Leonard
82 lines
2.5 KiB
Python
82 lines
2.5 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
Write batch summaries back to oracle.db.
|
|
|
|
Reads /tmp/athena_summarize_batch.json (array of entry dicts with 'summary' key added by sub-agent),
|
|
writes each summary JSON to entries.summary column.
|
|
|
|
Usage:
|
|
python3 write_summaries.py /tmp/athena_summarize_batch_result.json
|
|
"""
|
|
import json
|
|
import sqlite3
|
|
import sys
|
|
|
|
def main():
|
|
if len(sys.argv) < 2:
|
|
print("Usage: python3 write_summaries.py <result_json_file>", file=sys.stderr)
|
|
sys.exit(1)
|
|
|
|
result_path = sys.argv[1]
|
|
try:
|
|
with open(result_path) as f:
|
|
results = json.load(f)
|
|
except (FileNotFoundError, json.JSONDecodeError) as e:
|
|
print(f"Error reading {result_path}: {e}", file=sys.stderr)
|
|
sys.exit(1)
|
|
|
|
db_path = '/home/vpsadmin/oracle/oracle.db'
|
|
conn = sqlite3.connect(db_path)
|
|
cur = conn.cursor()
|
|
|
|
written = 0
|
|
skipped = 0
|
|
errors = 0
|
|
|
|
for item in results:
|
|
eid = item.get('id')
|
|
summary = item.get('summary')
|
|
if not eid or not summary:
|
|
skipped += 1
|
|
continue
|
|
|
|
# Validate summary has expected keys
|
|
if not all(k in summary for k in ('one_liner', 'key_technical_point', 'potential_use_case', 'confidence')):
|
|
print(f" ⚠ ID {eid}: missing required keys, skipping", file=sys.stderr)
|
|
skipped += 1
|
|
continue
|
|
|
|
# Quality gate: reject low-confidence or generic summaries
|
|
ol = summary.get('one_liner', '')
|
|
if len(ol) < 20:
|
|
print(f" ⚠ ID {eid}: one_liner too short ({len(ol)} chars), skipping", file=sys.stderr)
|
|
skipped += 1
|
|
continue
|
|
if any(generic in ol.lower() for generic in ('this article discusses', 'this paper presents', 'see full')):
|
|
print(f" ⚠ ID {eid}: generic one_liner, skipping", file=sys.stderr)
|
|
skipped += 1
|
|
continue
|
|
|
|
try:
|
|
cur.execute("UPDATE entries SET summary = ? WHERE id = ?",
|
|
(json.dumps(summary), eid))
|
|
written += 1
|
|
print(f" ✓ ID {eid}: {ol[:70]}...")
|
|
except Exception as e:
|
|
print(f" ✗ ID {eid}: {e}", file=sys.stderr)
|
|
errors += 1
|
|
|
|
conn.commit()
|
|
|
|
# Verify
|
|
cur.execute("SELECT COUNT(*) FROM entries WHERE summary IS NOT NULL")
|
|
total = cur.fetchone()[0]
|
|
conn.close()
|
|
|
|
print(f"\nResults: {written} written, {skipped} skipped, {errors} errors")
|
|
print(f"Total entries with summary: {total}")
|
|
return 0 if errors == 0 else 1
|
|
|
|
if __name__ == "__main__":
|
|
sys.exit(main())
|