@rig-c/parts 1.57.1 → 1.59.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/docs/AUTHORING.md CHANGED
@@ -47,7 +47,17 @@ with `flattenRig(built)`, not with the built file.
47
47
 
48
48
  The painting is yours to bring: a PNG of one character, front-facing, full body,
49
49
  taller than wide, on a plain light background. How it was made is not read by any
50
- stage (`config.generation` is recorded for the optional adapter only).
50
+ stage (`config.generation` is recorded for the optional adapter only). rig-parts
51
+ ships no weights and generates no images itself. The optional `comfy` adapter
52
+ drives models on the user's own ComfyUI host. The default OpenPose ControlNet
53
+ and upscaler sources, pinned model cards and declared terms are in
54
+ [NOTICE](../NOTICE.md#models-named-by-the-optional-painting-graph).
55
+ `comfy seethrough --offload` selects the wrapper's CUDA group offload on that
56
+ host; omit it for a local Mac/MPS host. A Mac CLI may drive a remote CUDA host.
57
+ This offload limitation does not establish MPS compatibility for the rest of the
58
+ model stack. The wrapper README and the adapter graph use different node names
59
+ and depth-model identifiers; [the adapter notes](../README.md#the-optional-comfyui-adapter)
60
+ name the graph actually sent.
51
61
 
52
62
  ## 2. The folder
53
63
 
@@ -108,9 +118,9 @@ read by no CPU stage).
108
118
  | `assemble.extend_below_crop` | assemble | **proposed** with the plan: `{part, run, tag}` — a head-run part that reaches the bottom of the head crop is continued from a full-run layer, whole connected components |
109
119
  | `assemble.patches` | assemble; `meshes`/`regions` keys may name them too | **authored**, optional: `[{name, box, alpha, draw}]` — an extra part cut from the **painting** itself, for figure no See-through layer holds (a hem both runs dropped). `box` is `[x0, y0, x1, y1]` in **rig** pixels, `x1`/`y1` exclusive — the space of `parts.json` and `recomposite_rig.png`, where the hole is found; `alpha` is `"silhouette"` (the painting's figure silhouette inside the box, the one `--seam silhouette` uses) or `"box"` (the whole box); `draw` is `"back"`, `"front"` or `{"before": "<plan part>"}`. A patch is always a region: its bone is `regions.<name>`, in the one place every region's bone lives, and a `meshes` entry for it is refused. `draw` sets assembled part order; an exclusively bound limb rider's final rig slot joins that limb's block (see the limbs subsection). See §5 for how to place one |
110
120
  | `assemble.cuts` | assemble; `meshes`/`regions` keys may name the new parts too, and must (each is a part like any other) | **authored**, optional (issue #170): `[{from, into, polygon, draw, overlap, under, under_fill}]`, `under` and `under_fill` optional — a plan part's layer cut by a polygon the author draws: the pixels of plan part `from` whose **centre** lies inside `polygon` (even-odd; a centre exactly on an edge is outside) leave `from` and make the new part `into`, colour and alpha, so every pixel of the layer ends in exactly one of the two. `polygon` is `[[x, y], …]`, at least three points, non-negative integer **rig** pixels — the space of `parts.json` and of the patches' boxes. The file that shows the layer in those coordinates is `assemble`'s own `rig/parts/<from>.png`, placed at its `parts.json` `x`, `y` (the PNG's pixel `(i, j)` is rig pixel `(x + i, y + j)`), or `render/recomposite_rig.png`, the whole rig 1:1; the contact sheets (`sheet`) are scaled and carry no coordinates. `draw` is `"back"`, `"front"` or `{"before": "<plan part>"}`, as for a patch (to keep the setup pose, draw the piece immediately in front of `from` — `{"before": "<the plan part after from>"}`, or `"front"` when `from` is the last plan part; `{"before": "<from>"}` draws it behind `from`, which `under` refuses as `ASSEMBLE_UNDER_DRAW` because the fill would hide it). `from` must be a plan part (`CONFIG_NAME_RESOLVES`; a patch or another cut's piece is refused; a part a limb takes may be cut where no segment of the limb holds it, issue #564, §5 *Limbs*), `into` a new name (`CONFIG_PART_UNIQUE`), the polygon three or more integer points (`CONFIG_FIELD_TYPE`). The piece's `parts.json` record carries `from`'s `<run>:<tag>` and its own counts, each split from the layer's by where the pixel lies; it is a mesh or a region under its own name, and the coverage rule holds it like a plan part. What the config cannot know the stage refuses: a polygon point past the rig (`ASSEMBLE_CUT_INSIDE`), two cuts of one part sharing a pixel (`ASSEMBLE_CUT_OVERLAP`), a cut that takes none of the layer's opaque pixels or leaves `from` none (`ASSEMBLE_CUT_PIXELS`), a cut of a part a limb takes that takes a pixel a segment of the limb holds (`ASSEMBLE_CUT_CROSSES_LIMB`). `overlap` is required, a non-negative integer of rig pixels with no default (`CONFIG_FIELD_TYPE`): the **band** — the polygon's pixels within `overlap` px of its edge by Chebyshev distance (the polygon's mask less its erosion by a `2 * overlap + 1` square) — stays in `from` as well as going to the piece, so the piece's resampled edge has art under it; the counts stay a strict partition and the band is the base's extra copy, printed on the `cut:` line. At `0` the two abut and `check`'s seam bar judges the edge between them (`CHECK_SEAM_WITHIN_BAR`, §5). A band only covers the edge at rest: a piece that swings well clear of where it was painted shows the hole it leaves, and a thin piece's band shows a ghost of the piece. `under` (issue #193), optional, is the author's image of what lies under the piece: a PNG path relative to the config's directory (`CONFIG_FIELD_TYPE` unless a non-empty string), in rig pixels over the polygon's **bounding box** — the least and greatest polygon x and y, so the image is `max x − min x` by `max y − min y` and its pixel `(i, j)` is rig pixel `(min x + i, min y + j)`. Where the cut removed pixels from `from` — the piece's pixels (alpha above 0) less the band — `from` takes the image's pixel, colour and alpha wherever the image's alpha is at least the alpha the cut removed there (**filled**); elsewhere `from` stays empty there (**left empty**): a transparent pixel is the author's statement that nothing lies under it (the background, or a part drawn behind), and a less opaque one is not half-filled. The image is read nowhere else, so its pixels over the band, over the box outside the polygon or where the layer had no art never change `from`. The `cut:` line then ends `; under <n> px filled, <m> px left empty` (`n + m` is every removed pixel), so an unintended hole is a number in the log, not a guess; the filled pixels count in `from`'s `opaque_px` and `occluded_px` (the piece is drawn over them), never as visible or projected. The stage refuses: a file that is not there (`ASSEMBLE_UNDER_PRESENT`) or not a PNG (`ASSEMBLE_UNDER_DECODES`), an image not exactly the box's size (`ASSEMBLE_UNDER_SIZE`, both sizes given), an image that fills none of the removed pixels — all transparent there, or another file (`ASSEMBLE_UNDER_EMPTY`; to leave the whole hole, drop `under`), and a piece drawn behind `from` (`ASSEMBLE_UNDER_DRAW`: the fill would hide it; draw it `"front"` or `{"before": "<the plan part after from>"}`). Producing the image is the author's, by whatever means; absent, `from` keeps only its band. `under_fill` (issue #360), optional, is the one value `"colour"` (`CONFIG_FIELD_TYPE` on any other): the same pixels are filled and left empty, and a filled pixel takes the image's **colour** at the alpha the cut removed there instead of the image's alpha, so `from`'s alpha is the uncut layer's wherever the image fills and its silhouette does not grow — an opaque image over a soft edge leaves the edge soft. It applies to the part's hidden-art layer as well as to a file; a cut with neither is refused (`ASSEMBLE_UNDER_FILL_SOURCE`). The `cut:` line's filled count then says how many took the colour only: `; under 190 px filled (190 colour only, at the alpha the cut removed), 0 px left empty`. Absent, a filled pixel takes the image's colour and alpha, as above |
111
- | `assemble.alternates` | assemble, rig, check (`--set`, `--all-sets`) | **authored**, optional (issues #454, #465, #472, #473, #476): `[{name, source?, full?, head?, head_box?, parts?, hides?, cuts?, skeleton?, drawn?, tips?}]` — alternate painting sets assembled into `<out>/rig/parts/<set>_<part>.png` (`parts.json` `alternates`), attached on each covered slot in `rig.json` as `<set>_<part>` beside the base attachment, and indexed in `<out>/rig/sets.json` (`rig-parts-sets/1`: `sets.<set>.slots`, `.hides`, `.drawn` when present, and `.threshold: {yaw: 15, pitch: 12}` when the set replaces `face` on a rig with `turn` or `turn`/`tilt` slider animations). `source`, `full` and `head` may instead be given on the CLI with `--alt <name>=<source>:<full>[:<head>]`. `parts` lists the subset of parts the set replaces (default: every part not in `hides`); a **subset set** (issue #473, e.g. mouth visemes `a`, `i`, `u` with `"parts": ["mouth"]`) cuts only its covered parts from the alternate runs (refusing with `ASSEMBLE_ALT_INPUT_PRESENT` if a covered part's layer is missing or empty), writes only `{slot: "<set>_<slot>"}` in `sets.json` so `@rig-c/play`'s `setSet(name)` switches those slots while leaving every other slot on its base attachment (and `setSet(null)` restores the base pose), may be keyed on a slot attachment track in `motion.animations_from` as `{"t": …, "set": "<name>" \| null}` (`CONFIG_SET_COVERS_SLOT` if `"set"` is undeclared or does not cover the track's slot), and under `check --set <name>` counts `ANIMATIONS_DO_NOT_FOLD` only on its own covered slots; `hides` lists slots the set hides (`null` in `sets.json`). A **gesture subset set** (issue #476) covers a limb chain declared in `config.limbs.<limb>` (all of its segment `part`s; with `drawn` alone, cut at the base's joints posed to the set's own `drawn` angles by `poseBoneEntriesForDrawn`, each segment keeping the base's length; with `skeleton` (issue #539), cut at the skeleton's own joint positions by `placeBoneEntriesForSkeleton` — each covered bone the body-18 table names (`sleeve_*`, `arm_*0`/`arm_*1`, `leg_*0`/`leg_*1`) at its skeleton origin with the skeleton segment's own length and direction (`drawn` overriding the direction of a bone it names), and a covered bone the table does not name (a hand link `arm_r2`, a single `foot_r`) placed from its parent's new place — its base offset from the parent, stretched along the parent by the parent's length change and turned by its turn, so a link or a foot at the parent's end lands on the skeleton's end — with its base length and the drawn or base direction, so a foreshortened shin or forearm is cut where it is painted; naming the unsplit `from` layer, covering only some segments of the limb, or declaring polygon `cuts` on a limb set is refused with `CONFIG_LIMB_SET_JOINTS`) and carries `drawn` (or `skeleton`) for the covered bones only (giving `drawn` for an uncovered bone is refused with `CONFIG_SET_COVERS_BONE` / `RIG_SET_COVERS_BONE`; uncovered bones keep the base's `drawn`). A subset set's limb part takes no pixel that a base part drawn after it, which the set neither covers nor hides, covers at setup (issue #587): the set's attachment is drawn in the base slot's place, so that base part is drawn over it and what the set's painting shows there is that base part (shorts over a leg set's thigh top, which See-through synthesised in the set's leg layer); `assemble` clears those pixels — the base part's PNG at alpha above 8, at its `parts.json` x, y, on the shared canvas — from the set's part once its limbs are cut, before any count — its `merged_px` and its cap's record too (issue #603): a cap pixel it clears leaves the count it was in (`layer_px`, `image_px`, `upper_px`, `lower_px`, `lower_kept_px`) for the cap's `excluded_px`, written only when above 0, so every count is the written PNG's — and prints `LINT set "<set>" <part>: <n> px under <base part> excluded` per part and base part that took a pixel (n counts alpha above 8); a set that carries that base part itself (`"parts": ["thigh_l", "shin_l", "bottomwear"]`) excludes nothing and writes what it wrote before. A subset set that names one limb side of a layer more limbs take (issue #588: `"parts": ["thigh_r", "shin_r"]` where `leg_r` and `leg_l` both take the one `legs` layer) would put every pixel of that layer on its one limb, so each of its parts would hold a piece of both legs; when the set's layer holds more than one island (8-connected pixels with alpha above 8) of at least 1 % of its largest (issue #608: the ghost clean-up's own share; a smaller island is a speck, which goes with the island nearest it — the one holding the pixel centre nearest one of the speck's, none when two are equally near — and is never an island of its own), each segment part keeps the island its joints fall on — its bone's origin and the next segment's bone's origin, at the set's own bones, a joint on the island of a pixel whose square holds the point, and a joint on no island on the one island whose nearest pixel centre lies within the radius of the cap cut at that joint (`cap.radius`: an ankle just past the legwear's edge, on the shoe) — and drops the rest, once its limbs are cut and before the exclusion above, and `assemble` prints `LINT set "<set>" <part>: <layer part> holds <n> islands (8-connected, alpha above 8)[, and <s> speck(s) under 1 % of the largest (<f> px), each with the island nearest it], kept the one under <bone> [x, y] and <bone> [x, y][ (<e> px off it, within its cap's radius <r>)] (<k> px), dropped <m> (<d> px)` per part (k and d count the part's pixels above alpha 8, m the islands it lost one from, and `dropped none` when it lost none); a joint on no island with none or two within that radius, or with no cap cut at it (a limb's first joint), and a segment whose joints fall on two, are refused with `ASSEMBLE_ALT_LIMB_ISLAND` rather than guessed. A layer of one island — specks aside — may hold both legs over each other (a folded shin across the standing leg; issue #609): each of the set's segment parts then keeps the pixels nearer its own limb's segment line — the polyline of only its segments taking that source, their bone origins and the last segment's supplied tip or following segment's joint (issue #638) — than the other limb's, both in the set's pose, and gives the rest away, once its limbs are cut and before the exclusion above; a pixel whose two distances differ by 0.5 px or less (the overlap band: half a pixel, the reach by which a pixel centre is read as on a line) goes to the limb whose part there is drawn later in the base's draw order, and the `LINT set` line above counts what each part kept, gave and held in the band. The other limb stands where the set's `skeleton` places it — its keypoints place every limb bone of the body-18 table, the other leg's as well as the set's own (its `drawn` and `drawnLength` stay the covered bones') — and never at the base's place: a set that does not place it (a `drawn` set, which may not name a bone it does not cover, or a set with neither) is refused with `ASSEMBLE_ALT_LIMB_POSE` when its layer has art where the base draws that limb, and when it has none nothing says the layer holds it and the layer is cut as before; two limbs whose segment lines run closer than the limb's half-width over most of where their art can meet are refused with `ASSEMBLE_ALT_LIMB_OVERLAP`. A set that carries every limb of the layer is cut as before. For a subset set, where a limb's own source layer holds art (`alpha > 8`) under another layer of that run drawn in front (`alpha > 8`, by the projection's plan/depth order), the limb keeps its own layer pixels through projection, limb cutting, island/nearest-line assignment, base-footprint exclusion, fringe clearing and seam correction (issue #637). Those pixels are occluded, never counted as painting projection; no base footprint or painting colour replaces them. After source-art island/nearest-line assignment, a source-transparent cap pixel is assigned before protection, in both doors: its centre is compared against only the segments taking the recipient's source, at the set's supplied bones (ankle-to-toe for a separate foot source), with the existing 0.5 px overlap band and base draw order. A missing last endpoint is refused with `ASSEMBLE_ALT_LIMB_POSE`; an actual following joint may supply that endpoint, never a guessed point. An unposed other limb with base art beneath these cap pixels is likewise refused; without such overlap the existing compatible path retains the cap without inventing the other limb's pose. A pixel cleared by this assignment is never restored. `CAP_FROM_UPPER` records a copy made before assignment and alone proves no ownership: only a retained copy is protected beneath the upper segment, which is drawn over it. The source-art island selection itself is unchanged. Other retained cap pixels (from the layer or a supplied cap image) are protected under an actual front layer even where the lower source is transparent. Protection follows each assigned piece through exclusion, fringe clearing and seam correction; it never restores a pixel ownership removed. Cap pixels removed during ownership leave their original counts for `excluded_px`, just as excluded pixels do, so the cap ledger describes the written PNG. Target assignment reconciles this ledger even when there are no source-transparent cap candidates and when a direct caller does not request cover protection. Unprotected source pixels still obey the base-footprint exclusion. Missing limb/cap art is never extended or guessed. Base builds and uncovered, unexcluded sets keep their existing bytes. `skeleton` (a pose file path relative to the config's directory, `{"width": W, "height": H, "points": {…}}` with all 18 body-18 keypoints on the canvas, validated by the same rules as `generation.control.skeleton`: `CONFIG_FILE_PRESENT`, `CONFIG_IS_JSON`, `CONFIG_FIELD_TYPE`, `CONFIG_KEY_KNOWN`, `CONFIG_KEY_UNIQUE`, `CONFIG_POSE_CANVAS`, `CONFIG_POSE_KEYPOINTS`, `CONFIG_POSE_IN_CANVAS`) or `drawn` (`{<bone>: <worldDegrees>}`, each bone resolved against `bones`, `CONFIG_NAME_RESOLVES`) declares the posed angles the alternate was painted at (issue #465); when `skeleton` is given, `assemble` derives `drawn` from its 18 keypoints (filtered to covered bones on a subset set, with `drawn` overriding any bone named in both), records both `skeleton` (scaled by `assemble.rig_scale` to rig px) and `drawn` in `parts.json` `alternates` — and, when the skeleton placed the set's limb bones (issue #539), `drawnLength` (`{<bone>: <ratio>}`, each covered table bone's skeleton segment length over its base length, 6 places) and `registration` (`{offset: [dx, dy], scale}`: the similarity without rotation, about the canvas origin, that maps the skeleton's torso — the mean of its shoulders `r_shoulder`/`l_shoulder` to the mean of its hips `r_hip`/`l_hip`, read where the base declares the bone rooted there, `sleeve_*0`/`arm_*0` and `leg_*0` — onto those base bones' origins, 6 places), both also written to `sets.json`; the set's painting must be registered, its torso within 1 rig px and 1 % in scale of the base's, or the set is refused by name before any joint is cut (`ASSEMBLE_ALT_SKELETON_REGISTERED`, with the offset and scale the painting would need) — and the `rig` stage writes `<set>_*` region and mesh attachments in the bone-local frame of `poseRigBonesForDrawn(bones, drawn)` (weighting limb segments directly from `skeleton`'s joint endpoints when `skeleton` is present) and emits `drawn` in `sets.json` so `rig-pose` and `check --set <name>` / `--all-sets` pose the bones at `drawn` without rotating the already-posed art a second time |
121
+ | `assemble.alternates` | assemble, rig, check (`--set`, `--all-sets`) | **authored**, optional (issues #454, #465, #472, #473, #476): `[{name, source?, full?, head?, head_box?, parts?, hides?, cuts?, skeleton?, drawn?, tips?, front?}]` — alternate painting sets assembled into `<out>/rig/parts/<set>_<part>.png` (`parts.json` `alternates`), attached on each covered slot in `rig.json` as `<set>_<part>` beside the base attachment, and indexed in `<out>/rig/sets.json` (`rig-parts-sets/1`: `sets.<set>.slots`, `.hides`, `.drawn` when present, and `.threshold: {yaw: 15, pitch: 12}` when the set replaces `face` on a rig with `turn` or `turn`/`tilt` slider animations). `source`, `full` and `head` may instead be given on the CLI with `--alt <name>=<source>:<full>[:<head>]`. `parts` lists the subset of parts the set replaces (default: every part not in `hides`); a **subset set** (issue #473, e.g. mouth visemes `a`, `i`, `u` with `"parts": ["mouth"]`) cuts only its covered parts from the alternate runs (refusing with `ASSEMBLE_ALT_INPUT_PRESENT` if a covered part's layer is missing or empty), writes only `{slot: "<set>_<slot>"}` in `sets.json` so `@rig-c/play`'s `setSet(name)` switches those slots while leaving every other slot on its base attachment (and `setSet(null)` restores the base pose), may be keyed on a slot attachment track in `motion.animations_from` as `{"t": …, "set": "<name>" \| null}` (`CONFIG_SET_COVERS_SLOT` if `"set"` is undeclared or does not cover the track's slot), and under `check --set <name>` counts `ANIMATIONS_DO_NOT_FOLD` only on its own covered slots; `hides` lists slots the set hides (`null` in `sets.json`). A **gesture subset set** (issue #476) covers a limb chain declared in `config.limbs.<limb>` (all of its segment `part`s; with `drawn` alone, cut at the base's joints posed to the set's own `drawn` angles by `poseBoneEntriesForDrawn`, each segment keeping the base's length; with `skeleton` (issue #539), cut at the skeleton's own joint positions by `placeBoneEntriesForSkeleton` — each covered bone the body-18 table names (`sleeve_*`, `arm_*0`/`arm_*1`, `leg_*0`/`leg_*1`) at its skeleton origin with the skeleton segment's own length and direction (`drawn` overriding the direction of a bone it names), and a covered bone the table does not name (a hand link `arm_r2`, a single `foot_r`) placed from its parent's new place — its base offset from the parent, stretched along the parent by the parent's length change and turned by its turn, so a link or a foot at the parent's end lands on the skeleton's end — with its base length and the drawn or base direction, so a foreshortened shin or forearm is cut where it is painted; naming the unsplit `from` layer, covering only some segments of the limb, or declaring polygon `cuts` on a limb set is refused with `CONFIG_LIMB_SET_JOINTS`) and carries `drawn` (or `skeleton`) for the covered bones only (giving `drawn` for an uncovered bone is refused with `CONFIG_SET_COVERS_BONE` / `RIG_SET_COVERS_BONE`; uncovered bones keep the base's `drawn`). A subset set's limb part takes no pixel that a base part drawn after it, which the set neither covers nor hides, covers at setup (issue #587): the set's attachment is drawn in the base slot's place, so that base part is drawn over it and what the set's painting shows there is that base part (shorts over a leg set's thigh top, which See-through synthesised in the set's leg layer); `assemble` clears those pixels — the base part's PNG at alpha above 8, at its `parts.json` x, y, on the shared canvas — from the set's part once its limbs are cut, before any count — its `merged_px` and its cap's record too (issue #603): a cap pixel it clears leaves the count it was in (`layer_px`, `image_px`, `upper_px`, `lower_px`, `lower_kept_px`) for the cap's `excluded_px`, written only when above 0, so every count is the written PNG's — and prints `LINT set "<set>" <part>: <n> px under <base part> excluded` per part and base part that took a pixel (n counts alpha above 8); a set that carries that base part itself (`"parts": ["thigh_l", "shin_l", "bottomwear"]`) excludes nothing and writes what it wrote before. A subset set that names one limb side of a layer more limbs take (issue #588: `"parts": ["thigh_r", "shin_r"]` where `leg_r` and `leg_l` both take the one `legs` layer) would put every pixel of that layer on its one limb, so each of its parts would hold a piece of both legs; when the set's layer holds more than one island (8-connected pixels with alpha above 8) of at least 1 % of its largest (issue #608: the ghost clean-up's own share; a smaller island is a speck, which goes with the island nearest it — the one holding the pixel centre nearest one of the speck's, none when two are equally near — and is never an island of its own), each segment part keeps the island its joints fall on — its bone's origin and the next segment's bone's origin, at the set's own bones, a joint on the island of a pixel whose square holds the point, and a joint on no island on the one island whose nearest pixel centre lies within the radius of the cap cut at that joint (`cap.radius`: an ankle just past the legwear's edge, on the shoe) — and drops the rest, once its limbs are cut and before the exclusion above, and `assemble` prints `LINT set "<set>" <part>: <layer part> holds <n> islands (8-connected, alpha above 8)[, and <s> speck(s) under 1 % of the largest (<f> px), each with the island nearest it], kept the one under <bone> [x, y] and <bone> [x, y][ (<e> px off it, within its cap's radius <r>)] (<k> px), dropped <m> (<d> px)` per part (k and d count the part's pixels above alpha 8, m the islands it lost one from, and `dropped none` when it lost none); a joint on no island with none or two within that radius, or with no cap cut at it (a limb's first joint), and a segment whose joints fall on two, are refused with `ASSEMBLE_ALT_LIMB_ISLAND` rather than guessed. A layer of one island — specks aside — may hold both legs over each other (a folded shin across the standing leg; issue #609): each of the set's segment parts then keeps the pixels nearer its own limb's segment line — the polyline of only its segments taking that source, their bone origins and the last segment's supplied tip or following segment's joint (issue #638) — than the other limb's, both in the set's pose, and gives the rest away, once its limbs are cut and before the exclusion above; a pixel whose two distances differ by 0.5 px or less (the overlap band: half a pixel, the reach by which a pixel centre is read as on a line) goes to the limb whose part there is drawn later in the base's draw order, and the `LINT set` line above counts what each part kept, gave and held in the band. The other limb stands where the set's `skeleton` places it — its keypoints place every limb bone of the body-18 table, the other leg's as well as the set's own (its `drawn` and `drawnLength` stay the covered bones') — and never at the base's place: a set that does not place it (a `drawn` set, which may not name a bone it does not cover, or a set with neither) is refused with `ASSEMBLE_ALT_LIMB_POSE` when its layer has art where the base draws that limb, and when it has none nothing says the layer holds it and the layer is cut as before; two limbs whose segment lines run closer than the limb's half-width over most of where their art can meet are refused with `ASSEMBLE_ALT_LIMB_OVERLAP`. A set that carries every limb of the layer is cut as before. For a subset set, where a limb's own source layer holds art (`alpha > 8`) under another layer of that run drawn in front (`alpha > 8`, by the projection's plan/depth order), the limb keeps its own layer pixels through projection, limb cutting, island/nearest-line assignment, base-footprint exclusion, fringe clearing and seam correction (issue #637). Those pixels are occluded, never counted as painting projection; no base footprint or painting colour replaces them. After source-art island/nearest-line assignment, a source-transparent cap pixel is assigned before protection, in both doors: its centre is compared against only the segments taking the recipient's source, at the set's supplied bones (ankle-to-toe for a separate foot source), with the existing 0.5 px overlap band and base draw order. A missing last endpoint is refused with `ASSEMBLE_ALT_LIMB_POSE`; an actual following joint may supply that endpoint, never a guessed point. An unposed other limb with base art beneath these cap pixels is likewise refused; without such overlap the existing compatible path retains the cap without inventing the other limb's pose. A pixel cleared by this assignment is never restored. `CAP_FROM_UPPER` records a copy made before assignment and alone proves no ownership: only a retained copy is protected beneath the upper segment, which is drawn over it. The source-art island selection itself is unchanged. Other retained cap pixels (from the layer or a supplied cap image) are protected under an actual front layer even where the lower source is transparent. Protection follows each assigned piece through exclusion, fringe clearing and seam correction; it never restores a pixel ownership removed. Cap pixels removed during ownership leave their original counts for `excluded_px`, just as excluded pixels do, so the cap ledger describes the written PNG. Target assignment reconciles this ledger even when there are no source-transparent cap candidates and when a direct caller does not request cover protection. Unprotected source pixels still obey the base-footprint exclusion. Missing limb/cap art is never extended or guessed. Base builds and uncovered, unexcluded sets keep their existing bytes. `skeleton` (a pose file path relative to the config's directory, `{"width": W, "height": H, "points": {…}}` with all 18 body-18 keypoints on the canvas, validated by the same rules as `generation.control.skeleton`: `CONFIG_FILE_PRESENT`, `CONFIG_IS_JSON`, `CONFIG_FIELD_TYPE`, `CONFIG_KEY_KNOWN`, `CONFIG_KEY_UNIQUE`, `CONFIG_POSE_CANVAS`, `CONFIG_POSE_KEYPOINTS`, `CONFIG_POSE_IN_CANVAS`) or `drawn` (`{<bone>: <worldDegrees>}`, each bone resolved against `bones`, `CONFIG_NAME_RESOLVES`) declares the posed angles the alternate was painted at (issue #465); when `skeleton` is given, `assemble` derives `drawn` from its 18 keypoints (filtered to covered bones on a subset set, with `drawn` overriding any bone named in both), records both `skeleton` (scaled by `assemble.rig_scale` to rig px) and `drawn` in `parts.json` `alternates` — and, when the skeleton placed the set's limb bones (issue #539), `drawnLength` (`{<bone>: <ratio>}`, each covered table bone's skeleton segment length over its base length, 6 places) and `registration` (`{offset: [dx, dy], scale}`: the similarity without rotation, about the canvas origin, that maps the skeleton's torso — the mean of its shoulders `r_shoulder`/`l_shoulder` to the mean of its hips `r_hip`/`l_hip`, read where the base declares the bone rooted there, `sleeve_*0`/`arm_*0` and `leg_*0` — onto those base bones' origins, 6 places), both also written to `sets.json`; the set's painting must be registered, its torso within 1 rig px and 1 % in scale of the base's, or the set is refused by name before any joint is cut (`ASSEMBLE_ALT_SKELETON_REGISTERED`, with the offset and scale the painting would need) — and the `rig` stage writes `<set>_*` region and mesh attachments in the bone-local frame of `poseRigBonesForDrawn(bones, drawn)` (weighting limb segments directly from `skeleton`'s joint endpoints when `skeleton` is present) and emits `drawn` in `sets.json` so `rig-pose` and `check --set <name>` / `--all-sets` pose the bones at `drawn` without rotating the already-posed art a second time. `front` (issue #664) lists covered limb segment parts the set draws on their **front slot** instead of the base slot: each must be in the set's `parts` and not its `hides`, and listed in some `config.limbs.<limb>.front.parts` (`CONFIG_SET_FRONT`); the set then maps `<part>_front` (never `<part>`) to `<set>_<part>` in `parts.json` and `sets.json` and hides `<part>` (and `<part>_still`), its #587 exclusion counts only base parts drawn after the front slot's target, and keying that part's set on a slot track is refused (`CONFIG_SET_FRONT_KEYED`) — *Front slots (issue #664)* below |
112
122
  | `assemble.head_face_underfill` | assemble, `assemble --propose-plan` | **authored**, optional (issue #492): `boolean` (`CONFIG_FIELD_TYPE` on any other). When both runs exist and `full:face` has at least `200` opaque run px (`PROPOSE_MIN_PX`), the stage maps `full:face` into the head canvas via `fullLayerToHead` and measures how much of its hole-filled silhouette (`alpha > 8`) is covered by `head:face` (`alpha > 8`). A head-run `face` covering less than `75 %` (`HEAD_FACE_COVERAGE_FLOOR`, measured `0.0 %..48.6 %` on hollow/omitted head runs vs `84.8 %..99.6 %` on clean head runs across the 4 public examples and 9-standee corpus) is refused with `SEETHROUGH_HEAD_RUN_HOLLOW` unless `head_face_underfill: true` is declared. When `head_face_underfill: true` is declared and `head:face` exists (`opaqueRunPx >= 200`), `fullLayerToHead(full:face)` is composited under `head:face` on the head canvas (`alpha = max(headAlpha, fullAlpha)`, RGB blended by `headAlpha / 255`) before `head:face` is placed in rig space (`underfill=<N>px(full:face)` on the `face` part log line and a `face: head_face_underfill composited …` note in `assemble --propose-plan`). The fill restores transparent pixels only and leaves the cut's own edge where it is — a hollow whose edge is not under hair stays visible, and that is the author's to fix at the cut. When `head:face` has fewer than `200` opaque pixels, `SEETHROUGH_HEAD_RUN_HOLLOW` still fires even under `head_face_underfill: true` because the head run omitted the face layer and cannot supply a jawline |
113
- | `limbs` | assemble (the cut and the caps), rig (through `regions`), check (`LIMB_ONE_BONE`) | **proposed** by `propose --keypoints` (issue #349), then checked: `{<limb>: {segments: [{part, from, bone}, {part, from, bone, joint, range?, cap?, edge?, cap_shape?, cut_band?}, …]}}`, the upper segment first. The rule is *Limbs: cut at their joints* below the table. `part` is a new part (`CONFIG_PART_NAME`, `CONFIG_PART_UNIQUE` against the plan, patches, cuts and every other limb part); `from` a plan part or a cut's piece (`CONFIG_NAME_RESOLVES`) — the layer the segment's pixels come from, which is cut into the limb parts and is not itself made: a `meshes` or `regions` entry for it is `CONFIG_NAME_RESOLVES`, and the segments of one limb that take one part are consecutive (`CONFIG_LIMB_FROM_RUN`); `bone` a declared bone, each segment's the child of the one before (`CONFIG_LIMB_BONES_CHAIN`); the upper segment names no `joint`, `range`, `cap`, `edge`, `cap_shape` or `cut_band` (`CONFIG_KEY_KNOWN`), every other names its `joint` (`CONFIG_FIELD_PRESENT`) — the joint is its bone's origin; `range` is `[min, max]` degrees from rest with `min <= 0 <= max` (`CONFIG_LIMB_RANGE`), which `assemble` writes into the lower part's `parts.json` record and `check` poses the joint at — its two ends and its middle — reading the seam there (`SEAM_OVER_RANGE`, issue #360; rig-c needs no rotation limit); `cap` a PNG path relative to the config's directory (`CONFIG_FIELD_TYPE`); `edge` `"feather"` (issue #360), the one value it takes, and `cap_shape` `"disc"` (issue #360) or `"ball"` (issue #579) (`CONFIG_FIELD_TYPE` on any other) — the cut's geometry at that joint, *Limbs: cut at their joints* below; absent, the cut is hard and the cap the layer's half-disc of half the limb's width; `cut_band` a whole number of px, 1 or more (`CONFIG_FIELD_TYPE`), on a segment that takes the segment above's part — between two layers nothing is cut, so a band there is `CONFIG_KEY_KNOWN` — places the cut within that many px of the joint line (issue #360, *Limbs: cut at their joints*); absent, the cut is the joint line. Each limb part is attached like any part — `regions.<part>`, which must be the segment's `bone` (`CONFIG_LIMB_BONE`), or a mesh, which `check` holds to that bone alone (`LIMB_ONE_BONE`). With `limbs`, `assemble` reads `bones` too (`CONFIG_FIELD_PRESENT` without it). Absent, nothing is cut at a joint and no byte of any output moves |
123
+ | `limbs` | assemble (the cut and the caps), rig (through `regions`), check (`LIMB_ONE_BONE`) | **proposed** by `propose --keypoints` (issue #349), then checked: `{<limb>: {segments: [{part, from, bone}, {part, from, bone, joint, range?, cap?, edge?, cap_shape?, cut_band?}, …], split?, layout?, front?}}`, the upper segment first. The rule is *Limbs: cut at their joints* below the table. `part` is a new part (`CONFIG_PART_NAME`, `CONFIG_PART_UNIQUE` against the plan, patches, cuts and every other limb part); `from` a plan part or a cut's piece (`CONFIG_NAME_RESOLVES`) — the layer the segment's pixels come from, which is cut into the limb parts and is not itself made: a `meshes` or `regions` entry for it is `CONFIG_NAME_RESOLVES`, and the segments of one limb that take one part are consecutive (`CONFIG_LIMB_FROM_RUN`); `bone` a declared bone, each segment's the child of the one before (`CONFIG_LIMB_BONES_CHAIN`); the upper segment names no `joint`, `range`, `cap`, `edge`, `cap_shape` or `cut_band` (`CONFIG_KEY_KNOWN`), every other names its `joint` (`CONFIG_FIELD_PRESENT`) — the joint is its bone's origin; `range` is `[min, max]` degrees from rest with `min <= 0 <= max` (`CONFIG_LIMB_RANGE`), which `assemble` writes into the lower part's `parts.json` record and `check` poses the joint at — its two ends and its middle — reading the seam there (`SEAM_OVER_RANGE`, issue #360; rig-c needs no rotation limit); `cap` a PNG path relative to the config's directory (`CONFIG_FIELD_TYPE`); `edge` `"feather"` (issue #360), the one value it takes, and `cap_shape` `"disc"` (issue #360) or `"ball"` (issue #579) (`CONFIG_FIELD_TYPE` on any other) — the cut's geometry at that joint, *Limbs: cut at their joints* below; absent, the cut is hard and the cap the layer's half-disc of half the limb's width; `cut_band` a whole number of px, 1 or more (`CONFIG_FIELD_TYPE`), on a segment that takes the segment above's part — between two layers nothing is cut, so a band there is `CONFIG_KEY_KNOWN` — places the cut within that many px of the joint line (issue #360, *Limbs: cut at their joints*); absent, the cut is the joint line. Each limb part is attached like any part — `regions.<part>`, which must be the segment's `bone` (`CONFIG_LIMB_BONE`), or a mesh, which `check` holds to that bone alone (`LIMB_ONE_BONE`). With `limbs`, `assemble` reads `bones` too (`CONFIG_FIELD_PRESENT` without it). Absent, nothing is cut at a joint and no byte of any output moves. Optional `front` (issue #664), `{parts: [<own segment part>, …], after: "<produced part>"}`, gives each named segment a second slot `<part>_front`, empty at setup, placed after `after`'s slot family, for a gesture set to draw on (`CONFIG_FRONT_SCOPE`, `CONFIG_FRONT_NAME`, `CONFIG_FRONT_LAYOUT`; *Front slots (issue #664)* below) |
114
124
  | `bones` | rig; `propose --from-config` | **proposed** by `propose`, then corrected. A single bone `{name, parent, at, tip?}` or a chain `{chain, parent, points, tip, link_inherit?}` whose links are named `<chain>0 … <chain>n`. Parents come before children. In `rig.json` each link is **turned along its chain** — `rotation` is the direction from its origin to the next link's (the last link's to `tip`) and `length` that distance (issue #73), so a physics constraint added downstream finds a lever; under `--idle-keys ctl` the link's `<link>_ctl` carries the turn and the length and the link sits at local rotation 0 beneath it. A single bone with a `tip` is turned the same way, towards its tip, with that distance as its length (issue #239); a tip on the bone's own origin is refused (`RIG_TIP_DISTINCT`). The five face bones `eye_r`, `eye_l`, `brow_r`, `brow_l` and `mouth` (by name, when one has no tip of its own) are turned to the eye line — the line through the origins of `eye_r` and `eye_l` as `bones` writes them, read image-left to image-right, drawn when parts.json holds both an `eyewhite-r` and an `eyewhite-l` part — so the blink's `scaley` squashes across the eye line on a tilted head; move an eye bone and the line moves with it. With one eyewhite or none, or an eye bone missing, the rig stage's `face axis:` line says so and they keep the page's axes. A bone under a face bone keeps the face's axis. With the line, `rig.json` also declares `invariants.axes` — one entry, `"face"`, through `eye_r` and `eye_l`, holding the turned face bones (and their controls) — so rig-c's gate holds the build to the line; the line is the one rigc draws through the same two bones, so the tolerance is only the float floor — the largest gap the emitted (float32) bones have to it, rounded up to 6 places plus one 6-place unit — and the `face axis:` line prints it. With the line the rig stage also adds `head_up` (issue #250): a slotless bone under `head`, at head's origin, turned to the eye line + 90 degrees and as long as the eye midpoint lies from head's origin along that axis — the head's frame, which a rule that needs it reads; `head` keeps its own frame, so no key on it or under it changes meaning, and a config bone of that name is refused (`RIG_HEAD_UP_NAME_FREE`). A face bone with its own `tip` is held too (issue #258), in a second entry, `"face_tips"`, when its tip lies along the line within whole-pixel rounding — a tip written in whole pixels sits at most sqrt(2)/2 px off the point of the line it stands for, so its direction may sit up to asin((sqrt(2)/2) / length) degrees off — with its tolerance measured the same way; a tip further off is the bone's own axis, and the `face axis:` line names it with its gap and that bound. A failure is rig-c's `A51_BONES_LIE_ALONG_THEIR_DECLARED_AXIS` line, which names the bone and the fix (rig-c's `docs/AUTHORING.md` §3.7 and §5.2); with no line no entry is written and A51 reports SKIP. Every region on a turned bone is turned back, so the setup pose does not move. Any other single bone keeps world rotation 0; one whose parent is a link is turned back upright. A single bone may declare `inherit` (issue #348): what it takes from its parent's transform, one of rig-c's own mode names exactly as rig-c's table spells them — `normal`, `onlyTranslation`, `noRotationOrReflection`, `noScale`, `noScaleOrReflection` (rig-c's `RigBone.inherit`) — written to that bone in `rig.json` as declared; any other spelling, `"NoScale"` included (the runtime would fold its first letter; nothing here folds or maps), is refused (`CONFIG_BONE_INHERIT`), and a chain takes no `inherit` (`CONFIG_KEY_KNOWN`). A chain may instead declare `link_inherit` (issue #369), an object mapping its own expanded link names to those same modes, for example `"link_inherit": {"leg_r1": "noScale"}`. Only the named link changes; omitted links inherit normally. A key outside this chain is `CONFIG_NAME_RESOLVES`, with this chain's valid link names in the detail; a non-object or non-string mode is `CONFIG_FIELD_TYPE`, and another mode spelling is `CONFIG_BONE_INHERIT`. The mode is forwarded to the expanded link, or to its control under `--idle-keys ctl`, by the same rule as a single bone. This is an authored opt-in: `propose` does not add it. Its setup rotation is the turn from where rig-c's own poser puts the bone with no rotation of its own under that mode, so it still stands upright (or along its tip) at the setup pose. Where the idle gives the bone a `<bone>_ctl`, the mode goes on the control: the control is the bone the parent's rotation reaches, and the bone under it inherits normally, so the idle's keys moved onto the control still turn it. A foot that does not take its shin's rotation (`noRotationOrReflection` or `onlyTranslation`) keeps its shoe level while the leg is posed (`FEET_STAY_LEVEL`, §7); which mode is the author's. `noScale` keeps a shin or shoe's shape under a scaled parent but still takes the parent's rotation: a shoe that must also stay level needs a separate rotation constraint, for example a transform constraint from an unrotated bone. This does not change the proposed feet's existing inheritance or `FEET_STAY_LEVEL` default. An automatic mesh weighted through a bone with a mode other than `normal` gets no derived amplitude (its row names the bone) and no residual envelope (rig-c's `SKINNING_RANGE_UNSUPPORTED`). Absent, `rig.json` carries no `inherit` and no byte moves |
115
125
  | `meshes.<part>` | rig | **proposed**, then corrected. Exactly one of `grid` (the lattice: cell size, px — what `propose` writes: by default one spacing for the figure, round(H / 37) rig px with H the height of every part's box together, a part whose tenth of its short side is already coarser keeping that, and under `propose --grid-rule part` the reference's tenth of the part's short side clamped to 5..36, issue #265), `contour` (the outline mode, authored; the next row and "A contour mesh" below) or `auto` (the automatic mode, authored; "An automatic mesh" below) — two, or none, is `CONFIG_MESH_MODE`. In every mode: `r` (added to every distance before weighting: `w = 1/(d + r)²`), `segments` (a chain name, a bone name with a `tip`, or `[bone, [x0,y0], [x1,y1]]`), and, **authored**, optional, `exponent` (issue #161): a number above 0 that replaces the 2 in that rule, `w = 1/(d + r)^exponent` — a larger one hands each vertex more fully to its nearest bone; absent is 2 and changes no byte, a value that is not a number above 0 is `CONFIG_FIELD_TYPE` at `meshes.<part>.exponent`, and a declared one is echoed as `exponent` in the mesh's `mesh_report.json` row; the exponent is the distance rule's, so declared beside `rule: "heat"` it is refused (`CONFIG_FIELD_TYPE` at `meshes.<part>.exponent`). The segment list is the one authored decision about a layer: which bones may pull it. The slot's bone is the first segment's. Optional `rule` (issue #161): `"distance"` — the rule above, and what an absent `rule` means (every byte is the same) — or `"heat"`, bone heat over the part's own silhouette, in every mode: the lattice's art (alpha above 8, less any island a declared `stray` left out) with holes filled, or for `contour` and `auto` the silhouette their outline is traced from (that mode's art — alpha above 8 for `contour`, alpha 1 and above for `auto` — less any island a declared `stray` left out, holes filled, grown by its `margin`); each bound bone's sources are the silhouette pixels whose centre lies within 0.75 px of its segment (a pixel two bones reach goes to the nearer, a tie to the first listed), held at 1 for that bone and 0 for the others; the field is solved by row-major Gauss–Seidel over-relaxation to a sweep change under 1e-7 within 20000 sweeps; a vertex reads the nearest silhouette pixel and takes the same cap and floor (and `regions`) as the distance rule; `r` is not read. The mesh's `mesh_report.json` row gains `heat` (the constants, each bone's sources and sweeps, the residual). Any other value is `CONFIG_FIELD_TYPE`; a bound bone with no source pixel is `RIG_HEAT_BONE_SOURCE`, an island no bone has a source in `RIG_HEAT_ISLAND_SOURCE`, a solve past its sweep bound `RIG_HEAT_CONVERGED` — none is weighted 0. The rule is chosen per mesh, so `RIG_HEAT_BONE_SOURCE` names three remedies: move the segment onto the art, take the bone out of `segments`, or keep this part on the distance rule — the one to take when the bone anchors the part from outside its art (a torso bone under a neck), since the first moves a bone another part placed and the second drops the anchor (issue #303). On a lattice, `RIG_HEAT_ISLAND_SOURCE` on an island that is dust — at most 4 px and at most 0.1 % of the part's art px, `mesh-draft`'s line — names `meshes.<part>.stray` and the figure that leaves it out, with the value declared; an island above that line keeps the segment remedy, a real piece being a real authoring question. In every mode the mesh's `mesh_report.json` row's `bones` names the bones the emitted model binds for the attachment — the sorted set its weights in the written `rig.json` name, so a `turn` twin (`turn_<part>_<bone>`) or a stopped link's `<link>_d` where those rebind it, never the segment bone it stands for — the same list in every mode, `remesh --delta`'s included (issue #536) |
116
126
  | `meshes.<part>.influence` | rig | **authored**, optional (issue #378), a mode beside `grid`, `contour` and `auto`: `{cell, radius, across, max, shape, hold_tip?}`, every field but `hold_tip` required, none defaulted. rig-c's `segments` lattice over the part at `cell` px, each vertex pulled by every link of the mesh's `segments` within `radius` px of it — a pull that falls to exactly 0 at `radius` (`shape` `"linear"` or `"smooth"`), across the `segments` entries or only within the nearest one (`across`), the strongest `max` kept. The lattice and the shares are computed by rig-c's own promised functions (`buildSegmentsLattice`, `segmentDistance`, `influenceShares`), so the rule is rig-c's (its AUTHORING §3.4.7), and the mesh is written with explicit weights like every other mode. `radius` is a length on this part: the rig stage measures the farthest lattice vertex's distance to its nearest link and refuses a radius at or under it (`RIG_INFLUENCE_REACH`), naming the distance. `propose --cloth` drafts it when a recipe declares `influence` and no `radius` (issue #518): the farthest vertex's distance plus 3 of the part's lattice cells, printed on a `LINT` line with both figures. Measured on the cloth fixture and the public `wind` example (#513: 8 covered parts, `render --folds` with the physics on under wind and a hip sway of 8° and 30°): a radius just above the floor folds **more** than the lattice (one skirt went from 0 to 55 folded frames), and a larger one is **not** safer (one fringe folded again at about 3–4× its farthest distance); +2 to +6 cells read 0 on every part, +2 at the edge. That reading counted folds only; read against `check` as well (the rig as built, under the wind schedule +1 → −1 → calm, `build --environment`, the three covered `wind` parts at the same number of cells), the window that holds is **+2 to +3 cells**, and +4 is the ceiling: `wind` `bottomwear`'s `CHECK_TIP_OVER_ROOT` ratio is 1.898 on the lattice, 1.801 at +2 cells, 1.590 at +3 and **1.471 at +4**, under the floor 1.4725, so the build fails there, and 1.351 at +6 (#513, rig-parts 1.52.0, rigc 2.60.0; the lattice, +3 and +4 read again on rig-parts 1.54.0, rigc 2.60.0, the same figures, #599). The drafted +3 sits in that window, one cell under the ceiling (no radius between +3 and +4 was read), and the window is a hypothesis until a third subject — each edge rests on one part — so state `radius` in the recipe to keep your own. `r`, `exponent`, `rule`, `influences`, `stray` and `ribs` belong to the distance rule or another mode and are refused beside it (`CONFIG_INFLUENCE_MODE`). `hold_tip` (issue #615) — `true` only, or absent: the pulled shares are smoothed by the lattice's held-fork passes (the cloth recipe's `split.hold_tip` sentence below: four passes of 1-ring mesh-edge averaging, self weight 0.5, floor 0.005) capped at the declared `max`, so no vertex binds more than `max` bones; absent, the shares are rig-c's as pulled and every byte is the one written before the field existed (`false`, a second spelling of absent, is `CONFIG_FIELD_TYPE`). The mesh's `mesh_report.json` row echoes it in `influence`. `propose --cloth` writes it, with `max` raised to at least 8 (the lattice's smoothing cap), on the parts a recipe with `influence` and `split.hold_tip: true` meshes; the rig stage reads it off the mesh and never the recipe. Measured on `wind` `bottomwear` and the cloth fixture's skirt under a 60° hip sway (#615, `render --folds`, physics on): `wind` `bottomwear` folds 41 frames as drafted before the field (max 6) and **0** with `hold_tip` and max 8, against its lattice's 5; with `hold_tip` at max 6 it reads 23, and at max 8 without `hold_tip` 3; the fixture skirt (recipe given `split.hold_tip`) reads 20 before and 0 with it, against its lattice's 0, and 0 at max 8 without it. So the raised `max` carries most of the gap and the smoothing the rest on `wind` (rig-parts 1.54.0, rigc 2.60.0; a split skirt without `hold_tip` is not smoothed and reads its 44 as before). For cloth and hair that should bend as one surface across its chains |
@@ -867,6 +877,59 @@ The first character built this way (the exam-01 test rig, issue #349) read, per
867
877
  missing cap pixels at the elbows and knees and 468 and 562 at the ankles, where the shoe starts
868
878
  below the ankle line.
869
879
 
880
+ ### Front slots (issue #664)
881
+
882
+ A gesture set (`assemble.alternates`, issue #476) replaces a limb's attachments
883
+ but cannot change where they are drawn: a set maps slot to attachment and never
884
+ reorders slots. A forearm brought across the chest is drawn where the base arm
885
+ is, behind the topwear and the face. A **front slot** is a second slot for one
886
+ segment, declared on its limb:
887
+
888
+ ```json
889
+ "limbs": { "arm_r": { "segments": [ … ], "front": { "parts": ["forearm_r"], "after": "face" } } }
890
+ ```
891
+
892
+ and drawn on by a set that names the part in its own `front`:
893
+
894
+ ```json
895
+ { "name": "chest", "parts": ["upperarm_r", "forearm_r"], "drawn": { "sleeve_r1": 0 }, "front": ["forearm_r"] }
896
+ ```
897
+
898
+ - `front.parts` is a nonempty list of this limb's own segment parts, each once
899
+ (`CONFIG_FRONT_SCOPE`). `front.after` names one produced part — the same
900
+ notion a layout's `part` target uses, so a consumed source layer is not one —
901
+ and not a segment of this same limb (`CONFIG_FRONT_NAME`).
902
+ - The rig stage adds `<part>_front` on the bone the part's base slot binds, with
903
+ **no setup attachment** (`"attachment": null`), and places it immediately
904
+ after `after`'s whole slot family (its slot and its `<part>_still`) in the
905
+ final setup order; several front slots after one family keep the config's
906
+ order. The slot rides stops and the turn control with the rest, and stays out
907
+ of the limb closure (#646). An empty front slot renders nothing: with no set
908
+ using it, the setup render is pixel-identical, and a config with no `front`
909
+ anywhere writes the same bytes as before.
910
+ - A set's `front` part is drawn from `slots["<part>_front"]` and its base slot
911
+ `<part>` is hidden while the set is active (`sets.json` `hides`, with
912
+ `<part>_still` when there is one); its attachment is filed under the front
913
+ slot's skin key, with the same bones and weights. A part in a set's `front`
914
+ must be covered by that set and declared on its limb (`CONFIG_SET_FRONT`); a
915
+ `parts.json` set record mapping both `<part>` and `<part>_front` is refused
916
+ (`RIG_SET_FRONT_BOTH`).
917
+ - `assemble` places the front part after its target when it clears the
918
+ set's part of what base parts cover (#587) and when it measures the set's
919
+ recomposite; `check --set` stacks the set's flat composite the same way, at
920
+ the front slot's place in the rig's final slot order, so the seam reads the
921
+ forearm against the parts now behind it.
922
+ - Version 1 does not compose with an explicit limb layout (#647): a config that
923
+ declares `front` on one limb and `layout` on any limb is refused
924
+ (`CONFIG_FRONT_LAYOUT`). Keyed draw-order changes are out of scope: a `{"t", "set"}`
925
+ key that switches a set drawing the track's part on its front slot is refused
926
+ (`CONFIG_SET_FRONT_KEYED`) — show such a set with `setSet`.
927
+ - `assemble` reads the front part's place off `parts.json`'s order (after
928
+ `after`), while the rig and `check --set` read it off the final setup slot
929
+ order. The two agree unless the limb closure (#646) moves the target's
930
+ family; then only the set's #587 exclusion and recomposite figures read the
931
+ parts order, and the render and its seam bar follow the rig.
932
+
870
933
  ### The weight cut (`meshes.<part>.influences`, issue #284)
871
934
 
872
935
  This section and the mesh sections after it are for parts with no joints; a limb is cut at
@@ -2046,10 +2109,13 @@ moved on the public `wind` example with its keypoints, whose tracked config the
2046
2109
  wrote back byte for byte (measured once, when the flag landed). The text is written as stated: a relative path in it
2047
2110
  (`motion.animations_from`, `generation.control.skeleton`) resolves against the directory
2048
2111
  the written config sits in, so the out path belongs beside the input, and when it is not
2049
- a note names each such path. Refused before anything is written: `--write-config`
2050
- without `--cloth` or beside `--from-config`, and `--from` without `--write-config`
2051
- (`USAGE`), an out path that exists or that
2052
- the run writes itself (`PROPOSE_WRITE_CONFIG_FRESH`), a hand edit to an entry the
2112
+ a note names each such path. These are the cloth mode's rules; the three other
2113
+ `--write-config` modes are described in §4. Refused before anything is written:
2114
+ `--write-config` beside `--from-config`, and `--from` without both `--cloth` and `--write-config`
2115
+ (`USAGE`), an out path that exists or collides with any of the four artifacts
2116
+ under `--out` (`proposal.json`, `basis.json`, `render/landmarks.png`,
2117
+ `render/landmarks_head.png`), including directory-symlink and hardlink aliases
2118
+ (`PROPOSE_WRITE_CONFIG_FRESH`), a hand edit to an entry the
2053
2119
  recipe owns (`PROPOSE_WRITE_CONFIG_CONFLICT`), and a text whose splice does not
2054
2120
  parse to the config with the sections set — a key stated twice
2055
2121
  (`PROPOSE_WRITE_CONFIG_SPLICE`); a recipe that drafts nothing is
@@ -2446,21 +2512,28 @@ full loader before its assemble stage writes anything.
2446
2512
  a painting padded to a square by hand, whose padding becomes part of the rig.
2447
2513
  Config: the `layers` row, `head_box` not yet. A translucent painting is refused
2448
2514
  (`INPUTS_PAINTING_OPAQUE`).
2449
- 2. **See-through, full run** (external).
2515
+ 2. **See-through, full run** (external), on `inputs/st_input_full.png`. Upload
2516
+ that white-padded square to the Hugging Face Space too. Its [current source](https://huggingface.co/spaces/24yearsold/see-through-demo/blob/049a1ab07908e404aad03f73431dd1a98f88202c/common/utils/cv.py#L254-L274)
2517
+ pads non-square images before resizing; the earlier centre-crop report is not
2518
+ the current source path. The explicit white square keeps the input consistent
2519
+ with rig-parts' full-run coordinate map.
2450
2520
  3. `rig-parts layers inputs/layers/full` — read it. A refusal here is about the
2451
2521
  files, not the art (§6, *Reading the inputs*). A `WARN` line is about the art: a
2452
2522
  layer See-through made that is not plausibly part of the figure, which step 6
2453
2523
  leaves out (§6, *Plausibility*). It refuses nothing.
2454
- 4. `rig-parts propose --head-box --full inputs/layers/full --canvas <W>x<H>` →
2455
- `seethrough.head_box`. Run `rig-parts inputs` again: it now also writes
2524
+ 4. `rig-parts propose --head-box --full inputs/layers/full --canvas <W>x<H> --write-config config.json`
2525
+ saves `seethrough.head_box` in the existing config. `<W>x<H>` is the original
2526
+ painting's size, not the padded square's. The existing optional `--no-back-hair`
2527
+ leaves the back-hair layer out of the head-box proposal; without it that layer
2528
+ contributes to the box. Run `rig-parts inputs` again: it now also writes
2456
2529
  `st_input_head.png`, the painting cropped to that box at its exact size.
2457
2530
  5. **See-through, head run** (external), on that crop. See-through's own second,
2458
2531
  head stage does not replace it: that stage's layers are pasted back at the full
2459
2532
  run's scale on every route that was read (README *Getting See-through layers*),
2460
2533
  so the full run's eyes stay at its density — 22×13 px for reference-demo's right eye
2461
2534
  white at `resolution` 1024, against 60×38 px in the head run.
2462
- 6. `rig-parts assemble --propose-plan --source … --full … --head … --config config.json`
2463
- → paste `plan` and `extend_below_crop` into `config.assemble`. Config: the
2535
+ 6. `rig-parts assemble --propose-plan --source … --full … --head … --config config.json --write-config config.json`
2536
+ saves `assemble.plan` and `assemble.extend_below_crop`. Config: the
2464
2537
  `layers` row, with `head_box`.
2465
2538
  Read `notes`: every layer not planned — each tag a layer of either run holds that no
2466
2539
  row takes — is named there once, `<tag>: in no plan row — …`, with why, run by run: the
@@ -2491,8 +2564,8 @@ full loader before its assemble stage writes anything.
2491
2564
  7. `rig-parts assemble … --out work` — read its `uncovered hole` and `uncovered:` lines and look at
2492
2565
  `work/render/recomposite_error_rig.png` (§5): red is painting no part holds, which
2493
2566
  no later stage can see — then
2494
- `rig-parts propose --parts work/rig --source inputs/painting.png --out work` →
2495
- `work/proposal.json` and `work/render/landmarks.png` (+ `_head`). Copy
2567
+ `rig-parts propose --parts work/rig --source inputs/painting.png --out work --write-config config.json` →
2568
+ `work/proposal.json` and `work/render/landmarks.png` (+ `_head`). The flag saves
2496
2569
  each of `bones`, `meshes`, `regions`, `limbs`, `motion`, `constraints`, `turn` and `proposed_rig_scale` it
2497
2570
  wrote into the config: `limbs` is written only with `--keypoints`, `constraints`
2498
2571
  only with `--cloth` and `turn` only with `--full`/`--head` (`PROPOSAL_CONFIG_SECTIONS` in `src/propose.ts` is the list,
@@ -2508,6 +2581,13 @@ full loader before its assemble stage writes anything.
2508
2581
  `[bone, from, to]` segment (the off-art check reads only chain links), a
2509
2582
  region, a mesh that names no part, a torso bone absent, a chain the joints
2510
2583
  declare no line for — and a `coverage:` summary counts the three apart.
2584
+ Before the totals, `coverage leg-chain choice:` reports the `leg_r` and
2585
+ `leg_l` chains declared by the supplied proposal or config, their count and
2586
+ recorded source/part bindings, or why none was selected. A missing source is
2587
+ `not recorded`, a chain with no bound parts says `none bound`; nothing is
2588
+ inferred to fill either. The line is selection only, not a checked verdict.
2589
+ Read it and the per-part notes again when changing keypoint sources: a bone
2590
+ map for two leg chains does not cover a proposal that bound only one.
2511
2591
  A `not checked` bone is one nothing verified: look at it on
2512
2592
  `landmarks.png`. `work/basis.json` (spec `spine-parts-basis/1`) says, per
2513
2593
  proposed bone, what its origin and its stated tip rest on: `joint` (which,
@@ -2637,9 +2717,13 @@ full loader before its assemble stage writes anything.
2637
2717
  relative `motion.animations_from` or `generation.control.skeleton` keeps its
2638
2718
  text, so it resolves against `draft/`; the command says so.
2639
2719
 
2640
- The selftest runs this order, with the README's flags, on each fetched example from
2641
- a config holding only `key`, `seethrough` and `assemble.rig_scale`, pasting each
2642
- proposal in as the steps above say (`RL01`); every step must exit 0.
2720
+ The command-order selftest runs these stages on each fetched example from a
2721
+ config holding only `key`, `seethrough` and `assemble.rig_scale` (`RL01`). For
2722
+ the proposal step, copy `bones`, `meshes`, `regions`, `limbs`, `motion`,
2723
+ `constraints`, `turn` and `proposed_rig_scale` into the config, each section the
2724
+ proposal wrote. `RL01` exercises that manual route and does not execute
2725
+ `--write-config`. The save forms above write the fields owned by their mode;
2726
+ a saved draft still needs validation.
2643
2727
 
2644
2728
  ### Counting the triangles before a build (`mesh-stats`, issue #566)
2645
2729
 
@@ -2796,6 +2880,43 @@ file's figure, and every part is the clean file's but `hair_back`, which keeps a
2796
2880
  one of them occluded, so nothing shows it at rest; paint it there if the face will
2797
2881
  move off it (`AS245`).
2798
2882
 
2883
+ **Saving proposed config fields (`--write-config`, issue #356).** The flag is a
2884
+ valued destination path. Its modes have different ownership:
2885
+
2886
+ | proposal mode | destination | fields written |
2887
+ | --- | --- | --- |
2888
+ | `propose --head-box` | existing regular config | `seethrough.head_box` |
2889
+ | `assemble --propose-plan` | existing regular config; must name `--config` itself | `assemble.plan`, `assemble.extend_below_crop` |
2890
+ | ordinary `propose --parts` without `--cloth` | existing regular config | each present `PROPOSAL_CONFIG_SECTIONS` value: `bones`, `meshes`, `regions`, `limbs`, `motion`, `constraints`, `turn`, `proposed_rig_scale` |
2891
+ | `propose --cloth <input.json>` | new file | the recipe's edits spliced into the cloth config; §3, *The config written*, gives entry ownership, hand-edit conflicts and `--from <previous.json>` |
2892
+
2893
+ For the first three modes, each selected field replaces its whole value; other
2894
+ fields and optional proposal sections not emitted are retained. Changed whole
2895
+ values are spliced into the original text through `splicedEdits`, preserving
2896
+ untouched bytes. `notes` stays in the proposal. The stage's loader validates a same-directory candidate before
2897
+ replacement, so relative file references keep their base and invalid retained
2898
+ references are refused. A scale mismatch retains `CONFIG_PROPOSED_RIG_SCALE`;
2899
+ the writer does not change `assemble.rig_scale` to hide it. An unchanged update
2900
+ keeps the original bytes and mtime. A final source-change check precedes atomic
2901
+ replacement; this is a config-file commit, not a transaction over every artifact
2902
+ of the proposing command. A refusal leaves the target config unchanged. The
2903
+ writer lists changes on stderr, keeping the existing head-box and plan JSON on
2904
+ stdout. Without the flag, no config is changed. `--from-config` proposes nothing
2905
+ and refuses this flag; `--from` belongs to the cloth mode only.
2906
+ All four modes use one file-writing adapter. Fresh cloth output keeps its
2907
+ already-checked splice text and intentional partial-config contract; the full
2908
+ loader is not imposed on it. Publication links the complete same-directory
2909
+ staged file into a destination that must still be absent, so an intervening
2910
+ creator is refused without overwrite. Removing the staged temporary happens
2911
+ after publication: if that cleanup fails, `CONFIG_WRITE_FAILED` may accompany
2912
+ an already-published complete config. Inspect both named paths before retrying. A different
2913
+ `--write-config` destination for `assemble --propose-plan` is a usage error.
2914
+ For ordinary and cloth proposals, the destination must be separate from all four
2915
+ artifacts under `--out`: `proposal.json`, `basis.json`, `render/landmarks.png` and
2916
+ `render/landmarks_head.png`. Matching paths, directory-symlink aliases and
2917
+ existing hardlinks are refused before artifact output; the ordinary mode names
2918
+ `CONFIG_WRITE_TARGET`, and the cloth mode `PROPOSE_WRITE_CONFIG_FRESH`.
2919
+
2799
2920
  ## 5. After each stage: what to read
2800
2921
 
2801
2922
  | stage | read | green looks like | a common red |
@@ -3301,9 +3422,12 @@ See-through with another seed.
3301
3422
 
3302
3423
  | rule | means | change |
3303
3424
  | --- | --- | --- |
3425
+ | <a id="CONFIG_WRITE_TARGET"></a>`CONFIG_WRITE_TARGET` | ordinary `propose … --write-config` would overwrite its config through one of its own four artifact paths under `--out`: `proposal.json`, `basis.json`, `render/landmarks.png`, `render/landmarks_head.png`; the collision check resolves existing path ancestors and compares existing file identity, so directory-symlink aliases and hardlinks count too. Refused before artifact output; the cloth mode names the same collision `PROPOSE_WRITE_CONFIG_FRESH` | choose a config outside all four artifact paths and their aliases |
3426
+ | <a id="CONFIG_WRITE_CHANGED"></a>`CONFIG_WRITE_CHANGED` | the existing destination no longer matches the snapshot read for `--write-config` when the writer is ready to commit: it is absent, no longer a regular file, or its permissions or text differ. The writer refuses to replace the current destination | keep the current config, reconcile any concurrent edit, and rerun against it |
3427
+ | <a id="CONFIG_WRITE_FAILED"></a>`CONFIG_WRITE_FAILED` | a proposed-value splice moved an unintended field, or a config-writing operation failed: reading, checking collisions, staging, writing proposal artifacts, publishing or replacing the destination, or removing the staged temporary. The message names the action and path. Cleanup failures retain the primary failure and name the stranded temporary. Fresh cloth publication precedes temporary cleanup, so a cleanup failure may leave the complete new config present; there is no transaction over every proposal artifact | inspect the named action, destination and temporary; restore access or permissions as needed; remove only the specifically reported stranded temporary, and inspect any published config before retrying |
3304
3428
  | <a id="CONFIG_FILE_PRESENT"></a><a id="CONFIG_IS_JSON"></a>`CONFIG_FILE_PRESENT`, `CONFIG_IS_JSON` | no such file, or not JSON; for a file the config names (`motion.animations_from`, `requirements`), the path is read relative to the config's directory, and the `motion.animations_from` line names the path as typed, that directory and what the two resolved to (issue #569) | `--config`; a file the config names: put it at the resolved path, or write the path from the config's directory |
3305
3429
  | <a id="CONFIG_KEY_KNOWN"></a><a id="CONFIG_KEY_RETIRED"></a><a id="CONFIG_FIELD_PRESENT"></a><a id="CONFIG_FIELD_TYPE"></a>`CONFIG_KEY_KNOWN`, `CONFIG_KEY_RETIRED`, `CONFIG_FIELD_PRESENT`, `CONFIG_FIELD_TYPE` | an unknown key (a retired one says what replaces it), a missing one (a `motion.animations_from` file with no `spec` is told to write `"spec": "rigc-motion/1"` and the whole file's shape, issue #569), a wrong type (`bones[i].link_inherit` must be an object whose values are strings); a project's own record under a plain key, or an annotation (`note`, `<name>_note`) holding something other than a string; in `meshes.<part>.auto` (§3, *An automatic mesh*), a missing or out-of-range number, a Stage B opt-in that is not exactly a value rig-c accepts, or a field the mode does not have — a bound on one of the five allocation rows, or an authored `motionAmplitude` (it is derived, never written); under `comfy paint`, `CONFIG_FIELD_PRESENT` on `config.generation`: a config with no `generation` block (the line names the fields it holds) | the named field; a record goes under a key beginning `x-` (any JSON, read by nothing), a remark under `note` or `<name>_note` (a string) — §3; under `comfy paint`, `config.generation` |
3306
- | <a id="CONFIG_KEY_UNIQUE"></a>`CONFIG_KEY_UNIQUE` | a key written twice in one object of a file the config names — the pose file `generation.control.skeleton` names, or the `motion.animations_from` file; JSON keeps only the last value, so every other would be dropped without a word | that file: each key once |
3430
+ | <a id="CONFIG_KEY_UNIQUE"></a>`CONFIG_KEY_UNIQUE` | a key written twice in an existing config updated by `--write-config`, or in one object of a file the config names — the pose file `generation.control.skeleton` names, or the `motion.animations_from` file; JSON keeps only the last value, so every other would be dropped without a word | that file: each key once |
3307
3431
  | <a id="CONFIG_ANIMATION_NAME_FREE"></a>`CONFIG_ANIMATION_NAME_FREE` | the `motion.animations_from` file holds an animation named `idle`, the animation `config.motion` writes | rename that animation |
3308
3432
  | <a id="CONFIG_HEAD_BOX_SQUARE"></a>`CONFIG_HEAD_BOX_SQUARE` | `seethrough.head_box` is not a non-empty square | `head_box` — take `propose --head-box`'s |
3309
3433
  | <a id="CONFIG_TAG_KNOWN"></a><a id="CONFIG_PART_NAME"></a><a id="CONFIG_PART_UNIQUE"></a>`CONFIG_TAG_KNOWN`, `CONFIG_PART_NAME`, `CONFIG_PART_UNIQUE` | a plan or extend entry names neither a v3 tag nor `author:<label>` (an `author:` tag with a malformed label says what the label must be), a part name that is not a file name, or a part or layer twice — a patch named like a plan part or another patch, and a cut's `into` named like a plan part, a patch or another cut's, included | `assemble.plan`, `assemble.patches`, `assemble.cuts` |
@@ -3439,7 +3563,7 @@ says why each requirement exists and what breaks without it.
3439
3563
  | <a id="PROPOSE_FIGURE_LAYER"></a>`PROPOSE_FIGURE_LAYER` | with `--keypoints` (issue #577): a limb is drafted from the figure layer (no handwear or legwear part over 500 px for it), and two or more `topwear` parts each hold every joint and the trunk bones' origins within 15 px with the same opaque px in the 31x31 squares about them, so nothing says which is the figure; or the only `topwear` parts that hold them are named by a cloth recipe (`config.cloth.<tag>.part` or `.parts`, issue #631), whose mesh a limb would take (§3, *A figure of one material*) | leave the others out of `assemble.plan`, or give the arms and legs their own handwear and legwear layers; for a recipe's part, take it out of the recipe |
3440
3564
  | <a id="PROPOSE_CLOTH_DRAFTED"></a>`PROPOSE_CLOTH_DRAFTED` | with `--cloth` (issue #364): the config's `cloth` block holds a recipe for a tag the proposer drafts no chain on — no part has that tag, or every `back hair` part ends above the shoulders (a bun bone, no chain). A `bottomwear` recipe with no `bottomwear` part names every `topwear` part shaped like a skirt (issue #402) | leave that recipe out, or assemble the layer it is for, or declare the named part in `cloth.bottomwear.part` |
3441
3565
  | <a id="PROPOSE_CLOTH_PART"></a>`PROPOSE_CLOTH_PART` | with `--cloth` (issues #402, #437): `cloth.bottomwear.part` (a string) names no part, a part or cut piece whose source is not `topwear` (a `bottomwear` part takes the recipe by its tag already), or a target whose box ends above the hip + 0.25 face heights, the topwear rule's waist line; or `cloth.<tag>.parts` (or `part` as a list, issue #437) names a part not in `parts.json`, a part no link of the drafted tree crosses (`spanCrosses == 0`), or parts out of root-to-tip depth order along the chain | name the `topwear` part or its cut skirt piece that holds the skirt (or leave `part` out), or list only parts in `parts.json` that the drafted tree links cross in root-to-tip order |
3442
- | <a id="PROPOSE_WRITE_CONFIG_FRESH"></a>`PROPOSE_WRITE_CONFIG_FRESH` | with `--write-config` (issue #537): the out path exists, or is a file the run writes itself (`proposal.json`, `basis.json` under `--out`); the config is written to a new file, never over the `--cloth` config or an earlier one | name a file that does not exist yet, beside the `--cloth` config |
3566
+ | <a id="PROPOSE_WRITE_CONFIG_FRESH"></a>`PROPOSE_WRITE_CONFIG_FRESH` | with `--cloth … --write-config` (issues #537, #356): the out path exists or collides with one of the four artifacts under `--out` — `proposal.json`, `basis.json`, `render/landmarks.png`, `render/landmarks_head.png` — including directory-symlink and hardlink aliases; the config is written to a new file, never over the `--cloth` config or an earlier one | name a file that does not exist yet, beside the `--cloth` config and separate from all four artifact paths and their aliases |
3443
3567
  | <a id="PROPOSE_WRITE_CONFIG_CONFLICT"></a>`PROPOSE_WRITE_CONFIG_CONFLICT` | with `--write-config` (issue #552): an entry of `bones`, `constraints` or `motion.tracks` the recipe re-drafts or removes was changed by hand in the `--cloth` config — it equals neither the draft with the recipe nor the draft without it — so the write would drop the hand edit; the object is the entry (`config.constraints[40] "skirt0a_0_physics"` in CL207), the detail names the first field the hand edit holds and what the draft writes there. With `--from <previous config>` (issue #571) an entry equal to the previous recipe's draft is the recipe's too, and the conflict is an entry that equals neither the draft, the draft without the recipe, nor the previous recipe's draft: the detail says so and reads the first differing field against the previous recipe's draft. Also a name the config states twice where the recipe drafts it. Every conflict is named at once and nothing is written | take the draft by removing the hand edit (or the entry), or leave the recipe out for that part (its `cloth` block); state a name once |
3444
3568
  | <a id="PROPOSE_WRITE_CONFIG_SPLICE"></a>`PROPOSE_WRITE_CONFIG_SPLICE` | with `--write-config` (issue #537): the `--cloth` config's text with the drafted sections spliced in does not parse to that config with them set, and the message names the first key that moved — the text states a key twice, so the splice lands on the first and the parse keeps the last | state each key once in the `--cloth` config |
3445
3569
  | <a id="PROPOSE_CLOTH_BACK_HAIR_CHAINS"></a>`PROPOSE_CLOTH_BACK_HAIR_CHAINS` | with `--cloth` (issues #390, #445, #461): `cloth."back hair".chains` exceeds what the narrower side's opaque width at the eye-line root row holds at the part's mesh grid (`max(1, min(floor(w_r / grid), floor(w_l / grid)))`), or places a tree root outside the 15 px dilated art mask of the `back hair` part (e.g. an inner twin-tail root over the face where the hair has no pixel); the message names the part, each side's root-row width, the grid spacing, and any off-art root with its pixel alpha | lower `cloth."back hair".chains` to at most the max fit named in the message, or omit `chains` so `propose --cloth` derives the tree count per visible hair mass on each side from its root-row width |
@@ -3486,6 +3610,12 @@ says why each requirement exists and what breaks without it.
3486
3610
  | <a id="CONFIG_LAYOUT_TARGET"></a>`CONFIG_LAYOUT_TARGET` | `config.limbs.<name>.layout` declares both or neither `before`/`after`, or a target declares both or neither `limb`/`part` | write one immediate placement side and one typed named target |
3487
3611
  | <a id="CONFIG_LAYOUT_NAME"></a>`CONFIG_LAYOUT_NAME` | a layout names no declared target limb or produced part; a consumed source layer is not an emitted part | use the declared limb or produced segment/part names printed in the detail |
3488
3612
  | <a id="CONFIG_LAYOUT_SCOPE"></a>`CONFIG_LAYOUT_SCOPE` | layout `parts` is empty, duplicated, nonconsecutive, foreign, or names a rider instead of an own source-bound segment part | select distinct consecutive own segment-part references; inherited riders are proven separately |
3613
+ | <a id="CONFIG_FRONT_SCOPE"></a>`CONFIG_FRONT_SCOPE` | `config.limbs.<name>.front.parts` (issue #664) is empty, repeats a part, or names a part that is not one of this limb's own segment parts | list this limb's own segment parts, each once |
3614
+ | <a id="CONFIG_FRONT_NAME"></a>`CONFIG_FRONT_NAME` | `config.limbs.<name>.front.after` (issue #664) names no produced part (a consumed source layer is not one), or names a segment part of this same limb; or a part `front.parts` names would take the slot name `<part>_front` of a part the config already makes | name a produced part the limb is drawn behind, from the list in the detail; or rename the part holding that name |
3615
+ | <a id="CONFIG_FRONT_LAYOUT"></a>`CONFIG_FRONT_LAYOUT` | a config declares `config.limbs.<name>.front` (issue #664) while some limb declares `layout` (issue #647); version 1 does not compose the two | drop one of them on this rig |
3616
+ | <a id="CONFIG_SET_FRONT"></a>`CONFIG_SET_FRONT` | `assemble.alternates[i].front` (issue #664) is empty, repeats a part, or names a part the set does not cover (in its `parts`, not its `hides`) or that no `config.limbs.<limb>.front.parts` lists; the detail names the part and which held | cover the part in the set's `parts`, and declare its front slot on its limb |
3617
+ | <a id="CONFIG_SET_FRONT_KEYED"></a>`CONFIG_SET_FRONT_KEYED` | `motion.animations_from` (issue #664): a `{"t", "set"}` key on a slot attachment track names a set that draws the track's part on its front slot; that would change the draw order mid-clip, which is out of scope | show the set with `setSet`, or key a part the set draws on its base slot |
3618
+ | <a id="RIG_SET_FRONT_BOTH"></a>`RIG_SET_FRONT_BOTH` | a `parts.json` alternate record (issue #664) maps both a part's slot and its front slot `<part>_front`; the detail names the set and the part | re-run `assemble`, which maps one or the other |
3489
3619
  | <a id="RIG_LAYOUT_NONEXCLUSIVE"></a>`RIG_LAYOUT_NONEXCLUSIVE` | a requested segment family/target, or an accepted replay's bindings, span owners or leave the selected segment scope across base/alternate attachments | correct the named bindings or scope; mixed ownership is never forced and no stale proof is emitted |
3490
3620
  | <a id="RIG_LAYOUT_UNREPRESENTABLE"></a>`RIG_LAYOUT_UNREPRESENTABLE` | the named slot family is absent, a target is not contiguous, the accepted slot population/adjacency changed, or final rig text is stale | use an emitted contiguous target and the exact final accepted source; unrelated slots are not absorbed |
3491
3621
  | <a id="RIG_LAYOUT_CONFLICT"></a>`RIG_LAYOUT_CONFLICT` | layout declarations share a slot, target themselves, or form a reciprocal/chained pair; both field paths are printed | declare one disjoint pair, with each participating slot in only that placement |
package/docs/STABILITY.md CHANGED
@@ -37,7 +37,10 @@ this version does not implement. No command is unimplemented at 1.0.0 (`LATER`
37
37
  in `cli.ts` is empty). A refusal line has the shape
38
38
  ` FAIL <RULE>: <object> — <value found>; <value required> (see rig-parts guide <RULE>)`:
39
39
  the rule name is a stable identifier, the pointer at the end is fixed text, and the
40
- detail between them is prose that may be reworded. A command
40
+ detail between them is prose that may be reworded. Existing named or child core
41
+ guide pointers in relayed compile-error text are preserved verbatim. Only a bare
42
+ core compile-error line receives `(see rigc guide §5.1)` inside the enclosing
43
+ rig-parts refusal; no core error code is invented. A command
41
44
  also answers to the alias `spine-parts` as long as the install keeps it
42
45
  ([MIGRATION.md](MIGRATION.md)). `-h`, `help` and `-v` are the short forms of
43
46
  `--help` and `--version`. The flags' meanings are `rig-parts --help`'s; their
@@ -51,9 +54,9 @@ defaults are the help's too, and none is guessed from the art.
51
54
  | `inputs` | `--source`, `--config`, `--out` |
52
55
  | `comfy` | `paint`: `--config`, `--out`; optional `--seeds`, `--seed0`, `--host`, `--wait`, `--timeout`, `--poll` |
53
56
  | `comfy` | `seethrough`: `--image`, `--out`; optional `--resolution`, `--steps`, `--seed`, `--prefix`, `--offload`, `--lama`, `--nf4`, `--host`, `--wait`, `--timeout`, `--poll` |
54
- | `assemble` | `--propose-plan` with `--source`, `--full`, `--config`, optional `--head`; or `--source`, `--full`, `--config`, `--out`, optional `--head` (without it, one run: issue #359), `--seam` (near-white, silhouette), `--project` (core, visible), `--alt` (repeatable `<name>=<source>:<full>[:<head>]`, issue #454) |
55
- | `propose` | `--head-box` with `--full`, `--canvas`; optional `--no-back-hair` |
56
- | `propose` | `--parts`, `--source`, `--out`; optional `--compare` or `--from-config` (one of them), `--keypoints`, `--person` and `--wrist-snap` (each with `--keypoints`), `--full`, `--head` (the turn control's depths; not with `--from-config`), `--grid-rule` (`figure` or `part`, the drafted lattice's spacing; not with `--from-config`), `--cloth` (a config whose `cloth` block is the recipe, issue #364; not with `--from-config`), `--write-config` (a new file, with `--cloth` and not with `--from-config`, issue #537 — additive: without it nothing `propose` writes changes), `--from` (the config the recipe is changed from, only with `--write-config`, issue #571 — additive: without it nothing `--write-config` writes changes) |
57
+ | `assemble` | `--propose-plan` with `--source`, `--full`, `--config`, optional `--head`, `--write-config` (must name `--config` itself, issue #356); or `--source`, `--full`, `--config`, `--out`, optional `--head` (without it, one run: issue #359), `--seam` (near-white, silhouette), `--project` (core, visible), `--alt` (repeatable `<name>=<source>:<full>[:<head>]`, issue #454) |
58
+ | `propose` | `--head-box` with `--full`, `--canvas`; optional `--no-back-hair`, `--write-config` (existing config, issue #356) |
59
+ | `propose` | `--parts`, `--source`, `--out`; optional `--compare` or `--from-config` (one of them), `--keypoints`, `--person` and `--wrist-snap` (each with `--keypoints`), `--full`, `--head` (the turn control's depths; not with `--from-config`), `--grid-rule` (`figure` or `part`, the drafted lattice's spacing; not with `--from-config`), `--cloth` (a config whose `cloth` block is the recipe, issue #364; not with `--from-config`), `--write-config` (existing config for an ordinary proposal, issue #356; a new file with `--cloth`, issue #537; not with `--from-config` — additive: without it nothing `propose` writes changes), `--from` (the config the recipe is changed from, only with `--write-config`, issue #571 — additive: without it nothing `--write-config` writes changes) |
57
60
  | `compare` | `--left`, `--right`; optional `--map` |
58
61
  | `rig` | `--config`, `--parts`, `--out`; optional `--idle-keys` (ctl, direct), `--page-edges` (pot, free), `--pack-shape` (rect, polygon) |
59
62
  | `check` | `--rig`, `--out`; optional `--parts`, `--source`, `--page-edges`, `--pack-shape`, `--requirements`, `--seam-pairs`, `--keep-stills`, `--set` or `--all-sets` (issue #454), `--environment` (issue #538, additive) |
@@ -124,13 +127,14 @@ part name, `[i]` an entry of a list):
124
127
  - **seethrough** — required: `resolution`, `steps`, `seed`, `offload`; optional: `head_box`, `arm_split`
125
128
  - **seethrough.arm_split** — required: `order`, and one of `x` or `line` (issue #439)
126
129
  - **assemble** — required: `rig_scale`, `plan` (from plain `assemble` on); optional: `extend_below_crop`, `patches`, `cuts`, `alternates`, `head_face_underfill`
127
- - **assemble.alternates[i]** — required: `name`; optional: `source`, `full`, `head`, `parts`, `hides`, `head_box`, `cuts`, `skeleton`, `drawn`, `tips` (issues #454, #465, #652)
130
+ - **assemble.alternates[i]** — required: `name`; optional: `source`, `full`, `head`, `parts`, `hides`, `head_box`, `cuts`, `skeleton`, `drawn`, `tips`, `front` (issues #454, #465, #652, #664)
128
131
  - **assemble.extend_below_crop[i]** — required: `part`, `run`, `tag`
129
132
  - **assemble.patches[i]** — required: `name`, `box`, `alpha`, `draw`
130
133
  - **assemble.patches[i].draw**, when an object — required: `before`
131
134
  - **assemble.cuts[i]** — required: `from`, `into`, `polygon`, `draw`, `overlap`; optional: `under`, `under_fill`
132
135
  - **assemble.cuts[i].draw**, when an object — required: `before`
133
- - **limbs.&lt;limb&gt;** — required: `segments`; optional: `split`, `layout` (issues #349, #446, #647)
136
+ - **limbs.&lt;limb&gt;** — required: `segments`; optional: `split`, `layout`, `front` (issues #349, #446, #647, #664)
137
+ - **limbs.&lt;limb&gt;.front** — required: `parts` (own segment parts, issue #664), `after` (a produced part)
134
138
  - **limbs.&lt;limb&gt;.layout** — exactly one of `before`, `after`; optional: `parts` (own consecutive segment-part references, issue #647)
135
139
  - **limbs.&lt;limb&gt;.layout.before**, **limbs.&lt;limb&gt;.layout.after** — exactly one typed target: `limb` or `part`
136
140
  - **limbs.&lt;limb&gt;.segments[i]** — required: `part`, `from`, `bone`; required below a joint (i > 0): `joint`; optional below a joint: `range`, `cap`, `edge`, `cap_shape`, `cut_band`
@@ -201,7 +205,8 @@ every number) are AUTHORING §3's tables, and the loader refuses any other by na
201
205
  | the template `.psd` | `layers --template` | one empty layer per name `templateLayerNames` (`src/layers.ts`) holds, bottom first; the same canvas writes the same bytes |
202
206
  | `proposal.json`, `render/landmarks*.png` | `propose` | the config's four rig sections, `constraints` with `--cloth` (issue #364: a physics constraint per tree link named `<bone>_physics`, a dangling end's `<bone>_translate`, and a skirt swing's transforms `<swing>_<tree>`), `turn` with `--full`/`--head`, `proposed_rig_scale` (issue #354), and `notes` |
203
207
  | `basis.json` | `propose` | spec `spine-parts-basis/1` |
204
- | the `--write-config` file | `propose --cloth … --write-config` | the `--cloth` config read, byte for byte, but for the entries the recipe drafts (issue #537), measured against the same parts drafted without it: `meshes.<part>` and `regions.<part>` for each part the recipe moves, and `bones`, `constraints` and `motion.tracks` entry by entry by name (issue #552) — an entry the recipe drafts replaced or inserted, one it replaces removed, every other entry kept with its bytes and its place, and a hand edit to an entry the recipe owns refused by name (`PROPOSE_WRITE_CONFIG_CONFLICT`) rather than dropped; a section the input lacks is inserted; with `--from <previous config>` (issue #571) an entry equal to the previous recipe's draft is the recipe's as well |
208
+ | an existing config named by `--write-config` | `propose --head-box`, `assemble --propose-plan`, ordinary `propose` | issue #356, additive: `assemble --propose-plan` must update `--config` itself; only the selected proposed fields are replaced by whole-value text splices, preserving untouched bytes; absent optional proposal sections and other fields remain; the stage loader validates before atomic replacement, relative paths keep their base, an unchanged update preserves bytes and mtime, and stderr lists changes; no multi-artifact transaction is promised |
209
+ | a new `--write-config` file | `propose --cloth … --write-config` | complete checked splice text published without overwrite; an intervening creator is refused, while cleanup failure after publication may leave the complete new file present; intentional partial configs retain their existing contract; the `--cloth` config read, byte for byte, but for the entries the recipe drafts (issue #537), measured against the same parts drafted without it: `meshes.<part>` and `regions.<part>` for each part the recipe moves, and `bones`, `constraints` and `motion.tracks` entry by entry by name (issue #552) — an entry the recipe drafts replaced or inserted, one it replaces removed, every other entry kept with its bytes and its place, and a hand edit to an entry the recipe owns refused by name (`PROPOSE_WRITE_CONFIG_CONFLICT`) rather than dropped; a section the input lacks is inserted; with `--from <previous config>` (issue #571) an entry equal to the previous recipe's draft is the recipe's as well |
205
210
  | `config.json`, `mesh_draft.json` | `mesh-draft` | `config.json` is the config read, byte for byte, but for each drafted part's `meshes` entry; `mesh_draft.json` is spec `rig-parts-mesh-draft/1`: `rules`, `lattice_check`, `draft_check`, `figure_overdraw`, and per lattice part `part`, `mode`, `lattice`, `candidate` (with `bridges` — `units`, `joins` of `span` and `area_px2`, `area_px2`, `overdraw`, and `faint` only when an island was left out as faint — only when the candidate is a bridged one, issue #326), `delta` (under `--more-triangles` only), `readings`, `gate` (each line's `reading`, `pass`, `text`, `margin`), `note`; and `not_compared` (issue #307), only when the config holds an `auto` or `contour` entry: per such entry, in config order, `part`, `mode` and `note` — the part was neither drafted nor compared, and a config of lattices alone writes no such key |
206
211
  | `rig.json`, `motion.json`, `mesh_report.json`, `images/`, `remesh.json` | `remesh` | the input revision's files, `rig.json` and `mesh_report.json` with the part's attachment and row replaced (under `--delta` the row is the input's with its counts, `bones` and influences recomputed and `mode` "delta"), every other file byte for byte; `remesh.json` is spec `rig-parts-remesh/1`: `request` (`part`, `mesh`, `mesh_sha256`, `idle_keys`, `page_edges`, `pack_shape`; under `--delta`, issue #526: `part`, `delta` — the mesh edit as read — `delta_sha256`, `ops`, `applied` — `moved`, `added`, `removed`, `constrained` — `idle_keys`, `page_edges`, `pack_shape`), `versions` (`parts`, `rigc`), `input` (`rig` — `path`, `bytes`, `sha256` per file — `config_sha256`, `parts_json_sha256`), `baseline` (`reproduced`, `not_compared`; under `--delta` `reproduced` holds one sentence: no config is built, the input is the baseline), `target` (`before`, `after`: `vertices`, `triangles`, `hull`), `preserved` (`identical`, `spliced`), `revision` (per file as `input.rig`), `check` (`input`, `revision`: `pass`, `measured`, `not_met`, `skipped`, `physics`, then `line` and `status` — issue #541, additive, the spec stays `rig-parts-remesh/1`: `line` is the last line `check` prints on that tree, byte for byte, and `status` every block's status as that tree's `check.json` records it, by name in the file's order — `gate`, `loop` and `seam` read from their figures, every other block its own `status`, a block the file does not write absent; `scope`) |
207
212
  | `rig.json`, `motion.json`, `mesh_report.json`, `images/`, `retune.json` | `retune` | the input revision's files, `rig.json` with `constraints` and `invariants.detached` replaced and `motion.json` with `animations.idle` replaced by the config's build, every other file byte for byte; `retune.json` is spec `rig-parts-retune/1`: `request` (`idle_keys`, `page_edges`, `pack_shape`), `versions` (`parts`, `rigc`), `input` (`rig` — `path`, `bytes`, `sha256` per file — `config_sha256`, `parts_json_sha256`), `held` (`compared`, `not_compared`), `changed` (`constraints` — each `name`, `type`, `fields` — `added`, `removed`, `idle`: `tracks_changed`, `duration` as before and after), `preserved` (`identical`, `spliced`), `revision` (per file as `input.rig`), `check` (`input`, `revision`: `pass`, `measured`, `not_met`, `skipped`, `physics`, then `line` and `status` — issue #541, additive, the spec stays `rig-parts-retune/1`: `line` is the last line `check` prints on that tree, byte for byte, and `status` every block's status as that tree's `check.json` records it, by name in the file's order — `gate`, `loop` and `seam` read from their figures, every other block its own `status`, a block the file does not write absent; `scope`) |
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@rig-c/parts",
3
- "version": "1.57.1",
3
+ "version": "1.59.0",
4
4
  "description": "AI-authored 2D character parts from one anime painting and its See-through layer decomposition: measured parts, weighted-mesh rig specs, verified through rig-c before anything is written. A CLI for agents that cannot see the image.",
5
5
  "type": "module",
6
6
  "bin": {
@@ -39,6 +39,7 @@
39
39
  "files": [
40
40
  "bin/rig-parts.cjs",
41
41
  "cli.ts",
42
+ "tools/configwrite.ts",
42
43
  "src",
43
44
  "workflows",
44
45
  "docs/AUTHORING.md",