Add application file

app.py

@@ -44,10 +44,10 @@ if query:
     sql = """
         SELECT id, channel, video_id, video_title, speaker, start_time, end_time, text, pos_tags, upload_date, audio
         FROM data
-        WHERE LOWER(text) LIKE LOWER(?)
+        WHERE LOWER(text) LIKE '%' || LOWER(?) || '%'
         LIMIT 100
     """
-    df = con.execute(sql, [f"%{query}%"]).df()
+    df = con.execute(sql, [query]).df()
 else:
     df = con.execute("""
         SELECT id, channel, video_id, video_title, speaker, start_time, end_time, text, pos_tags, upload_date, audio
@@ -59,43 +59,44 @@ st.markdown(f"### Showing {len(df)} results")
 
 if len(df) == 0:
     st.warning("No matches found.")
-
-# Show table with inline audio players
-for i, row in df.iterrows():
-    col1, col2, col3 = st.columns([3, 5, 2])
-
-    col1.markdown(f"**ID:** {row['id']}")
-    col1.markdown(f"**Channel:** {row['channel']}")
-    col1.markdown(f"**Video ID:** {row['video_id']}")
-    col1.markdown(f"**Video Title:** {row['video_title']}")
-    col1.markdown(f"**Speaker:** {row['speaker']}")
-    col1.markdown(f"**Start Time:** {row['start_time']}")
-    col1.markdown(f"**End Time:** {row['end_time']}")
-    col1.markdown(f"**Upload Date:** {row['upload_date']}")
-
-    highlighted_text = row['text']
-    if query:
-        highlighted_text = re.sub(f'({re.escape(query)})', r'<mark>\1</mark>', highlighted_text, flags=re.IGNORECASE)
-
-    col2.markdown(f"**Text:** {highlighted_text}", unsafe_allow_html=True)
-    col2.markdown(f"**POS tags:** {row['pos_tags']}")
-
-    audio_data = row["audio"]
-    try:
-        if isinstance(audio_data, (bytes, bytearray, memoryview)):
-            audio_bytes = bytes(audio_data)
-        elif isinstance(audio_data, list):  # DuckDB sometimes gives list[int]
-            audio_bytes = bytes(audio_data)
-        else:
-            audio_bytes = None
-
-        if audio_bytes:
-            with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmpfile:
-                tmpfile.write(audio_bytes)
-                tmpfile.flush()
-                col3.audio(tmpfile.name, format="audio/mp3")
-        else:
-            col3.warning("Audio missing or invalid format.")
-    except Exception as e:
-        col3.error(f"Audio error: {e}")
-
+else:
+    def render_audio_cell(audio_bytes):
+        try:
+            if isinstance(audio_bytes, (bytes, bytearray, memoryview)):
+                data = bytes(audio_bytes)
+            elif isinstance(audio_bytes, list):
+                data = bytes(audio_bytes)
+            else:
+                return None
+            with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp:
+                tmp.write(data)
+                tmp.flush()
+                return tmp.name
+        except Exception:
+            return None
+
+    df["audio_file"] = df["audio"].apply(render_audio_cell)
+
+    # Build an interactive sortable table
+    st.markdown("### Results Table (Sortable)")
+    for i, row in df.iterrows():
+        with st.expander(f"? {row['speaker']} | {row['text'][:60]}..."):
+            col1, col2 = st.columns([2, 3])
+
+            with col1:
+                st.write(f"**ID:** {row['id']}")
+                st.write(f"**Channel:** {row['channel']}")
+                st.write(f"**Video ID:** {row['video_id']}")
+                st.write(f"**Video Title:** {row['video_title']}")
+                st.write(f"**Speaker:** {row['speaker']}")
+                st.write(f"**Start Time:** {row['start_time']}")
+                st.write(f"**End Time:** {row['end_time']}")
+                st.write(f"**Upload Date:** {row['upload_date']}")
+                st.write(f"**POS Tags:** {row['pos_tags']}")
+
+            with col2:
+                st.markdown(f"**Text:** {row['text']}")
+                if row['audio_file']:
+                    st.audio(row['audio_file'], format="audio/mp3")
+                else:
+                    st.warning("Audio not available or invalid format.")