get_rizzed / app.py
1littlecoder's picture
Update app.py
d92281e verified
raw
history blame
3.95 kB
import os
import google.generativeai as genai
import gradio as gr
import requests
from moviepy.editor import ImageClip, AudioFileClip
# Configure Google Gemini API
genai.configure(api_key=os.getenv("GEMINI_API_KEY"))
# Play.ht API keys
API_KEY = os.getenv('PLAY_API_KEY')
USER_ID = os.getenv('PLAY_USER_ID')
# theme selection
theme = gr.themes.Base(
primary_hue="emerald",
)
# Function to upload image to Gemini and get roasted text
def upload_to_gemini(path, mime_type="image/jpeg"):
file = genai.upload_file(path, mime_type=mime_type)
return file
def generate_roast(image_path):
# Upload the image to Gemini and get the text
uploaded_file = upload_to_gemini(image_path)
generation_config = {
"temperature": 1,
"top_p": 0.95,
"top_k": 40,
"max_output_tokens": 8192,
"response_mime_type": "text/plain",
}
model = genai.GenerativeModel(
model_name="gemini-1.5-flash-002",
generation_config=generation_config,
system_instruction="You are a professional satirist and fashion expert. You will be given a profile picture. Your duty is to roast in less than 50 words whatever is given to you in the funniest way possible!",
)
chat_session = model.start_chat(
history=[{"role": "user", "parts": [uploaded_file]}]
)
response = chat_session.send_message("Roast this image!")
return response.text
# Function to convert text to speech with Play.ht
def text_to_speech(text):
url = "https://api.play.ht/api/v2/tts/stream"
payload = {
"voice": "s3://voice-cloning-zero-shot/d9ff78ba-d016-47f6-b0ef-dd630f59414e/female-cs/manifest.json",
"output_format": "mp3",
"text": text,
}
headers = {
"accept": "audio/mpeg",
"content-type": "application/json",
"Authorization": API_KEY,
"X-User-ID": USER_ID
}
response = requests.post(url, json=payload, headers=headers)
if response.status_code == 200:
audio_path = "output_audio.mp3"
with open(audio_path, "wb") as audio_file:
audio_file.write(response.content)
return audio_path
else:
print(f"Error: {response.status_code} - {response.text}")
return None # Return None if there's an error
# Function to create video from image and audio
def generate_video(image_path, audio_path):
if audio_path is None:
return None # Skip video generation if there's no valid audio file
image_clip = ImageClip(image_path).set_duration(AudioFileClip(audio_path).duration)
audio_clip = AudioFileClip(audio_path)
video_clip = image_clip.set_audio(audio_clip)
video_output_path = "output_video.mp4"
video_clip.write_videofile(video_output_path, codec="libx264", audio_codec="aac")
return video_output_path
# Gradio Interface
with gr.Blocks(theme=theme) as demo:
gr.Markdown("# Image to Text-to-Speech Roasting App")
gr.Markdown("Upload an image, and the AI will roast it, convert the roast to audio, and create a video output.")
with gr.Row():
with gr.Column():
image_input = gr.Image(type="filepath", label="Upload Image")
with gr.Column():
output_text = gr.Textbox(label="Roast Text")
audio_output = gr.Audio(label="Roast Audio")
video_output = gr.Video(label="Roast Video")
def process_image(image):
roast_text = generate_roast(image)
audio_path = text_to_speech(roast_text)
if audio_path is None:
return "Error generating audio. Please try again.", None, None
video_path = generate_video(image, audio_path)
return roast_text, audio_path, video_path
submit_button = gr.Button("Generate Roast")
submit_button.click(process_image, inputs=image_input, outputs=[output_text, audio_output, video_output])
# Launch the app
demo.launch(debug=True)