Dropfleet data changed (2026-08-24)

[github-actions]

The Dropfleet data needs a look.

TTCombat published something (1)

DFC downloads scan: 28 files on page, 28 in baseline.

  REVISED  Resistance_Combined_Fleet_Stats_260731.pdf (re-uploaded; date stamp unchanged)
           https://cdn.shopify.com/s/files/1/0965/1274/files/Resistance_Combined_Fleet_Stats_260731.pdf?v=1787325876

1 change(s). Review, download updated PDFs into Rules-Mechanics-PDFs/, and integrate.

---

Integrating an edition: download the PDFs into Rules-Mechanics-PDFs/, diff them
page by page against the previous edition, update data/faction-*.json, and check
whether any new rule changes what is LEGAL to take. If it does, wire it into both
js/app.js and mobile/js/mobile.js, not just the card text. Then rebaseline with
python scripts/scan-dfc-files.py --update.

A REVISED entry means the same date stamp was re-uploaded. Usually cosmetic, but
not always: on 2026-08-02 a re-upload renamed a ship, so diff the text before
assuming nothing changed.

Run log

[github-actions]

The Dropfleet data needs a look.

TTCombat published something (1)

DFC downloads scan: 28 files on page, 28 in baseline.

  RESIZED  UCM_Combined_Fleet_Stats_260828.pdf (overwritten in place: 20083776 -> 20083785 bytes)
           https://cdn.shopify.com/s/files/1/0965/1274/files/UCM_Combined_Fleet_Stats_260828.pdf?v=1787908175

1 change(s). Review, download updated PDFs into Rules-Mechanics-PDFs/, and integrate.

BSData cross-check (1)

Either a ship whose number or rule differs from the community catalogue,
or one of theirs we have no counterpart for. Read the block: it says which.

The PDF settles a difference. If we turn out to be right, add the ship to
ADJUDICATED in scripts/compare-bsdata.py with the page reference so it
stops reappearing. An unmatched ship is usually a rename on one side or the
other, and often theirs, since TTCombat re-uploads reach us first.

WARN could not fetch Industrial.json: HTTP Error 404: Not Found
Cross-check vs BSData/dropfleet-commander-2024: 343 ships matched.

No new disagreements on points, stats or rules.

1 BSData ship(s) with no counterpart here (new ships, or naming that needs a manual look):
  Frances Mendoza - Flying Dutchman  (136 pts, United Colonies of Mankind.json)

The cards and our data disagree (1)

This one is about us, not TTCombat. Either a ship drifted from its card,
or a card was re-cut in place under an unchanged URL.

Open the PDF and read the block. If the data is wrong, fix it. If the parser
is wrong, confirm it against the card and move the line into accepted in
scripts/card-audit-baseline.json — that bucket is a claim someone read it.

warning: The `fitz` API is deprecated and will be removed in future. Use `import pymupdf` instead.
  bioficer    Bioficer_Combined_Fleet_Stats_260529.pdf       0 disagreement(s), 0 new
  -           Misc_Combined_Ship_Stats_250822.pdf (hash only)
  phr         PHR_Combined_Fleet_Stats_260828.pdf            0 disagreement(s), 0 new
  resistance  Resistance_Combined_Fleet_Stats_260731.pdf     2 disagreement(s), 0 new
  scourge     Scourge_Combined_Fleet_Stats_260626.pdf        2 disagreement(s), 0 new
  shaltari    Shaltari_Combined_Fleet_Stats_260731.pdf       0 disagreement(s), 0 new
  ucm         UCM_Combined_Fleet_Stats_260828.pdf            1 disagreement(s), 1 new

The cards and the data disagree in 1 new place(s):
  ucm
      Frances Mendoza - Flying Dutchman      on the card, no ship in the data matches it

The card settles it. Open the PDF, read the ship's block, and fix whichever
side is wrong. If the data is right and the parser is not, accept it with
`python scripts/audit-cards.py --update` so it stops reappearing.
Baseline: 4 verified against the card, 1 never read.
Those unread lines are parser noise as far as anyone knows, which is not
the same as knowing. `--verbose` lists them; the card settles each one.

data/fleet-data.json was hand-edited

It is generated. Regenerating it changed it, which means an edit went into
the mirror instead of into the faction-*.json file the app actually loads —
so the edit is not live, and the mirror now disagrees with the app.

Put the change in the faction file, then run python scripts/gen-fleet-data.py.

data/fleet-data.json | 276 +++++++++++++++++++++++++++++++++++++++------------
 1 file changed, 213 insertions(+), 63 deletions(-)

---

Integrating an edition: download the PDFs into Rules-Mechanics-PDFs/, diff them
page by page against the previous edition, update data/faction-*.json, and check
whether any new rule changes what is LEGAL to take. If it does, wire it into both
js/app.js and mobile/js/mobile.js, not just the card text. Then rebaseline with
python scripts/scan-dfc-files.py --update.

A REVISED entry means the same date stamp was re-uploaded. Usually cosmetic, but
not always: on 2026-08-02 a re-upload renamed a ship, so diff the text before
assuming nothing changed.

Run log

[github-actions]

The Dropfleet data needs a look.

TTCombat published something (1)

DFC downloads scan: 28 files on page, 28 in baseline.

  RESIZED  UCM_Combined_Fleet_Stats_260828.pdf (overwritten in place: 20083776 -> 20083785 bytes)
           https://cdn.shopify.com/s/files/1/0965/1274/files/UCM_Combined_Fleet_Stats_260828.pdf?v=1787908175

1 change(s). Review, download updated PDFs into Rules-Mechanics-PDFs/, and integrate.

The cards and our data disagree (1)

This one is about us, not TTCombat. Either a ship drifted from its card,
or a card was re-cut in place under an unchanged URL.

Open the PDF and read the block. If the data is wrong, fix it. If the parser
is wrong, confirm it against the card and move the line into accepted in
scripts/card-audit-baseline.json — that bucket is a claim someone read it.

warning: The `fitz` API is deprecated and will be removed in future. Use `import pymupdf` instead.
  bioficer    Bioficer_Combined_Fleet_Stats_260529.pdf       0 disagreement(s), 0 new
  -           Misc_Combined_Ship_Stats_250822.pdf (hash only)
  phr         PHR_Combined_Fleet_Stats_260828.pdf            0 disagreement(s), 0 new
  resistance  Resistance_Combined_Fleet_Stats_260731.pdf     2 disagreement(s), 0 new
  scourge     Scourge_Combined_Fleet_Stats_260626.pdf        2 disagreement(s), 0 new
  shaltari    Shaltari_Combined_Fleet_Stats_260731.pdf       0 disagreement(s), 0 new
  ucm         UCM_Combined_Fleet_Stats_260828.pdf            1 disagreement(s), 1 new

The cards and the data disagree in 1 new place(s):
  ucm
      Frances Mendoza - Flying Dutchman      on the card, no ship in the data matches it

The card settles it. Open the PDF, read the ship's block, and fix whichever
side is wrong. If the data is right and the parser is not, accept it with
`python scripts/audit-cards.py --update` so it stops reappearing.
Baseline: 4 verified against the card, 1 never read.
Those unread lines are parser noise as far as anyone knows, which is not
the same as knowing. `--verbose` lists them; the card settles each one.

---

Integrating an edition: download the PDFs into Rules-Mechanics-PDFs/, diff them
page by page against the previous edition, update data/faction-*.json, and check
whether any new rule changes what is LEGAL to take. If it does, wire it into both
js/app.js and mobile/js/mobile.js, not just the card text. Then rebaseline with
python scripts/scan-dfc-files.py --update.

A REVISED entry means the same date stamp was re-uploaded. Usually cosmetic, but
not always: on 2026-08-02 a re-upload renamed a ship, so diff the text before
assuming nothing changed.

Run log