import json
import sys
import subprocess

sys.stdout.reconfigure(encoding='utf-8')

# Run git diff between 02544795 and HEAD for work_item_master_2026-01-01.json
cmd = ["git", "diff", "02544795", "HEAD", "--", "resources/data_work_item_master/work_item_master_2026-01-01.json"]
res = subprocess.run(cmd, capture_output=True, text=True, encoding='utf-8')

diff_lines = res.stdout.splitlines()
print(f"Total diff lines in work_item_master: {len(diff_lines)}")

# Parse what tables changed
# Look for pum_table_id context
changed_tables = {}
cur_tid = None
cur_changes = []

for l in diff_lines:
    if '"pum_table_id":' in l:
        m = l.split('"pum_table_id":')[1].strip().replace('"', '').replace(',', '')
        cur_tid = m
        if cur_tid not in changed_tables:
            changed_tables[cur_tid] = []
    if (l.startswith('+') or l.startswith('-')) and not l.startswith('+++') and not l.startswith('---'):
        if cur_tid:
            changed_tables[cur_tid].append(l)

print(f"Number of tables with diffs: {len(changed_tables)}")
for tid, chs in list(changed_tables.items())[:15]:
    print(f"\n[{tid}] changes ({len(chs)} lines):")
    for c in chs[:4]:
        print(f"   {c}")