zfff

app.py

@@ -7,34 +7,23 @@ from gradio.outputs import Video
 from video_generator import generate_video, extract_last_frame
 
 def extract_lyrics(api_response):
+    words_timing = api_response["results"]["channels"][0]["alternatives"][0]["words"]
     lyrics_with_timing = []
     CHUNK_DURATION = 10
     current_chunk = ""
     current_chunk_start_time = 0
-    current_chunk_end_time = 0
-    
-    try:
-        words_timing = api_response["results"]["channels"][0]["alternatives"][0]["words"]
-        for word_info in words_timing:
-            word = word_info["word"]
-            start_time = word_info["start"]
+    for word_info in words_timing:
+        word = word_info["word"]
+        start_time = word_info["start"]
+        if start_time >= current_chunk_start_time + CHUNK_DURATION:
             end_time = word_info["end"]
+            lyrics_with_timing.append((current_chunk_start_time, end_time, current_chunk.strip()))
+            current_chunk = ""
+            current_chunk_start_time += CHUNK_DURATION
+        current_chunk += " " + word
+    lyrics_with_timing.append((current_chunk_start_time, words_timing[-1]["end"], current_chunk.strip()))
+    return lyrics_with_timing
 
-            if start_time >= current_chunk_start_time + CHUNK_DURATION:
-                lyrics_with_timing.append((current_chunk_start_time, current_chunk_end_time, current_chunk.strip()))
-                current_chunk = ""
-                current_chunk_start_time = start_time
-
-            current_chunk += " " + word
-            current_chunk_end_time = end_time
-
-        # Adding the last chunk
-        lyrics_with_timing.append((current_chunk_start_time, current_chunk_end_time, current_chunk.strip()))
-
-        return lyrics_with_timing
-    except KeyError:
-        print("Error in API response:", api_response)
-        return []
 
 def send_to_deepgram(audio_file_path):
     # Update with your Deepgram API endpoint and key
@@ -76,7 +65,8 @@ def analyze_audio(audio_file_path):
         prompt = generate_video_prompt(previous_mood, current_mood, next_mood, tempo, chroma_mean, spectral_contrast_mean, zero_crossing_rate_mean, mfcc_mean, lyrics_summary)
         description = f"Chunk starting at {start / sr} seconds:<br>Mood: {current_mood}<br>Video Prompt: {prompt}<br><br>"
         print(f"Generating video for chunk {i + 1}...")
-        video = generate_video(prompt, last_frame)
+        lyrics_with_timing = extract_lyrics(deepgram_response)
+        video = generate_video(lyrics_with_timing, last_frame)
         last_frame = extract_last_frame(video)
         print(f"Description for chunk {i + 1}: {description}")
         print(f"Video for chunk {i + 1}: {video}")