breakpilot-compliance/scripts/qa/apply_pdf_qa_results.py

"""Apply PDF QA results: update source_citation with correct article + article_type."""
import os
import json
import psycopg2
import urllib.parse

RESULTS_FILE = "/tmp/pdf_qa_results.json"

# Load results
with open(RESULTS_FILE) as f:
    results = json.load(f)
print(f"Loaded {len(results)} results")

# DB connection
db_url = os.environ['DATABASE_URL']
parsed = urllib.parse.urlparse(db_url)
conn = psycopg2.connect(
    host=parsed.hostname, port=parsed.port or 5432,
    user=parsed.username, password=parsed.password,
    dbname=parsed.path.lstrip('/'),
    options="-c search_path=compliance,public"
)

# Update in batches
cur = conn.cursor()
updated = 0
errors = 0
unchanged = 0

for i, r in enumerate(results):
    ctrl_id = r["ctrl_id"]
    article_label = r["article_label"]
    article_type = r["article_type"]  # preamble, article, annex, section, unknown

    try:
        # Update source_citation: set article and article_type
        cur.execute("""
            UPDATE compliance.canonical_controls
            SET source_citation = source_citation
                || jsonb_build_object('article', %s, 'article_type', %s),
                updated_at = now()
            WHERE id = %s::uuid
            AND (
                source_citation->>'article' IS DISTINCT FROM %s
                OR source_citation->>'article_type' IS DISTINCT FROM %s
            )
        """, (article_label, article_type, ctrl_id, article_label, article_type))

        if cur.rowcount > 0:
            updated += 1
        else:
            unchanged += 1

    except Exception as e:
        errors += 1
        if errors <= 5:
            print(f"  ERROR {ctrl_id}: {str(e)[:100]}")
        conn.rollback()
        continue

    if (i + 1) % 500 == 0:
        conn.commit()
        print(f"  Progress: {i+1}/{len(results)} (updated: {updated}, unchanged: {unchanged}, errors: {errors})")

conn.commit()
print(f"\nDone: {updated} updated, {unchanged} unchanged, {errors} errors out of {len(results)}")

# Verify: count by article_type
cur.execute("""
    SELECT source_citation->>'article_type' as art_type, count(*)
    FROM compliance.canonical_controls
    WHERE source_citation->>'article_type' IS NOT NULL
    GROUP BY 1
    ORDER BY count(*) DESC
""")
print("\nArticle type distribution in DB:")
for row in cur.fetchall():
    print(f"  {str(row[0]):12s}: {row[1]:5d}")

# Verify: sample preamble controls
cur.execute("""
    SELECT control_id, source_citation->>'article', source_citation->>'article_type',
           source_citation->>'source'
    FROM compliance.canonical_controls
    WHERE source_citation->>'article_type' = 'preamble'
    LIMIT 5
""")
print("\nSample preamble controls:")
for row in cur.fetchall():
    print(f"  {row[0]}: {row[1]} ({row[2]}) — {row[3][:40]}")

# Verify: sample annex controls
cur.execute("""
    SELECT control_id, source_citation->>'article', source_citation->>'article_type',
           source_citation->>'source'
    FROM compliance.canonical_controls
    WHERE source_citation->>'article_type' = 'annex'
    LIMIT 5
""")
print("\nSample annex controls:")
for row in cur.fetchall():
    print(f"  {row[0]}: {row[1]} ({row[2]}) — {row[3][:40]}")

conn.close()