Update app.py

app.py

@@ -5,15 +5,14 @@ import torch
 from PIL import Image, ImageDraw, ImageFont
 import scipy.io.wavfile
 from TTS.api import TTS  # Coqui TTS (open source)
-from moviepy.editor import CompositeVideoClip, ImageClip, AudioFileClip, concatenate_videoclips
+from moviepy.editor import ImageSequenceClip, AudioFileClip, concatenate_videoclips
 import os
-from groq import Groq
-from deepgram import Deepgram
 import subprocess
 
 # Initialize Clients
-groq_client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
-deepgram_client = Deepgram(api_key=os.environ.get("DEEGRAM_API_KEY"))
+# Replace with your actual API keys or methods of getting them
+# groq_client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
+# deepgram_client = Deepgram(api_key=os.environ.get("DEEGRAM_API_KEY"))
 
 # Use DistilGPT-2 for text generation
 script_generator = pipeline("text-generation", model="distilgpt2", truncation=True, max_length=100)
@@ -24,8 +23,9 @@ tts = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=False,
 # Use MusicGen for music generation
 music_generator = pipeline("text-to-audio", model="facebook/musicgen-small", device="cpu")
 
-# Use Stable Diffusion for image generation
-image_generator = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1-base", torch_dtype=torch.float32).to("cpu")
+# Use Fluently Anime (Stable Diffusion) for anime image generation
+model_id = "fluently/Fluently-anime"
+anime_image_generator = StableDiffusionPipeline.from_pretrained(model_id).to("cuda")
 
 # Generate Comedy Script using DistilGPT-2
 def generate_comedy_script(prompt):
@@ -38,13 +38,13 @@ def text_to_speech(script):
     tts.tts_to_file(text=script, file_path=output_audio)
     return output_audio
 
-# Create Images Using Stable Diffusion
+# Create Anime Images Using Fluently Anime
 def create_images_from_script(script):
     lines = script.split('. ')
     image_paths = []
     for i, line in enumerate(lines):
-        img = image_generator(line).images[0]
-        img_path = f'/tmp/image_{i}.png'
+        img = anime_image_generator(line).images[0]
+        img_path = f'/tmp/anime_image_{i}.png'
         img.save(img_path)
         image_paths.append(img_path)
     return image_paths
@@ -57,16 +57,18 @@ def generate_fun_music(prompt, output_music_file="fun_music.wav"):
     scipy.io.wavfile.write(output_music_file, rate=sampling_rate, data=audio_data)
     return output_music_file
 
-# Create Video from Generated Images Using PIA
+# Create Video from Generated Anime Images Using FFmpeg
 def generate_text_video(script):
     image_paths = create_images_from_script(script)
-    # PIA logic here
-    subprocess.run(["python", "-m", "pia", "generate", "--input", " ".join(image_paths), "--output", "/tmp/final_video.mp4"])
-    return "/tmp/final_video.mp4"
+    # Generate video using moviepy from the sequence of images
+    video_clip = ImageSequenceClip(image_paths, fps=24)
+    video_path = "/tmp/final_video.mp4"
+    video_clip.write_videofile(video_path, codec='libx264')
+    return video_path
 
 # Combine Audio and Video
 def combine_audio_video(video_path, audio_path):
-    video = VideoFileClip(video_path)
+    video = AudioFileClip(video_path)
     audio = AudioFileClip(audio_path)
     final_video = video.set_audio(audio)
     final_video.write_videofile("/tmp/final_comedy_video.mp4", fps=24)
@@ -93,8 +95,7 @@ def generate_kids_content(theme):
         frame_path = f'/tmp/kids_temp_{i}.png'
         img.save(frame_path)
         clips.append(ImageClip(frame_path).set_duration(1).set_position(('center', 'center')))
-    final_video = CompositeVideoClip(clips, size=(800, 400))
-    final_video = final_video.set_audio(AudioFileClip(music_file))
+    final_video = concatenate_videoclips(clips, method="compose").set_audio(AudioFileClip(music_file))
     final_video.write_videofile("/tmp/kids_animation.mp4", fps=24)
     return music_file, "/tmp/kids_animation.mp4"
 
@@ -131,3 +132,4 @@ with gr.Blocks() as app:
 
 app.launch()
 
+