@lmzhen/dsh-evolution-learning-graph 0.3.23 → 0.3.25

package/lib/index.js

@@ -28,6 +28,24 @@ function graphDensity(graph) {
 	};
 }
 /**
+* Render one node's display line for `/evolution graph`.
+*
+* A skill node's id IS its name, so `● <name>` already addresses it — an
+* isolated skill (no edge) is addressable by name alone. A memory node's id
+* is the `memory:<source>:<index>:<snapshot>` string `buildLearningGraph`
+* emits, so the id is appended in brackets: an isolated memory node (no edge
+* carrying its id) is otherwise unaddressable from the render. The full id is
+* shown (not the bare `memory:<source>:<index>` prefix) so it matches the id
+* form the edge lines render and so the trailing snapshot — the E-21 token the
+* mutation commands use to reject a stale index — survives the copy/paste
+* (F-322). A bare prefix would round-trip through `parseGraphNodeId` but would
+* silently skip the E-21 drift guard in `readMemoryIndex`.
+*/
+function renderNodeLine(node) {
+	if (node.kind === "skill") return "● " + node.label;
+	return "◆ " + node.label + "  [id: " + node.id + "]";
+}
+/**
 * Build a small deterministic graph from usage records and memory entries.
 * Memory node ids follow `memory:<source>:<index>` (source = memory|user,
 * index = position in that file's entries) — the SAME rule the parser
@@ -224,7 +242,7 @@ function apply(ctx) {
 						related.set(name, relatedSkillNames(content, name));
 					}
 					const graph = buildLearningGraph(usageMap, memoryEntries, userEntries, related);
-					const lines = graph.nodes.map((node) => (node.kind === "memory" ? "◆" : "●") + " " + node.label);
+					const lines = graph.nodes.map((node) => renderNodeLine(node));
 					const edges = graph.edges.map((edge) => edge.from + " --" + edge.type + "--> " + edge.to);
 					const density = graphDensity(graph);
 					const densityLine = `\n\nSkills: ${density.skillNodes} · related edges: ${density.relatedEdges} (${density.edgesPerNode}/node) · isolated: ${density.isolatedPct}%`;
@@ -319,4 +337,4 @@ function apply(ctx) {
 	});
 }
 //#endregion
-export { apply, buildLearningGraph, graphDensity, memorySnapshotOf, name, parseGraphNodeId, readMemoryIndex, resolveGraphNode };
+export { apply, buildLearningGraph, graphDensity, memorySnapshotOf, name, parseGraphNodeId, readMemoryIndex, renderNodeLine, resolveGraphNode };

package/lib/types/index.d.ts

@@ -35,6 +35,21 @@ export interface GraphDensity {
     isolatedPct: number;
 }
 export declare function graphDensity(graph: LearningGraph): GraphDensity;
+/**
+ * Render one node's display line for `/evolution graph`.
+ *
+ * A skill node's id IS its name, so `● <name>` already addresses it — an
+ * isolated skill (no edge) is addressable by name alone. A memory node's id
+ * is the `memory:<source>:<index>:<snapshot>` string `buildLearningGraph`
+ * emits, so the id is appended in brackets: an isolated memory node (no edge
+ * carrying its id) is otherwise unaddressable from the render. The full id is
+ * shown (not the bare `memory:<source>:<index>` prefix) so it matches the id
+ * form the edge lines render and so the trailing snapshot — the E-21 token the
+ * mutation commands use to reject a stale index — survives the copy/paste
+ * (F-322). A bare prefix would round-trip through `parseGraphNodeId` but would
+ * silently skip the E-21 drift guard in `readMemoryIndex`.
+ */
+export declare function renderNodeLine(node: GraphNode): string;
 /**
  * Build a small deterministic graph from usage records and memory entries.
  * Memory node ids follow `memory:<source>:<index>` (source = memory|user,

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@lmzhen/dsh-evolution-learning-graph",
   "description": "Learning graph over skills and memory (community build)",
-  "version": "0.3.23",
+  "version": "0.3.25",
   "publishConfig": {
     "access": "public"
   },
@@ -30,8 +30,8 @@
   ],
   "license": "MIT",
   "dependencies": {
-    "@lmzhen/dsh-evolution-approval": "^0.3.23",
-    "@lmzhen/dsh-evolution-core": "^0.3.23"
+    "@lmzhen/dsh-evolution-approval": "^0.3.25",
+    "@lmzhen/dsh-evolution-core": "^0.3.25"
   },
   "peerDependencies": {
     "@deepseek-ai/dsh-invariants": "^0.1.1-rc.2",