fix tests: replace extractCtxSize with parseCtxMapFromYaml + extractCtxFromRunningCmd

- Replace the stale extractCtxSize test (imported a deleted function) with
  real tests for parseCtxMapFromYaml and extractCtxFromRunningCmd
- parseCtxMapFromYaml: 5 tests covering extraction, comments, top-level
  key reset, empty input, missing ctx-size
- extractCtxFromRunningCmd: 3 tests covering parsing, undefined, missing
  --ctx-size flag

tests/router-utils.test.ts

@@ -7,19 +7,102 @@
 import assert from "node:assert/strict";
 import { test } from "node:test";
 import {
-	extractCtxSize,
+	parseCtxMapFromYaml,
+	extractCtxFromRunningCmd,
 	isReasoningModel,
 	isShardArtefact,
 } from "../ai-server/router-utils.ts";
 
-// ── extractCtxSize (llama-swap: ctx-size not exposed via API) ──
+// ── parseCtxMapFromYaml ─────────────────────────────────────────────────
 
-// llama-swap does not expose worker command lines, so extractCtxSize
+test("parseCtxMapFromYaml: extracts ctx-size from model blocks", () => {
-// always returns null. Callers use a default (32768) instead.
+	const yaml = `
-test("extractCtxSize: always returns null for llama-swap models", () => {
+models:
-	assert.equal(extractCtxSize({ id: "Qwen_Qwen3.6-35B-A3B-Q8_0" }), null);
+  Qwen_Qwen3.6-35B-A3B-Q8_0:
-	assert.equal(extractCtxSize({ id: "MiniMax-M2.7-IQ3_XXS", running: true }), null);
+    cmd: |
-	assert.equal(extractCtxSize({ id: "x", object: "model", running: false }), null);
+      /home/ai-server/llama.cpp/build/bin/llama-server
+      --model /home/ai-server/models/Qwen_Qwen3.6-35B-A3B-Q8_0.gguf
+      --ctx-size 262144
+      --temp 0.7
+  MiniMax-M2.7-IQ3_XXS:
+    cmd: |
+      /home/ai-server/llama.cpp/build/bin/llama-server
+      --model /home/ai-server/models/MiniMax-M2.7-UD-IQ3_XXS.gguf
+      --ctx-size 131072
+      --temp 1.0
+`;
+	const map = parseCtxMapFromYaml(yaml);
+	assert.equal(map.get("Qwen_Qwen3.6-35B-A3B-Q8_0"), 262144);
+	assert.equal(map.get("MiniMax-M2.7-IQ3_XXS"), 131072);
+	assert.equal(map.size, 2);
+});
+
+test("parseCtxMapFromYaml: skips comments and blank lines", () => {
+	const yaml = `
+# This is a comment
+models:
+
+  # Model with large context
+  Qwen_Qwen3.6-35B-A3B-Q8_0:
+    cmd: |
+      /path/to/server
+      --ctx-size 65536
+      --temp 0.7
+`;
+	const map = parseCtxMapFromYaml(yaml);
+	assert.equal(map.get("Qwen_Qwen3.6-35B-A3B-Q8_0"), 65536);
+});
+
+test("parseCtxMapFromYaml: resets on top-level keys", () => {
+	const yaml = `
+models:
+  Qwen_Qwen3.6-35B-A3B-Q8_0:
+    cmd: |
+      /path/to/server
+      --ctx-size 262144
+hooks:
+  on_startup:
+    preload:
+      - Qwen_Qwen3.6-35B-A3B-Q8_0
+`;
+	const map = parseCtxMapFromYaml(yaml);
+	assert.equal(map.get("Qwen_Qwen3.6-35B-A3B-Q8_0"), 262144);
+	// "preload" is not a valid model id pattern, but even if it were,
+	// it's under hooks: so should not be included.
+	assert.ok(!map.has("preload"));
+});
+
+test("parseCtxMapFromYaml: empty yaml returns empty map", () => {
+	const map = parseCtxMapFromYaml("");
+	assert.equal(map.size, 0);
+});
+
+test("parseCtxMapFromYaml: model without ctx-size is skipped", () => {
+	const yaml = `
+models:
+  SmallModel:
+    cmd: |
+      /path/to/server
+      --temp 0.7
+`;
+	const map = parseCtxMapFromYaml(yaml);
+	assert.equal(map.get("SmallModel"), undefined);
+	assert.equal(map.size, 0);
+});
+
+// ── extractCtxFromRunningCmd ────────────────────────────────────────────
+
+test("extractCtxFromRunningCmd: parses --ctx-size from cmd string", () => {
+	const cmd = "/home/ai-server/llama.cpp/build/bin/llama-server --model /home/ai-server/models/Qwen.gguf --ctx-size 262144 --temp 0.7";
+	assert.equal(extractCtxFromRunningCmd(cmd), 262144);
+});
+
+test("extractCtxFromRunningCmd: undefined cmd returns null", () => {
+	assert.equal(extractCtxFromRunningCmd(undefined), null);
+});
+
+test("extractCtxFromRunningCmd: cmd without --ctx-size returns null", () => {
+	assert.equal(extractCtxFromRunningCmd("/path/to/server --temp 0.7"), null);
 });
 
 // ── isShardArtefact ─────────────────────────────────────────────────────