Update app.py

app.py

@@ -20,26 +20,25 @@ 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
+# Updated function to split text into segments by punctuation or limit to 7-8 words without splitting words
 def split_text_into_segments(text):
     segments = []
-    raw_segments = re.split(r'([.!?])', text)
+    raw_segments = re.split(r'([.!?])', text)  # Split by punctuation with the delimiter preserved
+    combined_segments = []
+
+    # Combine text with punctuation back into full sentences
     for i in range(0, len(raw_segments) - 1, 2):
         sentence = raw_segments[i].strip() + raw_segments[i + 1]
+        combined_segments.append(sentence.strip())
+
+    # Further split sentences into 7-8 word segments without splitting words
+    for sentence in combined_segments:
         words = sentence.split()
-        
-        if len(words) > 8:
-            for j in range(0, len(words), 8):
-                segments.append(" ".join(words[j:j + 8]))
-        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]))
-    
+        while words:
+            segment = " ".join(words[:8])  # Take up to 8 words
+            segments.append(segment)
+            words = words[8:]  # Move to the next batch of words
+
     return segments
 
 # Function to generate SRT with accurate timing per batch
@@ -63,6 +62,7 @@ async def generate_accurate_srt(batch_text, batch_num, start_offset, pitch, rate
     for index, segment in enumerate(segments):
         end_time = start_time + segment_duration
         
+        # Ensure the end time does not exceed the total audio length
         if end_time > start_offset + actual_length:
             end_time = start_offset + actual_length
 
@@ -92,6 +92,7 @@ async def batch_process_srt_and_audio(script_text, pitch, rate, voice, progress=
         os.remove(audio_file)
         progress((batch_num + 1) / len(batches))
 
+    # Adjust the total length of the audio for the final cut-off
     total_audio_length = combined_audio.duration_seconds
     validated_srt_content = ""
     for line in all_srt_content.strip().splitlines():
@@ -99,8 +100,11 @@ async def batch_process_srt_and_audio(script_text, pitch, rate, voice, progress=
             start_str, end_str = line.split(' --> ')
             start_time = sum(x * float(t) for x, t in zip([3600, 60, 1, 0.001], start_str.replace(',', ':').split(':')))
             end_time = sum(x * float(t) for x, t in zip([3600, 60, 1, 0.001], end_str.replace(',', ':').split(':')))
+            
+            # Correct end time to ensure it does not exceed the total audio length
             if end_time > total_audio_length:
                 end_time = total_audio_length
+
             line = f"{format_time(start_time)} --> {format_time(end_time)}"
         validated_srt_content += line + "\n"
 
@@ -179,8 +183,7 @@ app = gr.Interface(
         gr.Audio(label="Audio Playback")
     ],
     title="HIVEcorp Text-to-Speech with SRT Generation",
-    description="Convert your script into audio and generate subtitles.",
-    theme="compact",
+    description="Convert your script into speech and generate synchronized subtitles (SRT)."
 )
 
-app.launch()
+app.launch(share=True)