Audited against the Mindhola GAD. Every new field exists because the 2D diagram or the component model needs it; nothing is captured "just in case".
bridge_pier_type.cable_stayedbridge_inventory. Set once at creation.| Field | Type | Values | Drives |
|---|---|---|---|
| bridge_pier_type | select | + cable_stayed (5th option) | Which SVG partial renders · which component-seeding branch runs |
pile_details[i]| Field | Type | Shown when | Drives in the 2D diagram |
|---|
super_structure_details[i]| Field | Type | Shown when | Drives |
|---|
| Drawn element | Source field(s) | Status |
|---|
The renderer spread cables across adjacentSpanPx × 0.45 — an arbitrary fraction — while the form was already capturing cable_anchor_spacing and ignoring it. Now the first anchor sits 1.5 × spacing from the pylon and the rest step out at exactly the entered spacing.
Result: at 6.5 m spacing with 11 cables, a main-span fan reaches 74.7 m into a 150 m span — the two cantilever arms meet exactly at midspan, as the construction method requires. That correctness fell out of using the real number.
Leg separation was a fixed 26–30 px per shape. The legs stand at the deck edges, so it must come from total_width_of_slab — already captured. Now derived, so a 19.5 m deck and a 12 m deck draw visibly different pylons.
The GAD legend defines EJ = EXPANSION JOINT and marks P1 and P6 as EJ piers — they bracket the cable-stayed section. The form only had a bridge-level no_of_expansion_joints count with no way to say where. Added pile_details[i][has_expansion_joint], and the diagram now marks them on the deck.
This matters beyond drawing: an EJ is where deck continuity breaks, so it defines the boundaries of each continuous deck unit.
Location keys are unchanged from today. What changes is which components are offered at each key, derived from the two JSON columns above.
| Component | Location key | Offered where | Item numbers |
|---|
The drawing's SPECIAL NOTE reads: "levels, RL, number of piles and dimensions of structural elements are tentative and subject to change based on detail design." Meanwhile every field in Carriage Way Details is set-once and greyed out on edit. If a bridge is registered from the GAD and detail design later changes the pile counts or pylon heights, some of it cannot be corrected.
Recommendation: allow Carriage Way Details to stay editable until the bridge has its first progress entry, then lock. That keeps the existing protection (no changing structure mid-construction) without blocking the normal GAD → detail-design revision that R&B expects.
One drawing settles three unknowns at once: the pylon shape (we've assumed H-frame), the real pylon height, and the cable count per fan. The GAD we have is plan + longitudinal elevation only, which cannot show any of them.
Published practice puts stay spacing at roughly 6–12 m for concrete decks and 10–20 m for longer steel spans; we've assumed 6.5 m, which is plausible for a 150 m RCC/PSC span but is a guess until confirmed.