Update app.py

app.py

@@ -1,15 +1,19 @@
 from fastapi import FastAPI, HTTPException
 from fastapi.responses import FileResponse
-from gtts import gTTS
-import cv2
-import numpy as np
+import pyttsx3  # Faster TTS library
+import moviepy.editor as mp  # Efficient video generation
 from PIL import Image, ImageDraw, ImageFont
+import numpy as np
 import os
 from concurrent.futures import ThreadPoolExecutor
 import asyncio
+from io import BytesIO
 
 app = FastAPI()
 
+# Initialize pyttsx3 TTS engine
+engine = pyttsx3.init()
+
 # Function to split the script into smaller chunks
 def split_script(script: str, max_words: int = 30):
     words = script.split()
@@ -19,10 +23,10 @@ def split_script(script: str, max_words: int = 30):
 # Function to create a video segment from a script chunk
 def create_video_segment(script_chunk: str, background_color: str, text_color: str, font_size: int):
     try:
-        # Step 1: Convert script chunk to audio using gTTS
-        tts = gTTS(script_chunk)
+        # Step 1: Convert script chunk to audio using pyttsx3
         audio_file = f"output_audio_{os.urandom(4).hex()}.mp3"
-        tts.save(audio_file)
+        engine.save_to_file(script_chunk, audio_file)
+        engine.runAndWait()
 
         # Step 2: Create a blank image with text
         width, height = 1280, 720  # HD resolution
@@ -45,53 +49,34 @@ def create_video_segment(script_chunk: str, background_color: str, text_color: s
         text_y = (height - text_height) // 2
         draw.text((text_x, text_y), script_chunk, font=font, fill=text_color_rgb)
 
-        # Convert the image to a numpy array for OpenCV
+        # Convert the image to a numpy array for moviepy
         frame = np.array(image)
-        frame = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
 
-        # Step 3: Create a video segment with the image and audio
+        # Step 3: Create a video segment with the image and audio using moviepy
         video_segment_file = f"video_segment_{os.urandom(4).hex()}.mp4"
-        fps = 24
-        fourcc = cv2.VideoWriter_fourcc(*"mp4v")
-        video_writer = cv2.VideoWriter(video_segment_file, fourcc, fps, (width, height))
-
-        # Calculate the number of frames based on audio duration
-        audio_duration = len(script_chunk.split()) * 0.5  # Approximate duration (adjust as needed)
-        num_frames = int(audio_duration * fps)
-
-        # Write the frames to the video
-        for _ in range(num_frames):
-            video_writer.write(frame)
-
-        # Release the video writer
-        video_writer.release()
-
-        # Step 4: Add audio to the video segment using ffmpeg (if available)
-        if os.path.exists(audio_file):
-            final_video_segment_file = f"final_{video_segment_file}"
-            os.system(f"ffmpeg -i {video_segment_file} -i {audio_file} -c:v copy -c:a aac {final_video_segment_file}")
-            os.remove(video_segment_file)
-            os.remove(audio_file)
-            return final_video_segment_file
-        else:
-            return video_segment_file
+        clip = mp.ImageClip(frame).set_duration(len(script_chunk.split()) * 0.5)  # Approximate duration
+        audio_clip = mp.AudioFileClip(audio_file)
+        final_clip = clip.set_audio(audio_clip)
+        final_clip.write_videofile(video_segment_file, fps=24)
+
+        # Clean up temporary audio file
+        os.remove(audio_file)
+
+        return video_segment_file
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))
 
 # Function to combine video segments into a single video
 def combine_video_segments(video_segment_files: list, output_file: str = "final_output_video.mp4"):
     try:
-        with open("video_list.txt", "w") as f:
-            for segment in video_segment_files:
-                f.write(f"file '{segment}'\n")
-
-        # Combine video segments using ffmpeg
-        os.system(f"ffmpeg -f concat -safe 0 -i video_list.txt -c copy {output_file}")
+        # Combine video segments using moviepy
+        clips = [mp.VideoFileClip(segment) for segment in video_segment_files]
+        final_clip = mp.concatenate_videoclips(clips)
+        final_clip.write_videofile(output_file)
 
         # Clean up video segments
         for segment in video_segment_files:
             os.remove(segment)
-        os.remove("video_list.txt")
 
         return output_file
     except Exception as e: