"""Oracle soft-cap archival. Bounds live `entries` growth by moving old / excess rows into `entries_archive` (preserving data — soft cap, not hard delete). """ import os import sqlite3 import time from oracle.config import DB_PATH ARCHIVE_SCHEMA = """ CREATE TABLE IF NOT EXISTS entries_archive ( id INTEGER PRIMARY KEY, source TEXT, source_id TEXT, url TEXT, title TEXT, extracted_text TEXT, summary TEXT, category_tags TEXT, signal_score REAL, raw_metadata TEXT, first_seen TEXT, last_updated TEXT, archived_at TEXT DEFAULT (strftime('%Y-%m-%dT%H:%M:%SZ', 'now')) ); """ def archive(days=30, cap=5000, dry_run=False, db_path=None): """Archive entries older than N days or beyond the cap. Returns (total_live, archived_count). """ path = db_path or str(DB_PATH) if not os.path.exists(path): print("No oracle.db — nothing to archive") return 0, 0 conn = sqlite3.connect(path) conn.execute(ARCHIVE_SCHEMA) cutoff = time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime(time.time() - days * 86400)) cur = conn.cursor() cur.execute("SELECT id, last_updated FROM entries") rows = cur.fetchall() n_total = len(rows) old_ids = [r[0] for r in rows if (r[1] or "") < cutoff] beyond = max(0, n_total - cap) if beyond > 0: ordered = sorted(rows, key=lambda r: r[1] or "")[:beyond] cap_ids = [r[0] for r in ordered] else: cap_ids = [] move_ids = sorted(set(old_ids) | set(cap_ids)) if not move_ids: print(f"Archive check: {n_total} live entries, none older than {days}d " f"or beyond cap {cap}. Nothing to archive.") conn.close() return n_total, 0 print(f"Archive check: {n_total} live entries -> would archive {len(move_ids)} " f"(old={len(old_ids)}, cap={len(cap_ids)}).") if dry_run: print("DRY RUN — no changes made.") conn.close() return n_total, 0 q = ",".join("?" * len(move_ids)) conn.execute( f"""INSERT OR REPLACE INTO entries_archive (id, source, source_id, url, title, extracted_text, summary, category_tags, signal_score, raw_metadata, first_seen, last_updated) SELECT id, source, source_id, url, title, extracted_text, summary, category_tags, signal_score, raw_metadata, first_seen, last_updated FROM entries WHERE id IN ({q})""", move_ids, ) conn.execute(f"DELETE FROM entries WHERE id IN ({q})", move_ids) conn.commit() conn.close() print(f"Archived {len(move_ids)} entries (live now {n_total - len(move_ids)}). " f"Preserved in entries_archive.") return n_total, len(move_ids)