From dd0d22084613ed303efe086fcad8367504f4585b Mon Sep 17 00:00:00 2001 From: Jeroen Schweitzer Date: Mon, 6 Jul 2026 22:26:16 +0200 Subject: [PATCH] =?UTF-8?q?style(tooling):=20ruff=20fixes=20in=20offset-sh?= =?UTF-8?q?ell=20script=20=E2=80=94=20unused=20import,=20ambiguous=20loop?= =?UTF-8?q?=20var?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Co-Authored-By: Claude Fable 5 --- .pql/changelog/ticket_history/2026-07.sql | 39 +++++++++++++++++++ .pql/changelog/tickets/2026-07.sql | 21 ++++++++++ .../blender_author_offset_shell.py | 3 +- 3 files changed, 61 insertions(+), 2 deletions(-) diff --git a/.pql/changelog/ticket_history/2026-07.sql b/.pql/changelog/ticket_history/2026-07.sql index 154603702..2894c3cc0 100644 --- a/.pql/changelog/ticket_history/2026-07.sql +++ b/.pql/changelog/ticket_history/2026-07.sql @@ -846,3 +846,42 @@ INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, chang RESOLVED (pending review). Root cause: fork bodies (thin/heavy/child) had a mesh<->armature mismatch. blender_segment_body.py apply_scale scaled mesh vertices ONLY, leaving the embedded armature at source scale; since character_visual.gd binds every segment onto ONE shared skeleton (skeleton/armature.glb) by bone name, each segment must be internally consistent with its own embedded armature — the mesh-only scale dropped the head ~0.35m below the Head bone and flung limbs apart on relocation. thin/heavy were additionally stale high-poly artifacts (torso 1419-1704v vs 375v) missing seg_hips entirely. Fix: replaced apply_scale with apply_fork_scale (tooling/blender_segment_body.py) which scales mesh AND armature together and BAKES via Blender transform_apply (edit-bone poking alone shears connected chains — arms/fingers still exploded). Rebuilt all 5 via new tooling/blender_rebuild_forks.py from Source-tier gltf; now 19 segments incl seg_hips, low-poly matching the healthy six. Bare-body QA (idle+walk, front+side, real CharacterVisual): all 5 coherent, head/hair attached, no explosion; healthy 6 untouched. Q-060: all 3 peasant garments Surface-Deform-fit onto all 5 forks (15/15 surface_deform, 0 shrinkwrap fallback) — Surface Deform HOLDS at the extremes on repaired bodies. No manifest change (body ids/paths unchanged).', NULL, '2026-07-06 20:02:30', '2026-07-06 20:02:30.990', '2026-07-06 20:02:30.990', NULL, '7606f8b27d88884fcb7d4a2e3a5663bb', 2) ON CONFLICT(hash) DO NOTHING; INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, changed_by, changed_at, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FKE0EW2617K9VE4QHSYSR9HW', 'status', 'in_progress', 'review', NULL, '2026-07-06 20:02:31', '2026-07-06 20:02:31.013', '2026-07-06 20:02:31.013', NULL, '942bea5d7db30ea972e6ba64341369a1', 2) ON CONFLICT(hash) DO NOTHING; INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, changed_by, changed_at, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FKE0EW2617K9VE4QHSYSR9HW', 'status', 'review', 'done', NULL, '2026-07-06 20:06:22', '2026-07-06 20:06:22.078', '2026-07-06 20:06:22.078', NULL, '10bec531a986c8b28b68229637a1297f', 2) ON CONFLICT(hash) DO NOTHING; +INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, changed_by, changed_at, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FKCPBF41RWK9PW9M22Q7961C', 'description', 'Fable-5-window sidequest workstream 2 (user directive 2026-07-06, extends T-1088 scope family): build the clothing ENGINE and a basic set now, park variety for later. Target bar: RimWorld''s original apparel breadth — varied enough to not feel limiting. Goal experience: being able to pick and SHOP for clothes — colorize + slap logos on casualwear. Basic set: t-shirt (logo-capable), hoodie (logo-capable), button-down shirt, shorts, jeans, formal pants, jacket, colorable uniform, black suit (moved up so officials are includable). Skipped by explicit call: capes, dresses/skirts (parked — difficulty tier). Engine deliverables: (1) multi-region tint masks — requires Quaternius Source tier files (USD5 Patreon month, one-time download, CC0 retained — user purchase action); (2) logo decal region on garment chest UV — flat 2D artwork per D-244, diegetic hook into economy generated brands (wiki/economics/corporations/generated_brands.toml) for brand-logo shirts; (3) manifest + CharacterVisualDescriptor category coverage (slots/tints already generic); (4) batch-fit each authored garment to all 11 body types via the validated Sprint-28 Blender Surface Deform pipeline (tooling/blender_*.py). Authoring routes per garment TBD by feasibility pass: derive from owned Fantasy pack meshes (62 parts, e.g. tunic->tshirt sleeve cut, pants->shorts hem cut) vs hand-author low-poly on the 65-bone rig (Garment Tool). Related: T-1088 locomotion sandbox doubles as the garment-under-animation QA rig (cycle outfits through Walk/Sprint/Crouch_Fwd). + +Correction (2026-07-06): Source tier is PARTIALLY owned already — docs/assets/downloads/ (gitignored vendor archives, present in both checkouts) holds Universal Base Characters[Source] (630MB: full .blend files for all 6 base bodies incl. FullBody + OnlyHead variants, hairstyle .blends origin-at-0 + rigged-to-head-bone, texture masters, Godot engine project) plus UBC[Standard], UAL 1+2[Standard], Modular Character Outfits Fantasy[Standard]. NOT owned: Fantasy Outfits[Source] — the garment recolor masks + named vertex groups. Consequence: authoring NEW garments is fully unblocked (author against source body .blends, create our own tint masks/vertex groups — no purchase needed); only pack-DERIVED garments'' clean masks would want one more USD5 Patreon month, decision deferred until the feasibility route table says how much derivation we actually use. + +Art direction (user, 2026-07-06): garments read CYBERPUNK/MODERN — near-future everyday civilian + corporate wear — NOT fantasy/medieval. The owned Fantasy pack is at most a topology donor (low-poly base shapes like pants/shirts are genre-neutral; genre lives in silhouette details, textures, trims, and accessories) — no medieval trim/textures may survive into shipped garments. Weight the authoring routes accordingly: texture identity matters as much as mesh. Canon hook: the wiki already ships a garment brand — thrds (wiki/star-systems/GJ-475/index.md, Braemar fiber cooperative, corridor fashion markets) — so branded clothing connects to Phase-1 lore, not only generated_brands.toml. This emphasis also feeds the Quaternius-pivot evaluation: commercial alternatives skew sci-fi/urban while Quaternius clothing is fantasy-only. + +Nuance (user, 2026-07-06): the cyberpunk/modern direction is the DEFAULT, not a prohibition — fashion in the Reach is cyclical, so fantasy/medieval-flavored garments may appear here and there as deliberate fashion statements (subculture pieces, luxury/revival trims, statement outerwear). Supersedes the previous note''s hard ''no medieval trim may survive'' line. Practical consequence: a small number of Fantasy-pack pieces may ship near-as-is where they read as intentional fashion; the basic set itself stays modern. + +Art direction CONSOLIDATED (user, 2026-07-06 — supersedes the two notes above): the Reach has no single clothing style — ALL styles coexist, with modern/cyberpunk probably the most prevalent. Style is a prevalence DISTRIBUTION, not a genre rule, and may eventually vary by culture/settlement (fits the simulation DNA; fashion-as-cultural-data is a later phase concern, not this ticket). Consequences now: (1) the engine is style-agnostic by construction (slots/tints/logos carry no genre); (2) the basic set targets the prevalent center — modern casual + corporate — because it dresses the most NPCs and the shopping experience; (3) garments get a lightweight style tag in the manifest (e.g. style: modern|formal|fantasy|utility) as the cheap seam so per-culture distributions can be authored later without retrofitting; (4) Fantasy-pack pieces are legitimate catalogue members under their own tag, prevalence-weighted low. + +Scope pin (user, 2026-07-06, final word on style): for simplicity, THIS ticket ships ONE style only — cyberpunk/modern. The pluralistic distribution model above remains the long-term truth (keep the manifest style tag as the seam), but the v1 basic set is single-style; other styles are catalogue growth, not scope. + +Route challenge (user, 2026-07-06): not sold on staying Quaternius-primary — ''Quaternius is limited and will move the wall at too slow a pace; I''d rather pick something that has what we need now.'' Fresh store verification (2026-07-06): Sidekick packs are one-time USD199.99, perpetual, engine-agnostic license (Godot legal, unsupported); NO raw FBX — extraction from unitypackage (community-proven: unitypackage_godot, third-party Sidekick Creator Godot plugin incl. body blend shapes); ''Sold out'' labels are a broken-theme artifact per Shopify JSON (needs one browser cart check); SyntyPass rejected (assets stop being developable after cancel). Relevant packs: Starter (FREE, 57+91 parts), Modern Civilians (165 parts = v1 basic set territory), Sci-Fi Civilians (103 parts, cyberpunk accents), Modern Police (uniforms/officials). DECISION NOW GATED ON THE FREE SPIKE: user downloads free Starter Pack -> Bridge-C spike (extract garment -> Robust Weight Transfer onto our rig -> 11-body batch -> animation QA in T-1088 sandbox). Pass -> buy Modern Civilians, Synty = primary garment supply (+UAL1/UAL2 Source ~USD30 for clips — no Synty substitute exists for Godot animations; Fantasy Outfits Source drops to low priority under the single-style pin). Fail -> hand-author route stands. Stay-vs-pivot governance record HELD until spike evidence lands. + +Scope guard (user concern, 2026-07-06): Synty intake must NOT displace the create-stuff-yourself pipeline — the user''s goal includes in-house garment creation capability, not just swapping vendor catalogs. Resolution: three-route portfolio sharing one engine — (a) Synty intake = variety bootstrap only; (b) derivation scripting (cuts/merges/retexture, agent-automatable) = create-yourself lite, turns any donor topology into original designs; (c) TRUE AUTHORING on owned UBC Source body .blends = the pipeline proper. GUARANTEE: the colorable uniform and black suit (no donor exists in any pack) are built via route (c) as the authoring pipeline''s end-to-end proof — non-negotiable deliverable of this ticket, so the pipeline exists even if Synty covers everything else. Spike status: extraction done inline by lead (158 FBX, sci-fi civilian theme confirmed); DECISIVE finding — Sidekick rig is UE-Mannequin-named, 52/65 bone-name overlap with our rig, sample garment binds to 12 vgroups of which 11 match ours exactly (only head->Head case rename) => transplant likely rename-and-rebind, far cheaper than the weight-transfer the pivot eval assumed; body blend shapes travel in the FBX. Transplant phase parked pending API recovery / token-window reset. + +ROUTE DECISION (user, 2026-07-06, final): Synty intake REJECTED despite the spike''s technical PASS — (1) USD199.99/pack is too expensive; (2) Sidekick modularity is baked body-segment swap (clothes fused to body chunks), incompatible with our layered clothing-over-body system (11 bodies, skin tones, slot swapping); (3) user dislikes the Synty style. The spike still paid: transplant/batch scripts are source-agnostic and become the G1-family production pipeline; T-1090 (5 bodies broken bare) discovered; the Synty question is closed with evidence, not parked. OPERATIVE ROUTE: (b) derive from owned Fantasy parts + (c) hand-author — with a new preferred authoring technique: OFFSET-SHELL FROM OUR OWN BODY MESHES (extract body segment surface, offset outward into garment shell, cut neckline/hem/sleeves; weights inherited by construction, style fully ours, USD0). Suit + uniform remain the full hand-author proof. Purchases still open (separate concerns): UAL1/UAL2 Source ~USD30 for animation clips; Fantasy Outfits Source USD20 low-priority convenience. + +Chromakey garment-QA harness DELIVERED (2026-07-06, commits eaca6c8c4/74fa16260/b09a50efb): two-pass depth-proximity clip detection — pass A keys covered body segments magenta, pass B re-renders garment-only with a depth-only 3cm epsilon bias toward camera; analyzer splits key pixels into exposed_skin (informational — sleeveless arms, collars) vs clip_through (gating — skin within 3cm in front of cloth). Run: tooling/garment-qa/run-garment-qa . Verified on the peasant set (avg_m/f x Walk/Sprint/Crouch_Fwd x 3 frames x 4 yaws): 0/18 front-view false positives; genuine findings = sleeveless armhole seam (avg_f, sides), deep-crouch waist gap (worst 150px, back), collar nape. First real fit observations already on record for the peasant set. Sensitivity: clip_epsilon_m + --min-pixels, to calibrate against the first modern garments. This is the acceptance gate for every garment the wardrobe pipeline produces; pair with the T-1088 sandbox for in-world walk QA.', 'Fable-5-window sidequest workstream 2 (user directive 2026-07-06, extends T-1088 scope family): build the clothing ENGINE and a basic set now, park variety for later. Target bar: RimWorld''s original apparel breadth — varied enough to not feel limiting. Goal experience: being able to pick and SHOP for clothes — colorize + slap logos on casualwear. Basic set: t-shirt (logo-capable), hoodie (logo-capable), button-down shirt, shorts, jeans, formal pants, jacket, colorable uniform, black suit (moved up so officials are includable). Skipped by explicit call: capes, dresses/skirts (parked — difficulty tier). Engine deliverables: (1) multi-region tint masks — requires Quaternius Source tier files (USD5 Patreon month, one-time download, CC0 retained — user purchase action); (2) logo decal region on garment chest UV — flat 2D artwork per D-244, diegetic hook into economy generated brands (wiki/economics/corporations/generated_brands.toml) for brand-logo shirts; (3) manifest + CharacterVisualDescriptor category coverage (slots/tints already generic); (4) batch-fit each authored garment to all 11 body types via the validated Sprint-28 Blender Surface Deform pipeline (tooling/blender_*.py). Authoring routes per garment TBD by feasibility pass: derive from owned Fantasy pack meshes (62 parts, e.g. tunic->tshirt sleeve cut, pants->shorts hem cut) vs hand-author low-poly on the 65-bone rig (Garment Tool). Related: T-1088 locomotion sandbox doubles as the garment-under-animation QA rig (cycle outfits through Walk/Sprint/Crouch_Fwd). + +Correction (2026-07-06): Source tier is PARTIALLY owned already — docs/assets/downloads/ (gitignored vendor archives, present in both checkouts) holds Universal Base Characters[Source] (630MB: full .blend files for all 6 base bodies incl. FullBody + OnlyHead variants, hairstyle .blends origin-at-0 + rigged-to-head-bone, texture masters, Godot engine project) plus UBC[Standard], UAL 1+2[Standard], Modular Character Outfits Fantasy[Standard]. NOT owned: Fantasy Outfits[Source] — the garment recolor masks + named vertex groups. Consequence: authoring NEW garments is fully unblocked (author against source body .blends, create our own tint masks/vertex groups — no purchase needed); only pack-DERIVED garments'' clean masks would want one more USD5 Patreon month, decision deferred until the feasibility route table says how much derivation we actually use. + +Art direction (user, 2026-07-06): garments read CYBERPUNK/MODERN — near-future everyday civilian + corporate wear — NOT fantasy/medieval. The owned Fantasy pack is at most a topology donor (low-poly base shapes like pants/shirts are genre-neutral; genre lives in silhouette details, textures, trims, and accessories) — no medieval trim/textures may survive into shipped garments. Weight the authoring routes accordingly: texture identity matters as much as mesh. Canon hook: the wiki already ships a garment brand — thrds (wiki/star-systems/GJ-475/index.md, Braemar fiber cooperative, corridor fashion markets) — so branded clothing connects to Phase-1 lore, not only generated_brands.toml. This emphasis also feeds the Quaternius-pivot evaluation: commercial alternatives skew sci-fi/urban while Quaternius clothing is fantasy-only. + +Nuance (user, 2026-07-06): the cyberpunk/modern direction is the DEFAULT, not a prohibition — fashion in the Reach is cyclical, so fantasy/medieval-flavored garments may appear here and there as deliberate fashion statements (subculture pieces, luxury/revival trims, statement outerwear). Supersedes the previous note''s hard ''no medieval trim may survive'' line. Practical consequence: a small number of Fantasy-pack pieces may ship near-as-is where they read as intentional fashion; the basic set itself stays modern. + +Art direction CONSOLIDATED (user, 2026-07-06 — supersedes the two notes above): the Reach has no single clothing style — ALL styles coexist, with modern/cyberpunk probably the most prevalent. Style is a prevalence DISTRIBUTION, not a genre rule, and may eventually vary by culture/settlement (fits the simulation DNA; fashion-as-cultural-data is a later phase concern, not this ticket). Consequences now: (1) the engine is style-agnostic by construction (slots/tints/logos carry no genre); (2) the basic set targets the prevalent center — modern casual + corporate — because it dresses the most NPCs and the shopping experience; (3) garments get a lightweight style tag in the manifest (e.g. style: modern|formal|fantasy|utility) as the cheap seam so per-culture distributions can be authored later without retrofitting; (4) Fantasy-pack pieces are legitimate catalogue members under their own tag, prevalence-weighted low. + +Scope pin (user, 2026-07-06, final word on style): for simplicity, THIS ticket ships ONE style only — cyberpunk/modern. The pluralistic distribution model above remains the long-term truth (keep the manifest style tag as the seam), but the v1 basic set is single-style; other styles are catalogue growth, not scope. + +Route challenge (user, 2026-07-06): not sold on staying Quaternius-primary — ''Quaternius is limited and will move the wall at too slow a pace; I''d rather pick something that has what we need now.'' Fresh store verification (2026-07-06): Sidekick packs are one-time USD199.99, perpetual, engine-agnostic license (Godot legal, unsupported); NO raw FBX — extraction from unitypackage (community-proven: unitypackage_godot, third-party Sidekick Creator Godot plugin incl. body blend shapes); ''Sold out'' labels are a broken-theme artifact per Shopify JSON (needs one browser cart check); SyntyPass rejected (assets stop being developable after cancel). Relevant packs: Starter (FREE, 57+91 parts), Modern Civilians (165 parts = v1 basic set territory), Sci-Fi Civilians (103 parts, cyberpunk accents), Modern Police (uniforms/officials). DECISION NOW GATED ON THE FREE SPIKE: user downloads free Starter Pack -> Bridge-C spike (extract garment -> Robust Weight Transfer onto our rig -> 11-body batch -> animation QA in T-1088 sandbox). Pass -> buy Modern Civilians, Synty = primary garment supply (+UAL1/UAL2 Source ~USD30 for clips — no Synty substitute exists for Godot animations; Fantasy Outfits Source drops to low priority under the single-style pin). Fail -> hand-author route stands. Stay-vs-pivot governance record HELD until spike evidence lands. + +Scope guard (user concern, 2026-07-06): Synty intake must NOT displace the create-stuff-yourself pipeline — the user''s goal includes in-house garment creation capability, not just swapping vendor catalogs. Resolution: three-route portfolio sharing one engine — (a) Synty intake = variety bootstrap only; (b) derivation scripting (cuts/merges/retexture, agent-automatable) = create-yourself lite, turns any donor topology into original designs; (c) TRUE AUTHORING on owned UBC Source body .blends = the pipeline proper. GUARANTEE: the colorable uniform and black suit (no donor exists in any pack) are built via route (c) as the authoring pipeline''s end-to-end proof — non-negotiable deliverable of this ticket, so the pipeline exists even if Synty covers everything else. Spike status: extraction done inline by lead (158 FBX, sci-fi civilian theme confirmed); DECISIVE finding — Sidekick rig is UE-Mannequin-named, 52/65 bone-name overlap with our rig, sample garment binds to 12 vgroups of which 11 match ours exactly (only head->Head case rename) => transplant likely rename-and-rebind, far cheaper than the weight-transfer the pivot eval assumed; body blend shapes travel in the FBX. Transplant phase parked pending API recovery / token-window reset. + +ROUTE DECISION (user, 2026-07-06, final): Synty intake REJECTED despite the spike''s technical PASS — (1) USD199.99/pack is too expensive; (2) Sidekick modularity is baked body-segment swap (clothes fused to body chunks), incompatible with our layered clothing-over-body system (11 bodies, skin tones, slot swapping); (3) user dislikes the Synty style. The spike still paid: transplant/batch scripts are source-agnostic and become the G1-family production pipeline; T-1090 (5 bodies broken bare) discovered; the Synty question is closed with evidence, not parked. OPERATIVE ROUTE: (b) derive from owned Fantasy parts + (c) hand-author — with a new preferred authoring technique: OFFSET-SHELL FROM OUR OWN BODY MESHES (extract body segment surface, offset outward into garment shell, cut neckline/hem/sleeves; weights inherited by construction, style fully ours, USD0). Suit + uniform remain the full hand-author proof. Purchases still open (separate concerns): UAL1/UAL2 Source ~USD30 for animation clips; Fantasy Outfits Source USD20 low-priority convenience. + +Chromakey garment-QA harness DELIVERED (2026-07-06, commits eaca6c8c4/74fa16260/b09a50efb): two-pass depth-proximity clip detection — pass A keys covered body segments magenta, pass B re-renders garment-only with a depth-only 3cm epsilon bias toward camera; analyzer splits key pixels into exposed_skin (informational — sleeveless arms, collars) vs clip_through (gating — skin within 3cm in front of cloth). Run: tooling/garment-qa/run-garment-qa . Verified on the peasant set (avg_m/f x Walk/Sprint/Crouch_Fwd x 3 frames x 4 yaws): 0/18 front-view false positives; genuine findings = sleeveless armhole seam (avg_f, sides), deep-crouch waist gap (worst 150px, back), collar nape. First real fit observations already on record for the peasant set. Sensitivity: clip_epsilon_m + --min-pixels, to calibrate against the first modern garments. This is the acceptance gate for every garment the wardrobe pipeline produces; pair with the T-1088 sandbox for in-world walk QA. + +ENGINE DELIVERED (2026-07-06, commits e22ea0fa4 + 278e8b7c2; earlier eaca6c8c4/74fa16260/b09a50efb chromakey, b54b8189d bodies): G1 skinned batch-fit (blender_batch_fit_skinned.py), offset-shell authoring (blender_author_offset_shell.py — route c create-yourself, weights by construction, procedural RGBA masks, UV2 chest channel), toon_garment.gdshader (channel-blended 4-region tint + logo after-tint/before-toon), compositor plumb (multi_region coverage flag -> garment shader; descriptor clothing_logos: per-character brand choice), thrds logo stub, manifest style tag + logo_capable. PROOF: tshirt_modern on 6 healthy bodies, 18-assertion suite green, 216-capture chromakey QA. ROUTE GUIDANCE from QA evidence (Q-060): single-reference SD-fit of offset-shells degrades with girth divergence (muscular_m worst 859px clip at 24mm standoff; reference body clean) — offset-shell garments should be authored PER BODY (needs per-body masks + runtime _mask.png loading, small change); SD-fit remains right for derived/hand-authored garments sharing one mask (suit, uniform, Fantasy cuts). REMAINING for this ticket: refit tshirt to the 5 repaired forks (T-1090 done — add to --only and rerun), per-body shell mode, then the basic set: hoodie, button-down, shorts, jeans, formal pants, jacket + the suit/uniform hand-author proof; creation-screen try-on UI for tints/logos.', NULL, '2026-07-06 20:22:08', '2026-07-06 20:22:08.752', '2026-07-06 20:22:08.752', NULL, 'b2e49bf9d2fcab1df28108db784e8b8f', 2) ON CONFLICT(hash) DO NOTHING; diff --git a/.pql/changelog/tickets/2026-07.sql b/.pql/changelog/tickets/2026-07.sql index 2525b5b87..7ac670845 100644 --- a/.pql/changelog/tickets/2026-07.sql +++ b/.pql/changelog/tickets/2026-07.sql @@ -722,3 +722,24 @@ INSERT INTO tickets (record_id, type, parent_record_id, title, description, stat INSERT INTO tickets (record_id, type, parent_record_id, title, description, status, priority, assigned_to, team, decision_ref, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FKE0EW2617K9VE4QHSYSR9HW', 'bug', '06FB0TNSRXX29H5X0EB0F0GBW4', '5 of 11 body types misrender bare in CharacterVisual (thin_f/m, heavy_f/m, child)', 'Found by the T-1089 Synty intake spike (2026-07-06), body-only controls with NO garment attached: thin_f/m, heavy_f/m, child misrender in CharacterVisual (exploded/spider arms, detached head/hair; child 303mm max warp, 87 collapsed edges; f-bodies bust-through). These are the newer high-poly bodies (no seg_hips, different segmentation) — exactly the 5 with no production peasant-clothing variants. Upstream of ANY clothing route; blocks T-1089''s 11-of-11 acceptance criterion and resolves part of Q-060. Evidence: spikes/synty-intake/out/bodies_fit_log.json + out/qa_bodies/ (44 screenshots) + out/bodies_recon.json. RESOLVED (pending review). Root cause: fork bodies (thin/heavy/child) had a mesh<->armature mismatch. blender_segment_body.py apply_scale scaled mesh vertices ONLY, leaving the embedded armature at source scale; since character_visual.gd binds every segment onto ONE shared skeleton (skeleton/armature.glb) by bone name, each segment must be internally consistent with its own embedded armature — the mesh-only scale dropped the head ~0.35m below the Head bone and flung limbs apart on relocation. thin/heavy were additionally stale high-poly artifacts (torso 1419-1704v vs 375v) missing seg_hips entirely. Fix: replaced apply_scale with apply_fork_scale (tooling/blender_segment_body.py) which scales mesh AND armature together and BAKES via Blender transform_apply (edit-bone poking alone shears connected chains — arms/fingers still exploded). Rebuilt all 5 via new tooling/blender_rebuild_forks.py from Source-tier gltf; now 19 segments incl seg_hips, low-poly matching the healthy six. Bare-body QA (idle+walk, front+side, real CharacterVisual): all 5 coherent, head/hair attached, no explosion; healthy 6 untouched. Q-060: all 3 peasant garments Surface-Deform-fit onto all 5 forks (15/15 surface_deform, 0 shrinkwrap fallback) — Surface Deform HOLDS at the extremes on repaired bodies. No manifest change (body ids/paths unchanged).', 'done', 'high', NULL, 'client', NULL, '2026-07-06 10:38:16.593', '2026-07-06 20:06:22.078', NULL, 'dd9ef4733867c6166098027e49a44569', 2) ON CONFLICT(record_id) DO UPDATE SET type=excluded.type, parent_record_id=excluded.parent_record_id, title=excluded.title, description=excluded.description, status=excluded.status, priority=excluded.priority, assigned_to=excluded.assigned_to, team=excluded.team, decision_ref=excluded.decision_ref, updated_at=excluded.updated_at, deleted_at=excluded.deleted_at, hash=excluded.hash, canonical_version=excluded.canonical_version WHERE excluded.updated_at > tickets.updated_at OR (excluded.updated_at = tickets.updated_at AND excluded.hash > tickets.hash); +INSERT INTO tickets (record_id, type, parent_record_id, title, description, status, priority, assigned_to, team, decision_ref, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FKCPBF41RWK9PW9M22Q7961C', 'story', '06FB0TNSRXX29H5X0EB0F0GBW4', 'Clothing catalogue engine + basic set — recolorable, logo-capable everyday garments', 'Fable-5-window sidequest workstream 2 (user directive 2026-07-06, extends T-1088 scope family): build the clothing ENGINE and a basic set now, park variety for later. Target bar: RimWorld''s original apparel breadth — varied enough to not feel limiting. Goal experience: being able to pick and SHOP for clothes — colorize + slap logos on casualwear. Basic set: t-shirt (logo-capable), hoodie (logo-capable), button-down shirt, shorts, jeans, formal pants, jacket, colorable uniform, black suit (moved up so officials are includable). Skipped by explicit call: capes, dresses/skirts (parked — difficulty tier). Engine deliverables: (1) multi-region tint masks — requires Quaternius Source tier files (USD5 Patreon month, one-time download, CC0 retained — user purchase action); (2) logo decal region on garment chest UV — flat 2D artwork per D-244, diegetic hook into economy generated brands (wiki/economics/corporations/generated_brands.toml) for brand-logo shirts; (3) manifest + CharacterVisualDescriptor category coverage (slots/tints already generic); (4) batch-fit each authored garment to all 11 body types via the validated Sprint-28 Blender Surface Deform pipeline (tooling/blender_*.py). Authoring routes per garment TBD by feasibility pass: derive from owned Fantasy pack meshes (62 parts, e.g. tunic->tshirt sleeve cut, pants->shorts hem cut) vs hand-author low-poly on the 65-bone rig (Garment Tool). Related: T-1088 locomotion sandbox doubles as the garment-under-animation QA rig (cycle outfits through Walk/Sprint/Crouch_Fwd). + +Correction (2026-07-06): Source tier is PARTIALLY owned already — docs/assets/downloads/ (gitignored vendor archives, present in both checkouts) holds Universal Base Characters[Source] (630MB: full .blend files for all 6 base bodies incl. FullBody + OnlyHead variants, hairstyle .blends origin-at-0 + rigged-to-head-bone, texture masters, Godot engine project) plus UBC[Standard], UAL 1+2[Standard], Modular Character Outfits Fantasy[Standard]. NOT owned: Fantasy Outfits[Source] — the garment recolor masks + named vertex groups. Consequence: authoring NEW garments is fully unblocked (author against source body .blends, create our own tint masks/vertex groups — no purchase needed); only pack-DERIVED garments'' clean masks would want one more USD5 Patreon month, decision deferred until the feasibility route table says how much derivation we actually use. + +Art direction (user, 2026-07-06): garments read CYBERPUNK/MODERN — near-future everyday civilian + corporate wear — NOT fantasy/medieval. The owned Fantasy pack is at most a topology donor (low-poly base shapes like pants/shirts are genre-neutral; genre lives in silhouette details, textures, trims, and accessories) — no medieval trim/textures may survive into shipped garments. Weight the authoring routes accordingly: texture identity matters as much as mesh. Canon hook: the wiki already ships a garment brand — thrds (wiki/star-systems/GJ-475/index.md, Braemar fiber cooperative, corridor fashion markets) — so branded clothing connects to Phase-1 lore, not only generated_brands.toml. This emphasis also feeds the Quaternius-pivot evaluation: commercial alternatives skew sci-fi/urban while Quaternius clothing is fantasy-only. + +Nuance (user, 2026-07-06): the cyberpunk/modern direction is the DEFAULT, not a prohibition — fashion in the Reach is cyclical, so fantasy/medieval-flavored garments may appear here and there as deliberate fashion statements (subculture pieces, luxury/revival trims, statement outerwear). Supersedes the previous note''s hard ''no medieval trim may survive'' line. Practical consequence: a small number of Fantasy-pack pieces may ship near-as-is where they read as intentional fashion; the basic set itself stays modern. + +Art direction CONSOLIDATED (user, 2026-07-06 — supersedes the two notes above): the Reach has no single clothing style — ALL styles coexist, with modern/cyberpunk probably the most prevalent. Style is a prevalence DISTRIBUTION, not a genre rule, and may eventually vary by culture/settlement (fits the simulation DNA; fashion-as-cultural-data is a later phase concern, not this ticket). Consequences now: (1) the engine is style-agnostic by construction (slots/tints/logos carry no genre); (2) the basic set targets the prevalent center — modern casual + corporate — because it dresses the most NPCs and the shopping experience; (3) garments get a lightweight style tag in the manifest (e.g. style: modern|formal|fantasy|utility) as the cheap seam so per-culture distributions can be authored later without retrofitting; (4) Fantasy-pack pieces are legitimate catalogue members under their own tag, prevalence-weighted low. + +Scope pin (user, 2026-07-06, final word on style): for simplicity, THIS ticket ships ONE style only — cyberpunk/modern. The pluralistic distribution model above remains the long-term truth (keep the manifest style tag as the seam), but the v1 basic set is single-style; other styles are catalogue growth, not scope. + +Route challenge (user, 2026-07-06): not sold on staying Quaternius-primary — ''Quaternius is limited and will move the wall at too slow a pace; I''d rather pick something that has what we need now.'' Fresh store verification (2026-07-06): Sidekick packs are one-time USD199.99, perpetual, engine-agnostic license (Godot legal, unsupported); NO raw FBX — extraction from unitypackage (community-proven: unitypackage_godot, third-party Sidekick Creator Godot plugin incl. body blend shapes); ''Sold out'' labels are a broken-theme artifact per Shopify JSON (needs one browser cart check); SyntyPass rejected (assets stop being developable after cancel). Relevant packs: Starter (FREE, 57+91 parts), Modern Civilians (165 parts = v1 basic set territory), Sci-Fi Civilians (103 parts, cyberpunk accents), Modern Police (uniforms/officials). DECISION NOW GATED ON THE FREE SPIKE: user downloads free Starter Pack -> Bridge-C spike (extract garment -> Robust Weight Transfer onto our rig -> 11-body batch -> animation QA in T-1088 sandbox). Pass -> buy Modern Civilians, Synty = primary garment supply (+UAL1/UAL2 Source ~USD30 for clips — no Synty substitute exists for Godot animations; Fantasy Outfits Source drops to low priority under the single-style pin). Fail -> hand-author route stands. Stay-vs-pivot governance record HELD until spike evidence lands. + +Scope guard (user concern, 2026-07-06): Synty intake must NOT displace the create-stuff-yourself pipeline — the user''s goal includes in-house garment creation capability, not just swapping vendor catalogs. Resolution: three-route portfolio sharing one engine — (a) Synty intake = variety bootstrap only; (b) derivation scripting (cuts/merges/retexture, agent-automatable) = create-yourself lite, turns any donor topology into original designs; (c) TRUE AUTHORING on owned UBC Source body .blends = the pipeline proper. GUARANTEE: the colorable uniform and black suit (no donor exists in any pack) are built via route (c) as the authoring pipeline''s end-to-end proof — non-negotiable deliverable of this ticket, so the pipeline exists even if Synty covers everything else. Spike status: extraction done inline by lead (158 FBX, sci-fi civilian theme confirmed); DECISIVE finding — Sidekick rig is UE-Mannequin-named, 52/65 bone-name overlap with our rig, sample garment binds to 12 vgroups of which 11 match ours exactly (only head->Head case rename) => transplant likely rename-and-rebind, far cheaper than the weight-transfer the pivot eval assumed; body blend shapes travel in the FBX. Transplant phase parked pending API recovery / token-window reset. + +ROUTE DECISION (user, 2026-07-06, final): Synty intake REJECTED despite the spike''s technical PASS — (1) USD199.99/pack is too expensive; (2) Sidekick modularity is baked body-segment swap (clothes fused to body chunks), incompatible with our layered clothing-over-body system (11 bodies, skin tones, slot swapping); (3) user dislikes the Synty style. The spike still paid: transplant/batch scripts are source-agnostic and become the G1-family production pipeline; T-1090 (5 bodies broken bare) discovered; the Synty question is closed with evidence, not parked. OPERATIVE ROUTE: (b) derive from owned Fantasy parts + (c) hand-author — with a new preferred authoring technique: OFFSET-SHELL FROM OUR OWN BODY MESHES (extract body segment surface, offset outward into garment shell, cut neckline/hem/sleeves; weights inherited by construction, style fully ours, USD0). Suit + uniform remain the full hand-author proof. Purchases still open (separate concerns): UAL1/UAL2 Source ~USD30 for animation clips; Fantasy Outfits Source USD20 low-priority convenience. + +Chromakey garment-QA harness DELIVERED (2026-07-06, commits eaca6c8c4/74fa16260/b09a50efb): two-pass depth-proximity clip detection — pass A keys covered body segments magenta, pass B re-renders garment-only with a depth-only 3cm epsilon bias toward camera; analyzer splits key pixels into exposed_skin (informational — sleeveless arms, collars) vs clip_through (gating — skin within 3cm in front of cloth). Run: tooling/garment-qa/run-garment-qa . Verified on the peasant set (avg_m/f x Walk/Sprint/Crouch_Fwd x 3 frames x 4 yaws): 0/18 front-view false positives; genuine findings = sleeveless armhole seam (avg_f, sides), deep-crouch waist gap (worst 150px, back), collar nape. First real fit observations already on record for the peasant set. Sensitivity: clip_epsilon_m + --min-pixels, to calibrate against the first modern garments. This is the acceptance gate for every garment the wardrobe pipeline produces; pair with the T-1088 sandbox for in-world walk QA. + +ENGINE DELIVERED (2026-07-06, commits e22ea0fa4 + 278e8b7c2; earlier eaca6c8c4/74fa16260/b09a50efb chromakey, b54b8189d bodies): G1 skinned batch-fit (blender_batch_fit_skinned.py), offset-shell authoring (blender_author_offset_shell.py — route c create-yourself, weights by construction, procedural RGBA masks, UV2 chest channel), toon_garment.gdshader (channel-blended 4-region tint + logo after-tint/before-toon), compositor plumb (multi_region coverage flag -> garment shader; descriptor clothing_logos: per-character brand choice), thrds logo stub, manifest style tag + logo_capable. PROOF: tshirt_modern on 6 healthy bodies, 18-assertion suite green, 216-capture chromakey QA. ROUTE GUIDANCE from QA evidence (Q-060): single-reference SD-fit of offset-shells degrades with girth divergence (muscular_m worst 859px clip at 24mm standoff; reference body clean) — offset-shell garments should be authored PER BODY (needs per-body masks + runtime _mask.png loading, small change); SD-fit remains right for derived/hand-authored garments sharing one mask (suit, uniform, Fantasy cuts). REMAINING for this ticket: refit tshirt to the 5 repaired forks (T-1090 done — add to --only and rerun), per-body shell mode, then the basic set: hoodie, button-down, shorts, jeans, formal pants, jacket + the suit/uniform hand-author proof; creation-screen try-on UI for tints/logos.', 'backlog', 'high', NULL, 'client', NULL, '2026-07-06 07:34:18.656', '2026-07-06 20:22:08.752', NULL, 'a8211523c2bef72893601e640205216e', 2) ON CONFLICT(record_id) DO UPDATE SET type=excluded.type, parent_record_id=excluded.parent_record_id, title=excluded.title, description=excluded.description, status=excluded.status, priority=excluded.priority, assigned_to=excluded.assigned_to, team=excluded.team, decision_ref=excluded.decision_ref, updated_at=excluded.updated_at, deleted_at=excluded.deleted_at, hash=excluded.hash, canonical_version=excluded.canonical_version WHERE excluded.updated_at > tickets.updated_at OR (excluded.updated_at = tickets.updated_at AND excluded.hash > tickets.hash); diff --git a/tooling/garment-fit/blender_author_offset_shell.py b/tooling/garment-fit/blender_author_offset_shell.py index fe04a76c3..1a04291cc 100644 --- a/tooling/garment-fit/blender_author_offset_shell.py +++ b/tooling/garment-fit/blender_author_offset_shell.py @@ -48,7 +48,6 @@ import os import bpy import bmesh import numpy as np -from mathutils import Vector # -------------------------------------------------------------------------- # Parameters @@ -304,7 +303,7 @@ def _classify_region(center): def _tris_from_face(face, uv_layer): """Fan-triangulate a bmesh face into (uv, uv, uv) tuples in [0,1] space.""" loops = face.loops[:] - uvs = [l[uv_layer].uv.copy() for l in loops] + uvs = [loop[uv_layer].uv.copy() for loop in loops] tris = [] for i in range(1, len(uvs) - 1): tris.append((uvs[0], uvs[i], uvs[i + 1]))