Add application file

app.py

@@ -4,6 +4,7 @@ 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"
@@ -36,37 +37,21 @@ except Exception as e:
     st.error(f"DuckDB connection failed: {e}")
     st.stop()
 
-# Search input
-query = st.text_input("Search text (case-insensitive)", "").strip()
+# Search
+query = st.text_input("Search text (case-insensitive)", "").strip().lower()
 
-# 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
+# Perform search
 if query:
     sql = """
-        SELECT id, channel, video_id, video_title, speaker, start_time, end_time, upload_date, text, pos_tags, audio
+        SELECT id, channel, video_id, speaker, start_time, end_time, upload_date, text, pos_tags, audio
         FROM data
         WHERE LOWER(text) LIKE ?
         LIMIT 100
     """
-    df = con.execute(sql, [f"%{query.lower()}%"]).df()
+    df = con.execute(sql, [f"%{query}%"]).df()
 else:
     df = con.execute("""
-        SELECT id, channel, video_id, video_title, speaker, start_time, end_time, upload_date, text, pos_tags, audio
+        SELECT id, channel, video_id, speaker, start_time, end_time, upload_date, text, pos_tags, audio
         FROM data
         LIMIT 100
     """).df()
@@ -76,11 +61,26 @@ 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 ""
+            with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp:
+                tmp.write(data)
+                tmp.flush()
+                return f'<audio controls style="height:20px;"> <source src="file://{tmp.name}" type="audio/mpeg"></audio>'
+        except Exception:
+            return ""
+
     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()
+    df_display = df[["id", "channel", "video_id", "speaker", "start_time", "end_time", "upload_date", "text", "pos_tags", "Audio"]].copy()
 
-    # Render table with inline audio column (HTML support)
+    # Adjust column widths to make room for audio
     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)
+    st.markdown("(Scroll right to view audio controls)")
+    st.dataframe(df_display.drop(columns=["audio"]))