dsh-livebench-panel 0.1.12 → 0.1.14

package/lib/client.js

@@ -92,7 +92,7 @@ window.__ModuleLoader__.load({
 		function LiveBenchView() {
 			const [config, setConfig] = useState(null);
 			const [configError, setConfigError] = useState(null);
-			const [sel, setSel] = useState({ models: [], reasoning: "default", cats: [], tasks: [], release: "2024-11-25", begin: "", end: "", maxTokens: "32000" });
+			const [sel, setSel] = useState({ models: [], efforts: {}, cats: [], tasks: [], release: "2024-11-25", begin: "", end: "", maxTokens: "32000" });
 			const [modelDdOpen, setModelDdOpen] = useState(false);
 			const modelDdRef = useRef(null);
 			const [busy, setBusy] = useState(false);
@@ -171,10 +171,6 @@ window.__ModuleLoader__.load({
 				return () => clearInterval(timer);
 			}, [refreshStatus, loadResults]);
 
-			useEffect(() => {
-				if (logRef.current) logRef.current.scrollTop = logRef.current.scrollHeight;
-			}, [log]);
-
 			const categories = useMemo(() => (config ? Object.keys(config.tasks) : []), [config]);
 			// 有效题数：LiveBench 会丢弃「发布晚于所选 release」和「在所选 release
 			// 前已退役」的题目，这里按题目桶 (发布日, 移除日) 精确复算。
@@ -203,24 +199,14 @@ window.__ModuleLoader__.load({
 				const next = list.includes(value) ? list.filter((v) => v !== value) : [...list, value];
 				return { ...prev, [key]: next };
 			});
-			// 已选模型（provider::model 值列表）的解析与 effort 并集
+			// 已选模型（provider::model 值列表）的解析与 per-model effort
 			const selectedModelEntries = useMemo(() => sel.models.map((value) => {
 				const [pid, ...rest] = value.split("::");
 				const mid = rest.join("::");
 				const provider = config?.providers.find((p) => p.id === pid) ?? null;
 				const model = provider?.models.find((m) => m.id === mid) ?? null;
-				return { providerId: pid, modelId: mid, efforts: model?.efforts ?? [] };
+				return { providerId: pid, modelId: mid, modelValue: value, efforts: model?.efforts ?? [] };
 			}).filter((entry) => entry.modelId !== ""), [sel.models, config]);
-			const effortOptions = useMemo(() => {
-				const seen = new Set();
-				const union = [];
-				for (const entry of selectedModelEntries) {
-					for (const e of entry.efforts) {
-						if (!seen.has(e)) { seen.add(e); union.push({ value: e, label: e === "off" ? "off（关闭思考）" : e }); }
-					}
-				}
-				return union;
-			}, [selectedModelEntries]);
 
 			const onStart = async () => {
 				setStartError(null);
@@ -243,7 +229,8 @@ window.__ModuleLoader__.load({
 				}
 				setBusy(true);
 				try {
-					// 每个 selected 模型并发启动一个评测（服务端有并发上限保护）
+					// 每个 selected 模型并发启动一个评测（服务端有并发上限保护），
+					// 推理强度按各模型在次级下拉中的选择逐个下发
 					const failures = [];
 					const launches = selectedModelEntries.map(async (entry) => {
 						const payload = await api("/start", {
@@ -252,7 +239,7 @@ window.__ModuleLoader__.load({
 							body: JSON.stringify({
 								provider: entry.providerId,
 								model: entry.modelId,
-								reasoningEffort: sel.reasoning,
+								reasoningEffort: sel.efforts[entry.modelValue] ?? "default",
 								benchNames,
 								release: sel.release,
 								begin: sel.begin,
@@ -412,10 +399,8 @@ window.__ModuleLoader__.load({
 									: `已选 ${sel.models.length} 个模型`),
 							modelDdOpen && h("div", {
 								style: {
-									position: "absolute", zIndex: 30, top: "calc(100% + 4px)", left: 0, right: 0,
-									maxHeight: "320px", overflowY: "auto", border: "1px solid var(--dsw-alias-border-l2)",
+									marginTop: "6px", maxHeight: "340px", overflowY: "auto", border: "1px solid var(--dsw-alias-border-l2)",
 									borderRadius: "8px", background: "var(--dsw-alias-bg-layer-1)", padding: "6px",
-									boxShadow: "var(--dsw-shadow-lv1, 0 4px 16px rgba(0,0,0,.25))",
 								},
 							},
 								(config?.providers ?? []).map((p) => h("div", { key: p.id, style: { marginBottom: "4px" } },
@@ -424,26 +409,36 @@ window.__ModuleLoader__.load({
 									p.models.map((m) => {
 										const value = p.id + "::" + m.id;
 										const checked = sel.models.includes(value);
-										return h("label", {
-											key: value,
-											style: { display: "flex", alignItems: "center", gap: "6px", padding: "3px 4px", cursor: "pointer", borderRadius: "6px" },
-										},
-											h("input", { type: "checkbox", checked, onChange: () => {
-												setSel((prev) => {
-													const set = new Set(prev.models);
-													if (set.has(value)) set.delete(value); else set.add(value);
-													return { ...prev, models: [...set] };
-												});
-											} }),
-											h("span", { style: { fontSize: "12.5px" } }, m.name));
+										return h("div", { key: value, style: { borderBottom: "1px solid var(--dsw-alias-border-l2)", padding: "3px 4px" } },
+											h("div", {
+												style: { display: "flex", alignItems: "center", gap: "8px", cursor: "pointer" },
+												onClick: () => {
+													setSel((prev) => {
+														const set = new Set(prev.models);
+														const nextEfforts = { ...prev.efforts };
+														if (set.has(value)) { set.delete(value); delete nextEfforts[value]; }
+														else set.add(value);
+														return { ...prev, models: [...set], efforts: nextEfforts };
+													});
+												},
+											},
+												h("input", { type: "checkbox", checked, readOnly: true, style: { cursor: "pointer" } }),
+												h("span", { style: { fontSize: "12.5px", flex: 1 } }, m.name)),
+											checked && m.efforts.length > 0 && h("div", { style: { display: "flex", alignItems: "center", gap: "6px", padding: "4px 0 2px 26px" } },
+												h("span", { style: { fontSize: "11px", color: "var(--dsw-alias-label-tertiary)", whiteSpace: "nowrap" } }, "推理强度"),
+												h("select", {
+													className: c("select"), style: { height: "28px", fontSize: "12px", flex: 1 },
+													value: sel.efforts[value] ?? "default",
+													onClick: (event) => event.stopPropagation(),
+													onChange: (event) => {
+														const v = event.target.value;
+														setSel((prev) => ({ ...prev, efforts: { ...prev.efforts, [value]: v } }));
+													},
+												},
+													h("option", { value: "default" }, "（模型默认）"),
+													m.efforts.map((e) => h("option", { key: e, value: e }, e === "off" ? "off（关闭思考）" : e)))));
 									}))),
 						),
-						h("div", { className: c("field") },
-							h("label", { className: c("label") }, "推理强度"),
-							h("select", { className: c("select"), value: sel.reasoning, onChange: setField("reasoning"), disabled: effortOptions.length === 0 },
-								h("option", { value: "default" }, effortOptions.length === 0 ? "（按模型默认）" : "（模型默认）"),
-								effortOptions.map((e) => h("option", { key: e.value, value: e.value }, e.label))),
-						),
 						h("div", { className: c("field") },
 							h("label", { className: c("label") }, "题集 release"),
 							h("select", { className: c("select"), value: sel.release, onChange: setField("release") },

package/package.json

@@ -1,43 +1,43 @@
-{
-  "name": "dsh-livebench-panel",
-  "version": "0.1.12",
-  "description": "DSH web plugin: a LiveBench tab in the Trajectory view (right of 对话/轨迹). Run LiveBench evaluations against every model configured in the DeepSeek Harness — pick provider/model, category, task, release and question range from dropdowns, watch progress, and read scores in place.",
-  "license": "MIT",
-  "type": "module",
-  "main": "lib/index.js",
-  "exports": {
-    ".": "./lib/index.js",
-    "./client": "./lib/client.js",
-    "./cordis.patch.yml": "./cordis.patch.yml",
-    "./package.json": "./package.json"
-  },
-  "engines": {
-    "node": ">=20"
-  },
-  "files": [
-    "lib",
-    "cordis.patch.yml",
-    "README.md"
-  ],
-  "keywords": [
-    "dsh-plugin",
-    "deepseek-harness",
-    "livebench",
-    "benchmark",
-    "evaluation"
-  ],
-  "dsh": {
-    "bundle": {
-      "patch": "./cordis.patch.yml"
-    },
-    "client": {
-      "platform": "web",
-      "inject": [
-        "@deepseek-ai/dsh-client-runtime",
-        "@deepseek-ai/dsh-client-locale",
-        "@deepseek-ai/dsh-client-ui-slots",
-        "@deepseek-ai/dsh-client-ui-conversation"
-      ]
-    }
-  }
-}
+{
+  "name": "dsh-livebench-panel",
+  "version": "0.1.14",
+  "description": "DSH web plugin: a LiveBench tab in the Trajectory view (right of 对话/轨迹). Run LiveBench evaluations against every model configured in the DeepSeek Harness — pick provider/model, category, task, release and question range from dropdowns, watch progress, and read scores in place.",
+  "license": "MIT",
+  "type": "module",
+  "main": "lib/index.js",
+  "exports": {
+    ".": "./lib/index.js",
+    "./client": "./lib/client.js",
+    "./cordis.patch.yml": "./cordis.patch.yml",
+    "./package.json": "./package.json"
+  },
+  "engines": {
+    "node": ">=20"
+  },
+  "files": [
+    "lib",
+    "cordis.patch.yml",
+    "README.md"
+  ],
+  "keywords": [
+    "dsh-plugin",
+    "deepseek-harness",
+    "livebench",
+    "benchmark",
+    "evaluation"
+  ],
+  "dsh": {
+    "bundle": {
+      "patch": "./cordis.patch.yml"
+    },
+    "client": {
+      "platform": "web",
+      "inject": [
+        "@deepseek-ai/dsh-client-runtime",
+        "@deepseek-ai/dsh-client-locale",
+        "@deepseek-ai/dsh-client-ui-slots",
+        "@deepseek-ai/dsh-client-ui-conversation"
+      ]
+    }
+  }
+}