Update app.py

app.py

@@ -27,27 +27,28 @@ async def text_to_speech(text, voice, rate, pitch):
     return tmp_path, None
 
 # Generate SRT file based on user preferences
-def generate_srt(words, audio_duration, srt_path, words_per_line, lines_per_paragraph, speech_rate):
-    total_words = len(words)
-    average_word_duration = (60 / (150 + speech_rate))  # Average duration of each word in seconds, based on typical speech rates
-    segment_duration = average_word_duration * words_per_line  # Duration for each line
+def generate_srt(paragraphs, audio_duration, srt_path, words_per_line, lines_per_paragraph):
+    total_paragraphs = len(paragraphs)
+    
+    # Calculate how long each segment will be displayed
+    segment_duration = audio_duration / total_paragraphs  # Total audio duration divided by total paragraphs
+    
     current_time = 0
-
     with open(srt_path, 'w', encoding='utf-8') as srt_file:
-        for i in range(0, total_words, words_per_line * lines_per_paragraph):  # Increment by lines per paragraph
-            segment_words = words[i:i + (words_per_line * lines_per_paragraph)]
-            lines = [segment_words[j:j + words_per_line] for j in range(0, len(segment_words), words_per_line)]
+        for i, paragraph in enumerate(paragraphs):
+            words = paragraph.split()
+            lines = [words[j:j + words_per_line] for j in range(0, len(words), words_per_line)]
             lines = [' '.join(line) for line in lines]
 
             start_time = current_time
-            end_time = min(start_time + segment_duration * lines_per_paragraph, audio_duration)  # Ensure it doesn't exceed audio duration
+            end_time = min(start_time + segment_duration, audio_duration)  # Ensure it doesn't exceed audio duration
             
             start_time_str = format_srt_time(start_time)
             end_time_str = format_srt_time(end_time)
-            srt_file.write(f"{i // (words_per_line * lines_per_paragraph) + 1}\n{start_time_str} --> {end_time_str}\n")
+            srt_file.write(f"{i + 1}\n{start_time_str} --> {end_time_str}\n")
             srt_file.write('\n'.join(lines) + '\n\n')
 
-            current_time += segment_duration * lines_per_paragraph  # Update current time for the next segment
+            current_time += segment_duration  # Update current time for the next segment
 
     return srt_path
 
@@ -72,8 +73,11 @@ async def text_to_audio_and_srt(text, voice, rate, pitch, words_per_line, lines_
     # Generate SRT file based on the entire text
     base_name = os.path.splitext(audio_path)[0]
     srt_path = f"{base_name}_subtitle.srt"
-    words = text.split()
-    generate_srt(words, audio_duration, srt_path, words_per_line, lines_per_paragraph, rate)
+    
+    # Split input text into paragraphs based on larger gaps (two consecutive newlines)
+    paragraphs = [p.strip() for p in text.split('\n\n') if p.strip()]
+    
+    generate_srt(paragraphs, audio_duration, srt_path, words_per_line, lines_per_paragraph)
 
     return audio_path, srt_path, None