Add application file

app.py

@@ -37,9 +37,6 @@ except Exception as e:
     st.error(f"DuckDB connection failed: {e}")
     st.stop()
 
-# Enable full-text search index on first run (one-time setup if not exists)
-con.execute("PRAGMA create_fts_index('data', 'text')")
-
 # Search
 query = st.text_input("Search text (case-insensitive)", "").strip()
 
@@ -47,10 +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 text % ?
+        WHERE LOWER(text) LIKE LOWER(?)
         LIMIT 100
     """
-    df = con.execute(sql, [query]).df()
+    df = con.execute(sql, [f"%{query}%"]).df()
 else:
     df = con.execute("""
         SELECT id, channel, video_id, video_title, speaker, start_time, end_time, text, pos_tags, upload_date, audio
@@ -79,26 +76,17 @@ else:
             return None
 
     df["audio_file"] = df["audio"].apply(render_audio_cell)
-    df_display = df.drop(columns=["audio"]).copy()
-
-    # Add HTML audio tag column
-    def audio_html(path):
-        if path:
-            return f'<audio controls preload="none" style="height:20px;"> <source src="file://{path}" type="audio/mpeg"> </audio>'
-        return ""
-
-    df_display["Audio"] = df["audio_file"].apply(audio_html)
 
-    # Reorder columns
-    column_order = ["id", "channel", "video_id", "video_title", "speaker", "start_time", "end_time", "upload_date", "text", "pos_tags", "Audio"]
-    df_display = df_display[column_order]
-
-    st.markdown("### Full Table View (Sortable)")
-    st.write("Note: Audio is embedded using HTML tags; not all browsers allow playback from local temp paths.")
-    st.dataframe(df_display.drop(columns=["Audio"]))
+    # Display table (sortable)
+    df_display = df[["id", "channel", "video_id", "video_title", "speaker", "start_time", "end_time", "upload_date", "text", "pos_tags"]].copy()
+    st.dataframe(df_display, use_container_width=True)
 
+    # Audio previews column (aligned separately)
     st.markdown("### Audio Previews")
-    for i, row in df_display.iterrows():
-        if row["Audio"]:
-            st.markdown(f"**{row['speaker']} | {row['text'][:80]}**", unsafe_allow_html=True)
-            st.markdown(row["Audio"], unsafe_allow_html=True)
+    for i, row in df.iterrows():
+        audio_path = row["audio_file"]
+        if audio_path:
+            st.audio(audio_path, format="audio/mp3")
+        else:
+            st.warning("Missing or unreadable audio.")
+