Updated offsets

app.py

@@ -25,9 +25,19 @@ def get_syllables_per_second(audio_file):
         transcription = processor.batch_decode(predicted_ids, output_char_offsets=True)
         offsets = transcription['char_offsets']
 
-    audio_duration = len(audio_input) / sample_rate
-    syllable_count = sum(1 for item in offsets[0] if item['char'] in ['p', 't', 'k'])
-    syllables_per_second = syllable_count / audio_duration
+    # Find the start and end time offsets of the syllables
+    syllable_offsets = [item for item in offsets[0] if item['char'] in ['p', 't', 'k']]
+    
+    if syllable_offsets:  # if any syllable is found
+        first_syllable_offset = syllable_offsets[0]['start_offset'] / sample_rate
+        last_syllable_offset = syllable_offsets[-1]['end_offset'] / sample_rate
+        # Duration from the first to the last syllable
+        syllable_duration = last_syllable_offset - first_syllable_offset
+    else:
+        syllable_duration = 0
+
+    syllable_count = len(syllable_offsets)
+    syllables_per_second = syllable_count / syllable_duration if syllable_duration > 0 else 0
 
     return syllables_per_second
 
@@ -36,4 +46,4 @@ uploaded_file = st.file_uploader("Choose an audio file", type=["wav"])
 if uploaded_file is not None:
     with st.spinner("Processing the audio file..."):
         result = get_syllables_per_second(uploaded_file)
-        st.write("Syllables per second: ", result)
+        st.write("Syllables per second: ", result)