7f86104c74
feat(m2): V64 K2 obstacle seeds - all 10 halls now obstacle-aware
...
Owner directive 2026-07-15: extend obstacle recognition to K2 (halls 6-10).
- tools/cad/extract_obstacle_seed_k2.py: same 1m-cell raster + row-strip
method as K1, plus hall-frame rotation transform (+-11..12deg wings)
- 31 measured strips: edge doors, service-room walls, and the full-width
midline door strips on H9/H10 = the A/B divider shutter tracks - the 1.0m
door buffer keeps the shutter travel line clear (operationally correct)
- Layers: ELEVATOR, STAIR, DOOR, gypsum/lightweight-steel/cement-panel walls
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 23:45:14 +09:00
ffa133aee5
feat(m2): V63 wall/stair/elevator/door obstacle recognition + avoidance
...
Owner directive 2026-07-15: recognize walls, stairs, elevators, doors, drywall.
- tools/cad/extract_obstacle_seed.py: rasterize hall floor into 1m cells over
obstacle layers (WALL/DRYWALL/masonry/panel, STAIR, ELEVATOR, DOOR), connected
components, row-strip decomposition (prevents bbox inflation of L-shaped
thin wall runs) -> 60 measured rects for K1 halls 1-5
- V63: hall_obstacle table (kind: stair|elevator|door|wall) + seeds
- Engine: auto-layout reserves obstacles with kind-based buffer (door 1.0m
approach clearance, others 0.3m)
- Canvas: structure layer renders obstacle rects (door tinted green);
HallInfo.obstacles contract extension (additive)
- Limits recorded: wall strips not filled room interiors (perimeter margin +
wall strips still block booth intrusion); K2 deferred (different layer set)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 23:38:29 +09:00
6ff692c03f
feat(m2): V62 measured pillar seed + auto-layout pillar avoidance
...
Owner directive 2026-07-15: extract pillars and make auto-placement avoid them.
- tools/cad/extract_pillar_seed.py: K1 pillars from CAD layer 4-COL-H
(cross-validated against official drawing dims 54/117m south wall pillars);
K2 H7-H10 from official drawing dimension annotations (one interior midline
pillar each, dia 2.3m); H6 not seeded (no source - recorded as limitation)
- V62: hall_pillar table + 74 seeds; also corrects V61 trench y-axis
(RT points were seeded in CAD-y without the top-of-drawing flip - now 63-y,
generator fixed as the canonical form)
- Engine: autoGenerate reserves pillar radius + 0.6m buffer, packing skips
overlapping cells; HallInfo.pillars contract extension (additive, NON_NULL)
- Canvas: structure layer renders pillar circles (editor passes hall pillars)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 23:09:11 +09:00
6be091aa7c
feat(m2): V61 measured trench seed for K1 halls 1-5 from CAD
...
Owner directive 2026-07-15: seed real trench coordinates.
- tools/cad/extract_trench_seed.py: deterministic generator - extracts trench
channels from CAD layer 5-tranchi (spelling differs from expected), merges
paired channel edges into centerlines, emits supply points every 6m
- Measured layout (identical across halls 1-5, cross-validating band alignment):
7 horizontal channels per hall, 9.0m pitch, y=5.85..59.8m local,
extent x[20.6..170.2]m
- V61: delete assumed grid points for H1-H5, insert 875 measured points
(is_assumed=false); supply flags keep V5 policy; K2 stays assumed (no trench
layer in its plan drawing - power trunk drawing analysis is follow-up)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 06:45:32 +09:00
832ab7f8a9
feat(m2): K2 halls 6-10 DWG underlays - DWG tab complete for all 10 halls
...
Owner directive 2026-07-15: render exhibition center 2 and complete the DWG tab.
- fast_render.py: --rotate/--center options (align tilted halls axis-straight);
slim-DXF workflow notes (K2 plan 106MB DXF, basement xref/detail layers dropped)
- Hall geometry measured programmatically (long-line angle clusters) and verified
visually: H6 -12deg 60x93, H7/H8 axis-aligned 90x126 stacked (block exactly
252m = 2x126, floor width exactly 90m), H9 -11deg / H10 +11deg 99x132
- public/media/floorplans/dwg/hall6..10.png: floor+3m crops, calibration stays
deterministic (same margin formula as K1)
- hallFloorplan.ts: DWG_HALL_DIMS extended to all 10 halls (rotation noted)
- docs/analysis/cad-extraction.md: K2 section with confirmed coordinates table
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 02:32:58 +09:00
d8ef5c53df
feat(m2): DWG/JPG floorplan underlay tabs + CAD extraction pipeline
...
Owner directives 2026-07-14~15: run the DWG pipeline; split floorplan into DWG and JPG tabs.
- tools/cad/dwg_extract.py: DWG(LibreDWG dwg2dxf) -> DXF -> layer inventory /
keyword extraction / grid-label anchor detection (ezdxf, recover fallback)
- tools/cad/render_hall_png.py: render hall-band crops from the K1 combined
plan+trench drawing (x0=293.5m, width 171m, Y bands 80.1m + 63m pitch,
single load -> multi-crop). Real trench layout confirmed: 7 horizontal runs
at ~9m pitch + central feeder (differs from assumed 6m grid seed - V61 later)
- public/media/floorplans/dwg/hall1..5.png: K1 hall underlays (3m margin crop,
deterministic calibration from render params)
- Editor left panel new tool group with JPG/DWG tabs; DWG disabled for
halls without CAD render yet (K2 pending). Canvas switches url+region.
- docs/analysis/cad-extraction.md: pipeline findings, limits (A3002 LibreDWG
crash, band-to-hall numbering direction to confirm), next steps
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 00:35:12 +09:00
37aeebdd28
솔루션명 수정
2026-07-13 04:23:07 +09:00
GUARDiA
6294627070
fix(schedule,ui): holidays on exhibition calendar, bilingual brand, neutral calendar borders
2026-07-12 19:16:48 +09:00
