Update app.py

app.py

@@ -7,6 +7,30 @@ import os
 import io
 import tempfile
 
+def estimate_audio_length(text):
+    # Estimate 150 words per minute
+    word_count = len(text.split())
+    return word_count / 150  # Returns length in minutes
+
+def trim_dialogue(dialogue, target_length_minutes):
+    trimmed_dialogue = []
+    current_length = 0
+    target_length_seconds = target_length_minutes * 60
+
+    for item in dialogue.dialogue:
+        item_length = estimate_audio_length(item.text)
+        if current_length + item_length > target_length_minutes:
+            # Trim this item to fit
+            words = item.text.split()
+            words_to_keep = int((target_length_minutes - current_length) * 150)
+            item.text = " ".join(words[:words_to_keep]) + "..."
+            trimmed_dialogue.append(item)
+            break
+        trimmed_dialogue.append(item)
+        current_length += item_length
+
+    return trimmed_dialogue
+
 def generate_podcast(file, tone, length):
     # Extract text from PDF
     if not file.name.lower().endswith('.pdf'):
@@ -31,11 +55,17 @@ def generate_podcast(file, tone, length):
     except Exception as e:
         raise gr.Error(f"Error generating script: {str(e)}")
     
+    # Determine target length in minutes
+    target_length = 2 if length == "Short (1-2 min)" else 5
+
+    # Trim dialogue to fit target length
+    trimmed_dialogue = trim_dialogue(script, target_length)
+    
     # Generate audio for each dialogue item
     audio_segments = []
     transcript = ""
     try:
-        for item in script.dialogue:
+        for item in trimmed_dialogue:
             audio_file = generate_audio(item.text, item.speaker)
             audio_segment = AudioSegment.from_mp3(audio_file)
             audio_segments.append(audio_segment)
@@ -47,11 +77,10 @@ def generate_podcast(file, tone, length):
     # Combine audio segments
     combined_audio = sum(audio_segments)
     
-    # Adjust length if needed
-    if length == "Short (1-2 min)":
-        combined_audio = combined_audio[:120000]  # 2 minutes max
-    else:  # "Medium (3-5 min)"
-        combined_audio = combined_audio[:300000]  # 5 minutes max
+    # Ensure audio doesn't exceed target length
+    target_length_ms = target_length * 60 * 1000
+    if len(combined_audio) > target_length_ms:
+        combined_audio = combined_audio[:target_length_ms]
     
     # Save combined audio to a temporary file
     with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as temp_audio:
@@ -60,6 +89,7 @@ def generate_podcast(file, tone, length):
     
     return temp_audio_path, transcript
 
+# Gradio interface setup remains the same
 instructions = """
 # Podcast Generator