sync: auto-sync from GURU-BEAST-ROG at 2026-04-29 13:29:17

Author: Mike Swanson
Machine: GURU-BEAST-ROG
Timestamp: 2026-04-29 13:29:17

projects/radio-show/audio-processor/import_to_sqlite.py

@@ -87,10 +87,15 @@ CREATE TABLE IF NOT EXISTS qa_pairs (
     caller_name TEXT,
     caller_role TEXT,
     topic TEXT,
-    topic_tags TEXT
+    topic_tags TEXT,
+    usefulness_score INTEGER,
+    topic_class TEXT,
+    is_banter INTEGER
 );
 CREATE INDEX IF NOT EXISTS idx_qa_episode ON qa_pairs(episode_id);
 CREATE INDEX IF NOT EXISTS idx_qa_caller ON qa_pairs(caller_name);
+-- Indexes on quality columns (usefulness_score, topic_class) are created by
+-- _migrate_qa_quality_columns() so they apply to both fresh and migrated DBs.
 
 CREATE VIRTUAL TABLE IF NOT EXISTS segments_fts USING fts5(
     text,
@@ -133,9 +138,42 @@ def init_schema(conn: sqlite3.Connection):
     conn.executescript(SCHEMA)
     conn.executescript(TRIGGERS)
     conn.execute("PRAGMA foreign_keys = ON")
+    _migrate_qa_quality_columns(conn)
     conn.commit()
 
 
+# Columns added by the Q&A quality classifier (Track 1).
+# Defined here so the migration is idempotent and runs on every invocation:
+# the SCHEMA above creates them on a fresh DB; this block ALTERs an existing
+# qa_pairs that pre-dates them. Names + types must match SCHEMA exactly.
+QA_QUALITY_COLUMNS = (
+    ("usefulness_score", "INTEGER"),
+    ("topic_class",      "TEXT"),
+    ("is_banter",        "INTEGER"),
+)
+
+
+def _migrate_qa_quality_columns(conn: sqlite3.Connection) -> None:
+    """Add Q&A quality columns to qa_pairs if they're missing.
+
+    Idempotent: existing column values are untouched; new columns default NULL.
+    Safe to call on fresh DBs (qa_pairs already has the columns from SCHEMA -
+    PRAGMA table_info reflects that and we no-op).
+    """
+    existing = {row[1] for row in conn.execute("PRAGMA table_info(qa_pairs)").fetchall()}
+    if not existing:
+        # qa_pairs hasn't been created yet (shouldn't happen post-SCHEMA, but be safe)
+        return
+    for col, col_type in QA_QUALITY_COLUMNS:
+        if col not in existing:
+            conn.execute(f"ALTER TABLE qa_pairs ADD COLUMN {col} {col_type}")
+    # Indexes on the new columns (CREATE INDEX IF NOT EXISTS is already idempotent,
+    # but on a brand-new DB they were created by SCHEMA; on a migrated DB they
+    # weren't, so create them here too).
+    conn.execute("CREATE INDEX IF NOT EXISTS idx_qa_usefulness ON qa_pairs(usefulness_score)")
+    conn.execute("CREATE INDEX IF NOT EXISTS idx_qa_topic_class ON qa_pairs(topic_class)")
+
+
 def sha256_file(path: Path) -> str:
     h = hashlib.sha256()
     with open(path, "rb") as f: