Update app.py

app.py

@@ -27,26 +27,26 @@ async def text_to_speech(text, voice, rate, pitch):
     return tmp_path, None
 
 # Generate SRT file based on user preferences
-def generate_srt(paragraphs, audio_duration, srt_path, words_per_line, lines_per_paragraph):
-    total_paragraphs = len(paragraphs)
+def generate_srt(words, audio_duration, srt_path, words_per_line, lines_per_paragraph):
+    total_words = len(words)
     
     # Calculate how long each segment will be displayed
-    segment_duration = audio_duration / total_paragraphs  # Total audio duration divided by total paragraphs
+    segment_duration = audio_duration / (total_words // words_per_line // lines_per_paragraph)  # Calculate duration based on total segments
     
     current_time = 0
     with open(srt_path, 'w', encoding='utf-8') as srt_file:
-        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]
+        for i in range(0, total_words, words_per_line):
+            # Gather lines based on the defined words per line
+            lines = words[i:i + words_per_line]
+            line_text = ' '.join(lines)
 
             start_time = current_time
             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 + 1}\n{start_time_str} --> {end_time_str}\n")
-            srt_file.write('\n'.join(lines) + '\n\n')
+            srt_file.write(f"{(i // words_per_line) + 1}\n{start_time_str} --> {end_time_str}\n")
+            srt_file.write(f"{line_text}\n\n")
 
             current_time += segment_duration  # Update current time for the next segment
 
@@ -63,7 +63,10 @@ def format_srt_time(seconds):
 
 # Text to audio and SRT functionality
 async def text_to_audio_and_srt(text, voice, rate, pitch, words_per_line, lines_per_paragraph):
-    audio_path, warning = await text_to_speech(text, voice, rate, pitch)
+    # Clean up input text: remove extra spaces and newlines
+    cleaned_text = ' '.join(text.split())
+
+    audio_path, warning = await text_to_speech(cleaned_text, voice, rate, pitch)
     if warning:
         return None, None, warning
 
@@ -74,10 +77,10 @@ async def text_to_audio_and_srt(text, voice, rate, pitch, words_per_line, lines_
     base_name = os.path.splitext(audio_path)[0]
     srt_path = f"{base_name}_subtitle.srt"
     
-    # 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()]
+    # Split input text into words
+    words = cleaned_text.split()
     
-    generate_srt(paragraphs, audio_duration, srt_path, words_per_line, lines_per_paragraph)
+    generate_srt(words, audio_duration, srt_path, words_per_line, lines_per_paragraph)
 
     return audio_path, srt_path, None