refactor(builder): extract shared tree-flattener into craft-tree.ts

PageContext.treeToState, apply-ai-response.buildNodeTree, and the dead
serializeTreeForCraft each re-implemented the same ColumnLayout-linkedNodes +
SHELL_INNER walk and had drifted (buildNodeTree gained guards the others
lacked). Move sanitizeAiTree (resolvedName allowlist + id/ROOT repair) here
too and add flattenTreeForCraft, the shared structural walk (linkedNodes,
SHELL_INNER wrapping, style:[]->{} normalization) both real callers will
consolidate onto.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-07-12 14:04:43 -07:00
parent 71e675489c
commit b2b9278d3c
2 changed files with 338 additions and 1 deletions
+224
View File
@@ -1,5 +1,229 @@
import type { NodeTree, Node, NodeId } from '@craftjs/core';
import { getRandomId } from '@craftjs/utils';
import type { SerializedTreeNode } from '../types/sitesmith';
import { componentResolver } from '../components/resolver';
/** Only Container is a "real" Craft.js canvas in serialized state. Layout
* shells (Section/HeroSimple/ColumnLayout/etc) use <Element canvas> linkedNodes
* internally — their own node must serialize with isCanvas:false or
* toNodeTree's walker hits an Invariant because the shell claims to be a
* canvas but its render ignores `data.nodes`. */
export const CANVAS_TYPES = new Set(['Container']);
/** Shells that wrap their content in a single <Element id="<key>" is={Container}>.
* When a tree puts content directly under one of these, the children end up
* orphaned (the shell ignores data.nodes — it renders via the linkedNode) and
* Craft.js auto-creates the linkedNode at render time with a botched type
* field, which then crashes toNodeTree. Pre-create the linkedNode ourselves
* to keep the state shape Craft.js expects. */
export const SHELL_INNER: Record<string, string> = {
Section: 'section-inner',
BackgroundSection: 'bg-section-inner',
FormContainer: 'form-inner',
};
/** True if `name` is a registered component in the resolver (i.e. safe to
* hand to `actions.addNodeTree`/`actions.deserialize` without it throwing
* "does not exist in resolver" at render time). */
function isKnownResolvedName(name: unknown): name is string {
return typeof name === 'string' && Object.prototype.hasOwnProperty.call(componentResolver, name);
}
/**
* Validate + sanitize a tree (AI-supplied, or otherwise untrusted) before it
* is handed to `flattenTreeForCraft`/`query.parseFreshNode`/`actions.deserialize`.
* Fails SOFT — never throws:
*
* - **resolvedName allowlist**: any node whose `type.resolvedName` is not a
* key of `componentResolver` is dropped (it and its subtree), with a
* `console.warn`. If the tree's ROOT is itself invalid, this returns
* `null` and the caller must skip the whole op/tree (or substitute a safe
* fallback state).
* - **id policy — REGENERATE (not skip)**: a node id that is `'ROOT'`,
* empty/non-string, or collides with an id already in `usedIds` is
* replaced with a fresh `ai-auto-N` id (and a warning is logged). The
* node itself is kept — only structurally invalid *types* are dropped;
* a merely-bad *id* is repaired so we never lose otherwise-valid content
* over an id collision.
*
* `usedIds` should be seeded with any ids that must not collide (e.g. a live
* document's existing node ids); it is also used internally to keep ids
* unique within the tree being sanitized.
*/
export function sanitizeAiTree(
tree: SerializedTreeNode,
usedIds: Set<string> = new Set(),
idCounter: { n: number } = { n: 0 },
): SerializedTreeNode | null {
const resolvedName = tree?.type?.resolvedName;
if (!isKnownResolvedName(resolvedName)) {
console.warn(`sitesmith: dropping node with unknown component "${String(resolvedName)}"`);
return null;
}
const rawId = tree.props?.node_id;
let id = typeof rawId === 'string' ? rawId.trim() : '';
if (!id) {
id = `ai-auto-${idCounter.n++}`;
} else if (id === 'ROOT') {
console.warn('sitesmith: rejecting AI-supplied node id "ROOT" (reserved); regenerating');
id = `ai-auto-${idCounter.n++}`;
} else if (usedIds.has(id)) {
console.warn(`sitesmith: AI-supplied node id "${id}" collides with an existing id; regenerating`);
id = `ai-auto-${idCounter.n++}`;
}
// Belt-and-suspenders: keep drawing fresh ids in the (pathological) case
// the freshly generated id itself collides.
while (usedIds.has(id)) id = `ai-auto-${idCounter.n++}`;
usedIds.add(id);
const children: SerializedTreeNode[] = [];
for (const child of tree.nodes ?? []) {
const sanitizedChild = sanitizeAiTree(child, usedIds, idCounter);
if (sanitizedChild) children.push(sanitizedChild);
}
return {
type: tree.type,
props: { ...tree.props, node_id: id },
nodes: children,
};
}
/** A single flattened Craft.js serialized-node entry, as produced by
* `flattenTreeForCraft`. Shape matches `@craftjs/core`'s `SerializedNode`. */
export interface FlatCraftNode {
type: { resolvedName: string };
isCanvas: boolean;
props: Record<string, unknown>;
displayName: string | undefined;
custom: Record<string, unknown>;
hidden: boolean;
parent: string | null;
nodes: string[];
linkedNodes: Record<string, string>;
}
export interface FlattenTreeResult {
rootNodeId: string;
nodes: Record<string, FlatCraftNode>;
/** ids of nodes synthesized by the flattener itself (currently: the
* auto-created SHELL_INNER wrapper for Section/BackgroundSection/
* FormContainer) rather than present in the input tree. Callers that
* materialize real Craft.js nodes (e.g. via `query.parseFreshNode`) need
* to know which ids these are so they can be built directly instead of
* run through the resolver (which would merge in the component's default
* `craft.props`, changing the synthetic wrapper's props). */
syntheticIds: Set<string>;
}
/**
* Flatten a `SerializedTreeNode` tree into a flat map of Craft.js
* serialized-node entries, applying the transforms every caller in this
* codebase needs:
*
* - `style: []` -> `{}` (the AI sometimes emits an empty array instead of
* an empty object for a CSSProperties slot; React/Craft.js choke on that).
* - `ColumnLayout` children are moved from `nodes` into fixed `col-0,
* col-1, …` `linkedNodes` (ColumnLayout renders via linkedNodes, not
* direct children — direct children would be silently orphaned).
* - `Section`/`BackgroundSection`/`FormContainer` children are wrapped in a
* single pre-created `Container` linkedNode (see `SHELL_INNER`) so
* Craft.js never has to auto-materialize one at render time.
*
* This function assumes the input tree is already valid (every node's
* `type.resolvedName` is a registered component, and `props.node_id` is a
* unique, non-'ROOT' string where present) — callers should run untrusted
* trees through `sanitizeAiTree` first. Ids are read from `props.node_id`,
* falling back to an auto-generated id if absent.
*/
export function flattenTreeForCraft(tree: SerializedTreeNode): FlattenTreeResult {
let counter = 0;
const nodes: Record<string, FlatCraftNode> = {};
const syntheticIds = new Set<string>();
const walk = (node: SerializedTreeNode, parent: string | null): string => {
const id = (node.props?.node_id as string | undefined) || `ai-auto-${counter++}`;
const typeName = node.type?.resolvedName;
const rawProps = node.props ?? {};
const props: Record<string, unknown> = { ...rawProps };
if (Array.isArray(props.style)) props.style = {};
const childIds: string[] = [];
nodes[id] = {
type: node.type,
// isCanvas must match the component's craft.rules — only layout
// wrappers accept children. Setting it true on leaf components
// (Heading, TextBlock, ButtonLink, etc) makes Craft.js render them
// as empty drop-canvas wrappers and the actual content disappears.
isCanvas: typeName ? CANVAS_TYPES.has(typeName) : false,
props,
displayName: typeName,
custom: {},
hidden: false,
parent,
nodes: childIds,
linkedNodes: {},
};
for (const child of node.nodes ?? []) {
childIds.push(walk(child, id));
}
// ColumnLayout uses Craft.js linkedNodes with fixed ids (col-0, col-1, ...).
// Direct children would be orphaned by ColumnLayout's render (it creates
// fresh column Elements and ignores data.nodes) and any subsequent
// toNodeTree walk would hit an Invariant. Move them into linkedNodes so
// they render in the columns actually shown to the user.
if (typeName === 'ColumnLayout' && childIds.length > 0) {
const linked: Record<string, string> = {};
childIds.forEach((cid, i) => {
linked[`col-${i}`] = cid;
if (nodes[cid]) nodes[cid].isCanvas = true; // columns are canvases
});
nodes[id].nodes = [];
nodes[id].linkedNodes = linked;
const curProps = nodes[id].props;
if (!curProps.columns || curProps.columns !== childIds.length) {
curProps.columns = childIds.length;
}
}
// Section/BackgroundSection/FormContainer each render a single
// <Element id="<key>" is={Container} canvas> ... </Element>. If content
// was nested as direct children, Craft.js would auto-create the
// linkedNode on first render — storing its type as the Container
// component class rather than {resolvedName:'Container'}, which then
// crashes toNodeTree with "type (undefined) does not exist in resolver".
// Pre-create the linkedNode ourselves with the correct serialized type
// so Craft.js never has to materialize it.
const innerKey = SHELL_INNER[typeName ?? ''];
if (innerKey && childIds.length > 0) {
const innerId = `${id}__${innerKey}`;
nodes[innerId] = {
type: { resolvedName: 'Container' },
isCanvas: true,
props: { tag: 'div' },
displayName: 'Container',
custom: {},
hidden: false,
parent: id,
nodes: [...childIds],
linkedNodes: {},
};
syntheticIds.add(innerId);
for (const cid of childIds) {
if (nodes[cid]) nodes[cid].parent = innerId;
}
nodes[id].nodes = [];
nodes[id].linkedNodes = { [innerKey]: innerId };
}
return id;
};
const rootNodeId = walk(tree, null);
return { rootNodeId, nodes, syntheticIds };
}
/**
* Returns a deep-cloned NodeTree where every node id in the tree