Update app.py

app.py

@@ -20,26 +20,33 @@ def format_time(seconds):
     secs = seconds % 60
     return f"{hrs:02}:{mins:02}:{secs:02},{millis:03}"
 
-# Function to split text into segments by punctuation or limit to 7-8 words
+# Function to split text into segments without cutting words awkwardly
 def split_text_into_segments(text):
     segments = []
-    raw_segments = re.split(r'([.!?])', text)
-    for i in range(0, len(raw_segments) - 1, 2):
-        sentence = raw_segments[i].strip() + raw_segments[i + 1]
+    sentences = re.split(r'([.!?])', text)
+    for i in range(0, len(sentences) - 1, 2):
+        sentence = sentences[i].strip() + sentences[i + 1]
         words = sentence.split()
-        
+
+        # Ensure full phrases by keeping each segment between 7 to 8 words
         if len(words) > 8:
-            for j in range(0, len(words), 8):
-                segments.append(" ".join(words[j:j + 8]))
+            segment = ""
+            for word in words:
+                if len(segment.split()) < 8:
+                    segment += " " + word
+                else:
+                    segments.append(segment.strip())
+                    segment = word
+            if segment:
+                segments.append(segment.strip())
         else:
             segments.append(sentence.strip())
-    
-    if len(raw_segments) % 2 == 1:
-        remaining_text = raw_segments[-1].strip()
-        words = remaining_text.split()
-        for j in range(0, len(words), 8):
-            segments.append(" ".join(words[j:j + 8]))
-    
+
+    # Handle any leftover sentence fragment
+    if len(sentences) % 2 == 1:
+        remaining_text = sentences[-1].strip()
+        segments.append(remaining_text)
+
     return segments
 
 # Function to generate SRT with accurate timing per batch
@@ -127,43 +134,8 @@ async def process_script(script_text, pitch, rate, voice):
 voice_options = {
     "Andrew Male": "en-US-AndrewNeural",
     "Jenny Female": "en-US-JennyNeural",
-    "Guy Male": "en-US-GuyNeural",
-    "Ana Female": "en-US-AnaNeural",
-    "Aria Female": "en-US-AriaNeural",
-    "Brian Male": "en-US-BrianNeural",
-    "Christopher Male": "en-US-ChristopherNeural",
-    "Eric Male": "en-US-EricNeural",
-    "Michelle Male": "en-US-MichelleNeural",
-    "Roger Male": "en-US-RogerNeural",
-    "Natasha Female": "en-AU-NatashaNeural",
-    "William Male": "en-AU-WilliamNeural",
-    "Clara Female": "en-CA-ClaraNeural",
-    "Liam Female ": "en-CA-LiamNeural",
-    "Libby Female": "en-GB-LibbyNeural",
-    "Maisie": "en-GB-MaisieNeural",
-    "Ryan": "en-GB-RyanNeural",
-    "Sonia": "en-GB-SoniaNeural",
-    "Thomas": "en-GB-ThomasNeural",
-    "Sam": "en-HK-SamNeural",
-    "Yan": "en-HK-YanNeural",
-    "Connor": "en-IE-ConnorNeural",
-    "Emily": "en-IE-EmilyNeural",
-    "Neerja": "en-IN-NeerjaNeural",
-    "Prabhat": "en-IN-PrabhatNeural",
-    "Asilia": "en-KE-AsiliaNeural",
-    "Chilemba": "en-KE-ChilembaNeural",
-    "Abeo": "en-NG-AbeoNeural",
-    "Ezinne": "en-NG-EzinneNeural",
-    "Mitchell": "en-NZ-MitchellNeural",
-    "James": "en-PH-JamesNeural",
-    "Rosa": "en-PH-RosaNeural",
-    "Luna": "en-SG-LunaNeural",
-    "Wayne": "en-SG-WayneNeural",
-    "Elimu": "en-TZ-ElimuNeural",
-    "Imani": "en-TZ-ImaniNeural",
-    "Leah": "en-ZA-LeahNeural",
-    "Luke": "en-ZA-LukeNeural"
-}  # All voice options
+    # Add other voices here...
+}
 
 app = gr.Interface(
     fn=process_script,