Update app.py

app.py

@@ -1,9 +1,10 @@
 import os
 import math
+import re
 import gradio as gr
 import torch
 from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
-from moviepy.editor import VideoFileClip, concatenate_audioclips
+from moviepy.editor import VideoFileClip
 
 def transcribe(video_file, transcribe_to_text, transcribe_to_srt, language):
     device = "cuda:0" if torch.cuda.is_available() else "cpu"
@@ -79,12 +80,15 @@ def transcribe(video_file, transcribe_to_text, transcribe_to_srt, language):
         srt_content = ""
         for i, sub in enumerate(transcription_srt, 1):
             srt_entry = f"{i}\n{format_time(sub['start'])} --> {format_time(sub['end'])}\n{sub['text']}\n\n"
-            output += srt_entry
             srt_content += srt_entry
         
+        # Remove duplicate captions
+        cleaned_srt_content = clean_srt_duplicates(srt_content)
+        
+        # Save SRT content to a file
         srt_file_path = "transcription.srt"
         with open(srt_file_path, "w", encoding="utf-8") as srt_file:
-            srt_file.write(srt_content)
+            srt_file.write(cleaned_srt_content)
         
         output += f"\nSRT file saved as: {srt_file_path}"
     
@@ -95,6 +99,30 @@ def format_time(seconds):
     h, m = divmod(m, 60)
     return f"{int(h):02d}:{int(m):02d}:{s:06.3f}".replace('.', ',')
 
+def clean_srt_duplicates(srt_content):
+    """
+    Function to remove consecutive duplicate captions in SRT format.
+    """
+    cleaned_srt = []
+    last_text = None
+
+    # Pattern to match each SRT block
+    srt_pattern = re.compile(r"(\d+)\n(\d{2}:\d{2}:\d{2},\d{3}) --> (\d{2}:\d{2}:\d{2},\d{3})\n(.+)", re.DOTALL)
+    
+    for match in srt_pattern.finditer(srt_content):
+        index, start_time, end_time, text = match.groups()
+        text = text.strip()
+        
+        # If this caption is identical to the last one, skip it
+        if text == last_text:
+            continue
+        
+        # Add cleaned entry to the list
+        cleaned_srt.append(f"{index}\n{start_time} --> {end_time}\n{text}\n\n")
+        last_text = text
+
+    return ''.join(cleaned_srt)
+
 iface = gr.Interface(
     fn=transcribe,
     inputs=[