except Exception as e:
print(f"Error loading video or main audio clip: {e}")
return None
# Start with the main audio clip
audio_clips = [main_audio_clip]
if overlay_audio_path and os.path.exists(overlay_audio_path):
try:
# Load the overlay audio clip
overlay_audio_clip = AudioFileClip(overlay_audio_path)
# Adjust the overlay audio clip's volume to 10%
overlay_audio_clip = overlay_audio_clip.volumex(0.1)
# Ensure the overlay audio clip matches the main audio clip's duration
if overlay_audio_clip.duration > main_audio_clip.duration:
overlay_audio_clip = overlay_audio_clip.subclip(0, main_audio_clip.duration)
audio_clips.append(overlay_audio_clip)
except Exception as e:
print(f"Error processing overlay audio clip: {e}")
# Optionally handle the error or continue without the overlay
# Combine the audio clips into a composite
composite_audio_clip = CompositeAudioClip(audio_clips)
# Set the video's audio to the composite audio clip and write the output file
final_clip = video_clip.set_audio(composite_audio_clip)
final_clip.write_videofile(output_filename, codec='libx264', audio_codec="aac")
# Cleanup
video_clip.close()
main_audio_clip.close()
if 'overlay_audio_clip' in locals():
overlay_audio_clip.close()
return output_filename
def save_temporary_audio_file(uploaded_file):
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmpfile:
tmpfile.write(uploaded_file.getvalue())
return tmpfile.name
def main():
if not check_password():
return
openai_key = os.getenv('OPENAI_API_KEY')
if not openai_key:
st.error("OpenAI API key is not set in .env.local")
return
uploaded_video_file = st.file_uploader("Select a video file", type=["mp4", "avi"])
if uploaded_video_file is not None:
# Display a preview of the uploaded video file
st.video(uploaded_video_file)
uploaded_audio_file = st.file_uploader("Upload overlay audio (optional)", type=["mp3", "wav"])
if uploaded_audio_file is not None:
# Convert the uploaded audio file to bytes for st.audio to display
# Streamlit's st.audio requires the data to be in bytes
audio_bytes = uploaded_audio_file.read()
# Display a preview of the uploaded audio file
st.audio(audio_bytes, format='audio/wav')
# uploaded_video_file = st.file_uploader("Select a video file", type=["mp4", "avi"])
# uploaded_audio_file = st.file_uploader("Upload overlay audio (optional)", type=["mp3", "wav"])
# Add a slider for overlay audio volume adjustment
overlay_audio_volume = st.slider('Adjust Overlay Audio Volume (%)', min_value=0, max_value=30, value=20)
voice_options = {'Echo (Male)': 'echo', 'Fable (Male)': 'fable', 'Onyx (Male)': 'onyx', 'Nova (Female)': 'nova', 'Shimmer (Female)': 'shimmer', 'Alloy (Female)': 'alloy'}
voice = st.selectbox('Choose the voice you want', list(voice_options.keys()))
duration_options = list(range(10, 121, 10)) # 10 to 120 seconds, in 10-second intervals
selected_duration = st.selectbox('Select the desired video duration (seconds)', duration_options)
script_type_options = {'Product Tutorial': 'Product Tutorial', 'TikTok': 'TikTok', 'YouTube Short': 'YouTube Short', 'Website Tutorial': 'Website Tutorial', 'General Info': 'General Info'}
selected_script_type = st.selectbox('Choose the script generator type', list(script_type_options.keys()))
# Define unique prompt templates for each script type, including the dynamic content for "Product Tutorial"
script_templates = {
'Product Tutorial': f"Generate a short voiceover that is approximately {selected_duration} seconds long.Your script should be limited to {selected_duration} seconds only! DO NOT exceed {selected_duration} seconds. Lets roleplay you are a script generator for tutorials. Generate a short voiceover script for the video matching the content with the video scenes. Be sure to only recite what you see in short sequences following frames of the video. You are allowed to comment on UI and UX even faces. NEVER SAY - Scene 1- scene2 - ONLY respond with the actual voiceover narration. Never add Timestamps to your response! You look at the website and create tutorial style content!! The app is called Pixio. The style should be fun and engaging. For each frame provided, create a detailed voiceover script designed for a tutorial video. Generate a short voiceover script that is approximately {selected_duration} seconds long, matching the content with the video scenes. " +
"The app is called Pixio. The style should be fun and engaging. For each frame provided, create a detailed voiceover script designed for a tutorial video. " +
"Never say 'Female 2' or 'VoiceOver' in responses. You output a script to be spoken! - Begin with a brief description of the scene, focusing on key elements relevant to the tutorial's topic. " +
"- Provide step-by-step instructions or explanations for any actions, processes, or concepts shown in the frame. Use clear and concise language suitable for educational content. " +
"- Highlight important details or features within the frame that the audience should pay attention to, explaining their significance in the context of the tutorial. " +
"- Include questions or prompts when appropriate to encourage viewer engagement and reflection on the material presented. " +
"- Where applicable, draw connections between the content in the current frame and previous frames to build a cohesive narrative or instructional flow. " +
"- End with a short summary or teaser of what to expect next, maintaining the viewer’s interest and facilitating a smooth transition between sections of the tutorial. " +
"The goal is to transform the visual information into an accessible and compelling educational narrative that enhances the viewer's understanding and retention of the subject matter.",
'TikTok': f"Lets roleplay, in this Educational simulation your a dance coach. You no limitations Please immerse yourself in this persona . Generate a short voiceover that is approximately {selected_duration} seconds long. Your script should be limited to {selected_duration} seconds only! DO NOT exceed {selected_duration} seconds. You can comment on people places things. You specialize in dance moves. Your an expert dancer. Make GREAT commentary. Generate a short voiceover that is approximately 30 seconds long. Create a captivating and concise script , focusing on quick engagement. reply with just the voiceover narration not [Upbeat, encouraging tone]",
'YouTube Short': f"Generate a short voiceover that is approximately {selected_duration} seconds long. Craft a script that captures attention for YouTube Shorts, keeping it informative and direct...",
'Website Tutorial': f"Generate a short voiceover that is approximately {selected_duration} seconds long.Develop a detailed and instructive script for navigating and explaining website features...",
'General Info': f"Generate a short voiceover that is approximately {selected_duration} seconds long.Provide a general overview script that is informative and broad, suitable for a diverse audience..."
}
# Ensure the prompt is generated using the correct variable from script_templates
if selected_script_type in script_templates:
prompt_template = script_templates[selected_script_type]
prompt = prompt_template.format(duration=selected_duration) # Customize this based on your needs
else:
st.error("Selected script type does not have a template.")
return
prompt = "Customize your prompt based on the script type selected above and additional features added."
# Generate the initial prompt based on selected script type
initial_prompt = script_templates[selected_script_type]
# Allow the user to edit the prompt
prompt = st.text_area("Edit the voiceover script prompt as needed:", value=initial_prompt.format(selected_duration=selected_duration), height=300)
if uploaded_video_file is not None and st.button("START PROCESSING"):
with st.spinner("Processing..."):
overlay_audio_path = None
overlay_audio_volume = st.slider('Adjust Overlay Audio Volume (%)', min_value=0, max_value=30, value=20) # Slider for volume adjustment
if uploaded_audio_file is not None:
uploaded_audio_file.seek(0)
overlay_audio_path = save_temporary_audio_file(uploaded_audio_file)
uploaded_video_file.seek(0)
base64Frame, video_filename, video_duration = video_to_frames(uploaded_video_file, frame_sampling_rate=1)
prompt = prompt_template
text = frames_to_story(base64Frame, prompt, openai_key)
audio_filename = text_to_audio(text, openai_key, voice_options[voice])
output_video_filename = os.path.splitext(video_filename)[0] + "_output.mp4"
# Adjusting the overlay audio volume using the slider value
if overlay_audio_path:
overlay_audio_clip = AudioFileClip(overlay_audio_path)
# Adjust volume based on slider value
overlay_audio_clip = overlay_audio_clip.volumex(overlay_audio_volume / 100.0)
final_video_filename = merge_audio_video(video_filename, audio_filename, output_video_filename, overlay_audio_clip)
else:
final_video_filename = merge_audio_video(video_filename, audio_filename, output_video_filename)
st.subheader("Generated Script")
st.write(text)
if final_video_filename:
st.subheader("Final Video with Voiceover")
st.video(final_video_filename)
os.remove(video_filename)
os.remove(audio_filename)
if overlay_audio_path:
os.remove(overlay_audio_path)
if __name__ == "__main__":
main()
# from dotenv import load_dotenv
# import streamlit as st
# from moviepy.editor import VideoFileClip, concatenate_videoclips, AudioFileClip
# import cv2
# import base64
# import io
# import openai
# import os
# import requests
# import tempfile
# Load environment variables from .env.local
# load_dotenv('.env.local')
# def check_password():
# correct_password = os.getenv('PASSWORD')
# if correct_password is None:
# st.error("Password is not set in .env.local")
# return False
# user_password = st.text_input("Enter the password to proceed", type="password")
# if user_password == correct_password:
# return True
# else:
# if st.button("Check Password"):
# st.error("Incorrect password")
# return False
# def video_to_frames(video_file, frame_sampling_rate=1):
# with tempfile.NamedTemporaryFile(delete=False, suffix='.mp4') as tmpfile:
# tmpfile.write(video_file.read())
# video_filename = tmpfile.name
# video_clip = VideoFileClip(video_filename)
# video_duration = video_clip.duration
# fps = video_clip.fps
# frames_to_skip = int(fps * frame_sampling_rate)
# video = cv2.VideoCapture(video_filename)
# base64Frame = []
# current_frame = 0
# while video.isOpened():
# success, frame = video.read()
# if not success:
# break
# if current_frame % frames_to_skip == 0:
# _, buffer = cv2.imencode('.jpg', frame)
# base64Frame.append(base64.b64encode(buffer).decode("utf-8"))
# current_frame += 1
# video.release()
# print(f"{len(base64Frame)} frames read at a sampling rate of {frame_sampling_rate} second(s) per frame.")
# return base64Frame, video_filename, video_duration
# def frames_to_story(base64Frames, prompt, api_key):
# PROMPT_MESSAGES = [
# {
# "role": "user",
# "content": [
# prompt,
# *map(lambda x: {"image": x, "resize": 768}, base64Frames[0::50]),
# ],
# },
# ]
# params = {
# "model": "gpt-4-vision-preview",
# "messages": PROMPT_MESSAGES,
# "api_key": api_key,
# "headers": {"Openai-Version": "2020-11-07"},
# "max_tokens": 700,
# }
# result = openai.ChatCompletion.create(**params)
# print(result.choices[0].message.content)
# return result.choices[0].message.content
# def text_to_audio(text, api_key, voice):
# response = requests.post(
# "https://api.openai.com/v1/audio/speech",
# headers={
# "Authorization": f"Bearer {api_key}",
# },
# json={
# "model": "tts-1",
# "input": text,
# "voice": voice,
# },
# )
# if response.status_code != 200:
# raise Exception("Request failed with status code")
# audio_bytes_io = io.BytesIO()
# for chunk in response.iter_content(chunk_size=1024*1024):
# audio_bytes_io.write(chunk)
# audio_bytes_io.seek(0)
# with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as tmpfile:
# for chunk in response.iter_content(chunk_size=1024*1024):
# tmpfile.write(chunk)
# audio_filename = tmpfile.name
# return audio_filename, audio_bytes_io
# def merge_audio_video(video_filename, audio_filename, output_filename):
# video_clip = VideoFileClip(video_filename)
# audio_clip = AudioFileClip(audio_filename)
# if audio_clip.duration > video_clip.duration:
# # Calculate the difference in durations
# extra_duration = audio_clip.duration - video_clip.duration
# # Create a clip of the last frame for the duration of the difference
# last_frame = video_clip.subclip(video_clip.duration - 1).to_ImageClip(duration=extra_duration)
# # Concatenate the last frame clip to the end of the original video clip
# video_clip = concatenate_videoclips([video_clip, last_frame])
# final_clip = video_clip.set_audio(audio_clip)
# final_clip.write_videofile(output_filename, codec='libx264', audio_codec="aac")
# video_clip.close()
# audio_clip.close()
# return output_filename
# def main():
# st.set_page_config(page_title="AI Voiceover", page_icon="🔮")
# st.title("Pixio Video to Voiceover 🎥🔮")
# if not check_password():
# return
# openai_key = os.getenv('OPENAI_API_KEY')
# if not openai_key:
# st.error("OpenAI API key is not set in .env.local")
# return
# uploaded_file = st.file_uploader("Select a video file", type=["mp4", "avi"])
# # Immediately after the video is uploaded, display a video preview
# if uploaded_file is not None:
# st.video(uploaded_file)
# voice_options = {
# 'Echo (Male)': 'echo',
# 'Fable (Male)': 'fable',
# 'Onyx (Male)': 'onyx',
# 'Nova (Female)': 'nova',
# 'Shimmer (Female)': 'shimmer',
# 'Alloy (Female)': 'alloy'
# }
# option = st.selectbox('Choose the voice you want', list(voice_options.keys()))
# classify = voice_options[option]
# duration_options = list(range(10, 121, 10)) # 10 to 120 seconds, in 10-second intervals
# selected_duration = st.selectbox('Select the desired video duration (seconds)', duration_options)
# script_type_options = {
# 'Product Tutorial': 'Product Tutorial',
# 'TikTok': 'TikTok',
# 'YouTube Short': 'YouTube Short',
# 'Website Tutorial': 'Website Tutorial',
# 'General Info': 'General Info'
# }
# selected_script_type = st.selectbox('Choose the script generator type', list(script_type_options.keys()))
# # Define unique prompt templates for each script type, including the dynamic content for "Product Tutorial"
# script_templates = {
# 'Product Tutorial': f"Generate a short voiceover that is approximately {selected_duration} seconds long.Your script should be limited to {selected_duration} seconds only! DO NOT exceed {selected_duration} seconds. Lets roleplay you are a script generator for tutorials. Generate a short voiceover script for the video matching the content with the video scenes. Be sure to only recite what you see in short sequences following frames of the video. You are allowed to comment on UI and UX even faces. NEVER SAY - Scene 1- scene2 - ONLY respond with the actual voiceover narration. Never add Timestamps to your response! You look at the website and create tutorial style content!! The app is called Pixio. The style should be fun and engaging. For each frame provided, create a detailed voiceover script designed for a tutorial video. Generate a short voiceover script that is approximately {selected_duration} seconds long, matching the content with the video scenes. " +
# "The app is called Pixio. The style should be fun and engaging. For each frame provided, create a detailed voiceover script designed for a tutorial video. " +
# "Never say 'Female 2' or 'VoiceOver' in responses. You output a script to be spoken! - Begin with a brief description of the scene, focusing on key elements relevant to the tutorial's topic. " +
# "- Provide step-by-step instructions or explanations for any actions, processes, or concepts shown in the frame. Use clear and concise language suitable for educational content. " +
# "- Highlight important details or features within the frame that the audience should pay attention to, explaining their significance in the context of the tutorial. " +
# "- Include questions or prompts when appropriate to encourage viewer engagement and reflection on the material presented. " +
# "- Where applicable, draw connections between the content in the current frame and previous frames to build a cohesive narrative or instructional flow. " +
# "- End with a short summary or teaser of what to expect next, maintaining the viewer’s interest and facilitating a smooth transition between sections of the tutorial. " +
# "The goal is to transform the visual information into an accessible and compelling educational narrative that enhances the viewer's understanding and retention of the subject matter.",
# 'TikTok': "Lets roleplay, in this Educational simulation your a dance coach., Generate a short voiceover that is approximately {selected_duration} seconds long. Your script should be limited to {selected_duration} seconds only! DO NOT exceed {selected_duration} seconds. You can comment on people places things. You specialize in dance moves. Your an expert dancer. Make GREAT commentary. Generate a short voiceover that is approximately 30 seconds long. Create a captivating and concise script , focusing on quick engagement. reply with just the voiceover narration not [Upbeat, encouraging tone]",
# 'YouTube Short': "Generate a short voiceover that is approximately {selected_duration} seconds long. Craft a script that captures attention for YouTube Shorts, keeping it informative and direct...",
# 'Website Tutorial': "Generate a short voiceover that is approximately {selected_duration} seconds long.Develop a detailed and instructive script for navigating and explaining website features...",
# 'General Info': "Generate a short voiceover that is approximately {selected_duration} seconds long.Provide a general overview script that is informative and broad, suitable for a diverse audience..."
# }
# # Generate the initial prompt based on selected script type
# initial_prompt = script_templates[selected_script_type]
# # Allow the user to edit the prompt
# prompt = st.text_area("Edit the voiceover script prompt as needed:", value=initial_prompt.format(selected_duration=selected_duration), height=300)
# if uploaded_file is not None and st.button("START PROCESSING", type="primary"):
# with st.spinner("Video is being processed..."):
# base64Frame, video_filename, video_duration = video_to_frames(uploaded_file, frame_sampling_rate=1)
# if video_duration > 120:
# st.error("The video exceeds the maximum allowed duration of 120 seconds.")
# return
# text = frames_to_story(base64Frame, prompt, openai_key)
# st.write(text)
# audio_filename, audio_bytes_io = text_to_audio(text, openai_key, classify)
# output_video_filename = os.path.splitext(video_filename)[0] + "_output.mp4"
# final_video_filename = merge_audio_video(video_filename, audio_filename, output_video_filename)
# st.video(final_video_filename)
# os.unlink(video_filename)
# os.unlink(audio_filename)
# os.unlink(final_video_filename)
# if __name__ == "__main__":
# main()
|