@huukhanhnguyen/model / lanes/views/viewChain

lanes/views/viewChain#

VIEW_FILTER_METHOD#

const VIEW_FILTER_METHOD: "viewFilter"

VIEW_HEAD_METHODS#

const VIEW_HEAD_METHODS: Set<string>

The method names that may head a view chain, type-pinned to the lane's one vocabulary declaration (schema/vocabulary/laneMethods.ts, which the catalog derives its nodeType "view" entries from): a head naming a method that table does not declare stops compiling (step 7, 2026-07-27).

VIEW_METHODS#

const VIEW_METHODS: Set<string>

Every declared view method — heads AND steps, straight off the vocabulary table, so a method added to the catalog is legal here the same moment.

VIEW_PLACEABLE_METHODS#

const VIEW_PLACEABLE_METHODS: Set<string>

Paper placement accepts every HEAD; only a HEADLESS chain (a filter/style preset, which frames nothing) is unplaceable.

view3d joined the set 2026-07-27. It used to be excluded on the theory that a perspective bookmark cannot be drawn — but the drawing pipeline projects in PARALLEL from whatever direction the camera looks (views.ts's basisFromForward handles an arbitrary forward vector), and a parallel projection down an oblique direction is exactly an AXONOMETRIC drawing: the isometric that every drawing set carries next to its plans. So a placed view3d renders as an axonometric; its fov/zoom are ignored on paper, since a drawing has no perspective.

VIEW_STEP_METHODS#

const VIEW_STEP_METHODS: Set<string>

VIEW_STYLE_METHOD#

const VIEW_STYLE_METHOD: "viewStyle"

chainView()#

chainView(key, record?): View

Test/UI helper: build a chain View from a record-shaped description.

Parameters#

key#

string

record?#

camera?#

{ far?: number; fov?: number; height?: number; near?: number; position: Point; target: Point; up?: Point; }

camera.far?#

number

camera.fov?#

number

camera.height?#

number

camera.near?#

number

camera.position#

Point

camera.target#

Point

camera.up?#

Point

filter?#

ViewFilterRule[]

head?#

"viewPlan" | "viewSection" | "viewElevation" | "view3d" | "viewDetail"

label?#

string

style?#

string

Returns#

View

clearViewArg()#

clearViewArg(node, key): void

Parameters#

node#

Operation

key#

string

Returns#

void

findSectionObject()#

findSectionObject(objects, key): SectionObject | undefined

A section object by its objects[] key. Walks blocks. Falls back to the first section object in the scene — same contract as findSectionPlane.

Parameters#

objects#

readonly SceneObject[]

key#

string

Returns#

SectionObject | undefined

findSectionPlane()#

findSectionPlane(entries, key): SectionData | undefined

The scene's section plane — authoring-time + viewSection.plane resolve. Matches by the owning object's key (a SectionData payload alone carries no identity — see SectionPlaneRef's doc); falls back to the first section plane in the scene when no entry's key matches, same contract as findSectionObject. Same helper views.ts re-exports.

Parameters#

entries#

readonly SectionPlaneRef[]

key#

string

Returns#

SectionData | undefined

foldViewChain()#

foldViewChain(view, options?): FoldedView

Config-fold a views[] chain into a resolved definition.

Parameters#

view#

View

options?#

FoldViewOptions = {}

Returns#

FoldedView

isFlatView()#

isFlatView(view): view is FlatView

Parameters#

view#

FoldedView

Returns#

view is FlatView

isProjectionView()#

isProjectionView(view): view is ProjectionView

Parameters#

view#

FoldedView

Returns#

view is ProjectionView

isSectionView()#

isSectionView(view): view is SectionView

Parameters#

view#

FoldedView

Returns#

view is SectionView

isViewPlaceable()#

isViewPlaceable(view): boolean

True when the chain has a paper-placeable head (plan/section/elevation).

Parameters#

view#

View

Returns#

boolean

operationsFromRecordView()#

operationsFromRecordView(record): Operation[]

Inverse of the fold: a resolved view definition (head + pose + filter/style; crop and scale belong to the sheet placement, never to the view) back into the chain operations that produce it. The authoring UI and tests build fixtures through it, so a fixture can still think in projection terms.

head names the head method — it is NOT inferred from camera fields any more (the type tag it used to read is gone, and a pose alone cannot say whether an author meant a plan or a 3D bookmark). Everything else about the head is still derived from the pose: which principal elevation it faces, whether a plan is cut, where the section plane sits.

Parameters#

record#

camera?#

{ far?: number; fov?: number; height?: number; near?: number; position: Point; target: Point; up?: Point; }

camera.far?#

number

camera.fov?#

number

camera.height?#

number

camera.near?#

number

camera.position#

Point

camera.target#

Point

camera.up?#

Point

filter?#

ViewFilterRule[]

head?#

"viewPlan" | "viewSection" | "viewElevation" | "view3d" | "viewDetail"

style?#

string

Returns#

Operation[]

reasonOf()#

reasonOf(error): string

Exported so Model._foldViewOptions reports an unreadable view expression in the SAME words this module does — one gate, one message.

Parameters#

error#

unknown

Returns#

string

sectionDataFromObjects()#

sectionDataFromObjects(objects): SectionPlaneRef[]

Section-plane objects from the scene-object stream (not tessellated leaves), each paired with its objects[] key. viewSection.plane resolves against these (findSectionPlane).

Parameters#

objects#

readonly SceneObject[]

Returns#

SectionPlaneRef[]

sectionDepthOf()#

sectionDepthOf(view, evaluateArg?): number | undefined

Depth from the authoring head — not stored on the fold (one extra field).

Parameters#

view#

View

evaluateArg?#

(input) => unknown

Returns#

number | undefined

setViewArg()#

setViewArg(node, key, input): void

Parameters#

node#

Operation

key#

string

input#

string

Returns#

void

setViewStyleStep()#

setViewStyleStep(view, styleName): void

Set/replace the viewStyle step (last wins; creates one when missing).

Parameters#

view#

View

styleName#

string | undefined

Returns#

void

viewArgInput()#

viewArgInput(node, key): string | undefined

Arg get/set on a view operation node.

Parameters#

node#

Operation

key#

string

Returns#

string | undefined

viewExprJson()#

viewExprJson(value): string

Parameters#

value#

unknown

Returns#

string

viewExprNumber()#

viewExprNumber(n): string

Expression-literal helpers for authoring ops (UI + tests + migration).

Parameters#

n#

number

Returns#

string

viewExprPoint()#

viewExprPoint(p): string

Parameters#

p#

[number, number, number]

Returns#

string

viewExprString()#

viewExprString(s): string

Parameters#

s#

string

Returns#

string

viewExprWindow()#

viewExprWindow(w): string

Parameters#

w#

[number, number, number, number]

Returns#

string

viewHeadOf()#

viewHeadOf(view): Operation | undefined

First head operation in a view chain (skipping containers / disabled not checked).

Parameters#

view#

View

Returns#

Operation | undefined

viewStyleNameLiteral()#

viewStyleNameLiteral(view): string | undefined

Style name from the last viewStyle step (literal or any — no eval).

Parameters#

view#

View

Returns#

string | undefined

VIEW_HEAD_METHOD_NAMES#

Re-exports VIEW_HEAD_METHOD_NAMES