Update app.py

app.py

@@ -115,7 +115,24 @@ def processFile(filePath):
     totalTimeInSeconds = int(waveform_gain_adjusted.shape[-1]/sampleRate)
     print("Time in seconds calculated")
     return annotations, totalTimeInSeconds
-
+    
+@st.cache_data
+def extract_speaker_clip(file_path, speaker, _annotation, clip_duration=5.0):
+    """Extract a short audio clip for a speaker from their first segment."""
+    waveform, sample_rate = torchaudio.load(file_path)
+    for segment, _, label in _annotation.itertracks(yield_label=True):
+        if label == speaker and segment.duration >= 1.0:
+            start_sample = int(segment.start * sample_rate)
+            end_sample = int(min(segment.start + clip_duration, segment.end) * sample_rate)
+            clip = waveform[:, start_sample:end_sample]
+            # Convert to wav bytes in memory
+            import io
+            buf = io.BytesIO()
+            torchaudio.save(buf, clip, sample_rate, format="wav")
+            buf.seek(0)
+            return buf.read()
+    return None
+    
 def addCategory():
     newCategory = st.session_state.categoryInput
     st.toast(f"Adding {newCategory}")
@@ -608,12 +625,19 @@ try:
 
         st.sidebar.divider()
         st.sidebar.subheader("Rename Speakers")
-        st.sidebar.caption("Replace SPEAKER_## labels with real names.")
+        st.sidebar.caption("Replace SPEAKER_## labels with real names. Listen to a sample clip to help identify each speaker.")
         current_renames = st.session_state.speakerRenames[currFileIndex]
+        currFilePath = st.session_state.file_paths[currFileIndex]
         for sp in speakerNames:
             current_label = current_renames.get(sp, "")
+            st.sidebar.markdown(f"**{sp}**")
+            # Only show audio player for real audio files (not rttm/csv/txt)
+            if currFilePath.lower().endswith(('.wav', '.mp3', '.mp4')):
+                clip_bytes = extract_speaker_clip(currFilePath, sp, currAnnotation)
+                if clip_bytes:
+                    st.sidebar.audio(clip_bytes, format="audio/wav")
             new_name = st.sidebar.text_input(
-                f"{sp}",
+                f"Rename {sp}",
                 value=current_label,
                 placeholder=f"e.g. John",
                 key=f"rename_{currFileIndex}_{sp}"