=== Part 1: click depth from the home page ===
URL                        depth   inlinks  verdict
/                              0         1  ok
/layanan                       1         1  ok
/harga                         1         2  ok
/cabang                        1         1  ok
/cabang/keputih                2         1  ok
/cabang/gebang                 2         1  ok
/tentang                       1         1  ok
/kontak                        1         3  ok
/blog                          1         1  ok
/blog/merawat-kain             2         2  ok
/blog/tips-setrika             2         2  ok
/promo-lama                  inf         0  ORPHAN - unreachable from /
pages: 12, orphans: 1, deeper than 2 clicks: 0

=== Part 2: near-duplicate clusters (threshold 30) ===
  minimum spanning tree over 6 pages, 15 measured pairs:
    /harga            /harga-2024           4  keep (same cluster)
    /cabang/keputih   /cabang/gebang        6  keep (same cluster)
    /cabang/gebang    /cabang/rungkut       7  keep (same cluster)
    /harga            /layanan             38  CUT
    /cabang/rungkut   /layanan             52  CUT
  tree weight 107; edges 5 = 6 - 1
  cluster 1: /cabang/keputih /cabang/gebang /cabang/rungkut   <- pick one canonical URL
  cluster 2: /harga /harga-2024   <- pick one canonical URL
  cluster 3: /layanan
  clusters: 3 (cutting the 2 heaviest tree edges)

=== Part 3: the remediation plan, topologically sorted ===
  step  task                 hours  earliest finish
     1  audit-inventory         2               2
     2  fix-robots.txt          1               3
     3  add-canonicals          3               5
     4  rewrite-thin-pages      8              10
     5  merge-duplicates        4               9
     6  add-schema.org          5              15
     7  submit-sitemap          1              10
     8  re-scan                 1              16
  one worker, tasks end to end: 25 h
  unlimited workers, critical path: 16 h  (audit-inventory -> rewrite-thin-pages -> add-schema.org -> re-scan)
