Add application file

app.py

@@ -4,7 +4,6 @@ import streamlit as st
 from huggingface_hub import hf_hub_download
 import pandas as pd
 import tempfile
-import re
 
 HF_REPO_ID = "stcoats/temp-duckdb-upload"
 HF_FILENAME = "ycsep.duckdb"
@@ -37,20 +36,37 @@ except Exception as e:
     st.error(f"DuckDB connection failed: {e}")
     st.stop()
 
-# Search
+# Search input
 query = st.text_input("Search text (case-insensitive)", "").strip()
 
+# Render audio inline in a column
+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 ""
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp:
+            tmp.write(data)
+            tmp.flush()
+            return f'<audio controls style="height:20px; width:100%"> <source src="file://{tmp.name}" type="audio/mpeg"></audio>'
+    except Exception:
+        return ""
+
+# Fetch data from DuckDB
 if query:
     sql = """
-        SELECT id, channel, video_id, video_title, speaker, start_time, end_time, text, pos_tags, upload_date, audio
+        SELECT id, channel, video_id, video_title, speaker, start_time, end_time, upload_date, text, pos_tags, audio
         FROM data
-        WHERE LOWER(text) LIKE LOWER(?)
+        WHERE LOWER(text) LIKE ?
         LIMIT 100
     """
-    df = con.execute(sql, [f"%{query}%"]).df()
+    df = con.execute(sql, [f"%{query.lower()}%"]).df()
 else:
     df = con.execute("""
-        SELECT id, channel, video_id, video_title, speaker, start_time, end_time, text, pos_tags, upload_date, audio
+        SELECT id, channel, video_id, video_title, speaker, start_time, end_time, upload_date, text, pos_tags, audio
         FROM data
         LIMIT 100
     """).df()
@@ -60,33 +76,11 @@ st.markdown(f"### Showing {len(df)} results")
 if len(df) == 0:
     st.warning("No matches found.")
 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)
+    df["Audio"] = df["audio"].apply(render_audio_cell)
+    df_display = df[["id", "channel", "video_id", "video_title", "speaker", "start_time", "end_time", "upload_date", "text", "pos_tags", "Audio"]].copy()
 
-    # 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.iterrows():
-        audio_path = row["audio_file"]
-        if audio_path:
-            st.audio(audio_path, format="audio/mp3")
-        else:
-            st.warning("Missing or unreadable audio.")
+    # Render table with inline audio column (HTML support)
+    st.markdown("### Results Table (Sortable with Audio Column)")
+    st.write("(Scroll right to view audio controls)")
+    st.write(df_display.to_html(escape=False, index=False), unsafe_allow_html=True)