pi-llama-cpp 0.14.0 → 0.15.0

package/src/ui/editors/server/utils.ts

@@ -1,8 +1,12 @@
+import { ENDPOINT_PREFIX } from "../../../constants";
 import { ServerStatus } from "../../../enums/serverStatus";
 import type { LlamaServer } from "../../../interfaces/settings";
 import { checkServerHealth } from "../../../utils/health";
 import { ServerIds } from "../../../utils/serverIds";
 
+/** Emoji shown when a server requires an API key (401 on /v1/models). */
+const UNAUTHORIZED_EMOJI = "⛔";
+
 /**
  * Display concerns for a server row in the top-level list: the dim
  * `(...)` suffix after the URL and the health-status icon.
@@ -33,6 +37,52 @@ export class ServerDisplay {
   static async healthEmoji(url: string, timeout: number): Promise<string> {
     return SERVER_STATUS_ICONS[await checkServerHealth(url, timeout)];
   }
+
+  /**
+   * Checks if the server requires an API key by probing `/v1/models`.
+   *
+   * @param url - The server URL to check
+   * @param apiKey - The API key to send (may be a placeholder)
+   * @param timeout - Maximum time (ms) to wait for the request
+   * @returns `true` if the server responded with 401, `false` otherwise
+   */
+  static async requiresApiKey(
+    url: string,
+    apiKey: string,
+    timeout: number,
+  ): Promise<boolean> {
+    try {
+      const response = await fetch(`${url}${ENDPOINT_PREFIX}/models`, {
+        signal: AbortSignal.timeout(timeout),
+        headers: { Authorization: `Bearer ${apiKey}` },
+      });
+      return response.status === 401;
+    } catch {
+      // Timeout or network error — the server is unreachable, so no auth
+      // indicator is needed (the health emoji will cover that).
+      return false;
+    }
+  }
+
+  /**
+   * Probes the server's `/v1/models` endpoint to check if authentication
+   * is required. Returns ⛔ when the server responds with 401, empty string
+   * otherwise.
+   *
+   * @param url - The server URL to check
+   * @param apiKey - The API key to send (may be a placeholder)
+   * @param timeout - Maximum time (ms) to wait for the request
+   * @returns The emoji indicator (⛔ if 401, "" otherwise)
+   */
+  static async authEmoji(
+    url: string,
+    apiKey: string,
+    timeout: number,
+  ): Promise<string> {
+    return (await this.requiresApiKey(url, apiKey, timeout))
+      ? UNAUTHORIZED_EMOJI
+      : "";
+  }
 }
 
 /**

package/src/ui/settings/index.ts

@@ -26,6 +26,7 @@ export enum Options {
   SORT_BY = "sortBy",
   POLLING_TIMEOUT = "pollingTimeout",
   SERVER_TIMEOUT = "serverTimeout",
+  SHOW_SERVER_URLS = "showServerUrls",
 }
 
 // ─── Constants ──────────────────────────────────────────────────────────────
@@ -110,6 +111,13 @@ export const buildSettingsItems = async (
       currentValue: formatMs(serverTimeout),
       values: SERVER_PRESETS.map(formatMs),
     },
+    {
+      id: Options.SHOW_SERVER_URLS,
+      label: "Show server URLs",
+      description: "Show [Server: <url>] in /models (next open)",
+      currentValue: (await settings.resolveShowServerUrls()) ? "on" : "off",
+      values: ["on", "off"],
+    },
   ];
 };
 
@@ -139,6 +147,9 @@ export const applySettingChange = async (
     case Options.SERVER_TIMEOUT:
       await settings.setLlamaSetting("serverTimeout", parseMs(newValue));
       return;
+    case Options.SHOW_SERVER_URLS:
+      await settings.setLlamaSetting("showServerUrls", newValue === "on");
+      return;
   }
 };
 

package/src/ui/strings.ts

@@ -1,3 +1,5 @@
+import { getAgentDir } from "@earendil-works/pi-coding-agent";
+
 import { LLAMA_SERVER_URL } from "../constants";
 
 /**
@@ -129,3 +131,18 @@ export const HINTS = {
   /** Row description for a server in the overrides editor */
   overrideServerRow: "Enter: edit this server's override entries",
 } as const;
+
+/**
+ * Generates a warning message for servers that require an API key.
+ * Centralized so both the runtime manager and the editor wizard use
+ * identical wording.
+ */
+export const authRequiredMessage = (
+  baseUrl: string,
+  providerId: string,
+): string =>
+  [
+    "[pi-llama-cpp]",
+    `Server at '${baseUrl}' requires a valid API key.`,
+    `Configure the key via '/login ${providerId}' or in '${getAgentDir()}/auth.json'.`,
+  ].join("\n");

package/src/utils/health.ts

@@ -1,4 +1,5 @@
 import { ServerStatus } from "../enums/serverStatus";
+import type { HealthEndpoint } from "../interfaces/endpoints/health";
 
 /**
  * Probes a llama-server's `/health` endpoint and classifies the outcome.
@@ -34,7 +35,7 @@ export const checkServerHealth = async (
       signal: AbortSignal.timeout(timeout),
       headers: apiKey ? { Authorization: `Bearer ${apiKey}` } : undefined,
     });
-    const data = (await response.json()) as { status?: string };
+    const data = (await response.json()) as HealthEndpoint;
     return data.status === "ok" ? ServerStatus.READY : ServerStatus.UNREACHABLE;
   } catch (error) {
     // `AbortSignal.timeout` rejects `fetch` with a `TimeoutError`

package/src/utils/urlResolver.ts

@@ -14,7 +14,7 @@ export interface UrlResolverDeps {
 }
 
 /**
- * Multi-source llama-server URL resolution chain (see REFACTOR.md #4).
+ * Multi-source llama-server URL resolution chain
  *
  * Resolves URLs in priority order:
  *

package/tests/{commandManager.test.ts → command/commandManager.test.ts}

@@ -2,17 +2,17 @@ import type { Theme } from "@earendil-works/pi-coding-agent";
 import { initTheme } from "@earendil-works/pi-coding-agent";
 import type { KeybindingsManager, TUI } from "@earendil-works/pi-tui";
 import { beforeEach, describe, expect, it, vi } from "vitest";
-import { Action } from "../src/enums/action";
-import { CommandManager } from "../src/managers/command";
-import { ServerManager } from "../src/managers/server";
-import type { LlamaSettingsManager } from "../src/managers/settings";
-import { ServerSettingsList } from "../src/ui/editors/server/serverEditor";
-import { ServerDisplay } from "../src/ui/editors/server/utils";
+import { Action } from "../../src/enums/action";
+import { CommandManager } from "../../src/managers/command";
+import { ServerManager } from "../../src/managers/server";
+import type { LlamaSettingsManager } from "../../src/managers/settings";
+import { ServerSettingsList } from "../../src/ui/editors/server/serverEditor";
+import { ServerDisplay } from "../../src/ui/editors/server/utils";
 import {
   applySettingChange,
   buildSettingsItems,
   formatMs,
-} from "../src/ui/settings";
+} from "../../src/ui/settings";
 import {
   createMockCtx,
   createMockModel,
@@ -20,14 +20,15 @@ import {
   createMockServer,
   makeSettingsStub,
   mockRpc,
-} from "./mocks";
+} from "../mocks";
 
 beforeEach(() => {
   initTheme();
   vi.clearAllMocks();
   mockRpc.mockResolvedValue({ data: [] });
-  // Mock health checks to return "healthy" immediately
+  // Mock health checks and auth probes to return immediately
   vi.spyOn(ServerDisplay, "healthEmoji").mockResolvedValue("🟢");
+  vi.spyOn(ServerDisplay, "authEmoji").mockResolvedValue("");
 });
 
 /**
@@ -163,6 +164,7 @@ describe("CommandManager", () => {
         "sortBy",
         "pollingTimeout",
         "serverTimeout",
+        "showServerUrls",
       ]);
       // Booleans are displayed as on/off
       expect(items[0].values).toEqual(["on", "off"]);
@@ -203,6 +205,18 @@ describe("CommandManager", () => {
         "serverTimeout",
         500,
       );
+
+      await applySettingChange("showServerUrls", "on", settingsStub);
+      expect(settingsStub.setLlamaSetting).toHaveBeenCalledWith(
+        "showServerUrls",
+        true,
+      );
+
+      await applySettingChange("showServerUrls", "off", settingsStub);
+      expect(settingsStub.setLlamaSetting).toHaveBeenCalledWith(
+        "showServerUrls",
+        false,
+      );
     });
   });
 
@@ -564,5 +578,62 @@ describe("CommandManager", () => {
       expect(choices).toHaveLength(2);
       expect(choices[1]).toContain("model-b");
     });
+
+    it("should omit server URLs when showServerUrls is false", async () => {
+      const models = [createMockModel("model-a")];
+      const mockPi = createMockPi();
+      const servers = models.map((model) =>
+        createMockServer({
+          baseUrl: model.serverUrl,
+          models: [model],
+        }),
+      );
+      const settingsStub = makeSettingsStub({
+        resolveServers: vi.fn(async () => servers),
+        resolveShowServerUrls: vi.fn(async () => false),
+      });
+      const serverManager = new ServerManager(settingsStub);
+      const commandManager = new CommandManager(serverManager, settingsStub);
+
+      const ctx = createMockCtx((prompt, options) => {
+        if (prompt.includes("models")) return options[0];
+        return Action.INFO;
+      });
+
+      await commandManager.handleCommand("", ctx as any, mockPi as any);
+
+      const choices = vi.mocked(ctx.ui.select).mock.calls[0][1] as string[];
+      expect(choices).toHaveLength(1);
+      expect(choices[0]).toBe("model-a  ");
+      expect(choices[0]).not.toContain("[Server:");
+    });
+
+    it("should include server URLs when showServerUrls is true", async () => {
+      const models = [createMockModel("model-a")];
+      const mockPi = createMockPi();
+      const servers = models.map((model) =>
+        createMockServer({
+          baseUrl: model.serverUrl,
+          models: [model],
+        }),
+      );
+      const settingsStub = makeSettingsStub({
+        resolveServers: vi.fn(async () => servers),
+        resolveShowServerUrls: vi.fn(async () => true),
+      });
+      const serverManager = new ServerManager(settingsStub);
+      const commandManager = new CommandManager(serverManager, settingsStub);
+
+      const ctx = createMockCtx((prompt, options) => {
+        if (prompt.includes("models")) return options[0];
+        return Action.INFO;
+      });
+
+      await commandManager.handleCommand("", ctx as any, mockPi as any);
+
+      const choices = vi.mocked(ctx.ui.select).mock.calls[0][1] as string[];
+      expect(choices).toHaveLength(1);
+      expect(choices[0]).toContain("[Server: http://127.0.0.1:8080]");
+    });
   });
 });

package/tests/{events.test.ts → events/events.test.ts}

@@ -1,10 +1,10 @@
 import { beforeEach, describe, expect, it, vi } from "vitest";
-import { THINKING_BUDGETS } from "../src/constants";
-import { Status } from "../src/enums/status";
-import { EventManager } from "../src/managers/events";
-import { ServerManager } from "../src/managers/server";
-import type { Server } from "../src/server";
-import { createMockModel, createMockServer, makeSettingsStub } from "./mocks";
+import { THINKING_BUDGETS } from "../../src/constants";
+import { Status } from "../../src/enums/status";
+import { EventManager } from "../../src/managers/events";
+import { ServerManager } from "../../src/managers/server";
+import type { Server } from "../../src/server";
+import { createMockModel, createMockServer, makeSettingsStub } from "../mocks";
 
 /**
  * Injected settings stub (EventManager and — in the live-list test — the

package/tests/mocks.ts

@@ -7,6 +7,7 @@ import {
   POLLING_TIMEOUT,
   REACT_TO_MODEL_SELECT,
   SERVER_TIMEOUT,
+  SHOW_SERVER_URLS,
   SORT_BY,
   THINKING_BUDGETS,
 } from "../src/constants";
@@ -40,6 +41,7 @@ export const makeSettingsStub = (
     })),
     resolveServers: vi.fn(async () => []),
     resolveSortBy: vi.fn(async () => SORT_BY),
+    resolveShowServerUrls: vi.fn(async () => SHOW_SERVER_URLS),
     resolveApiKey: vi.fn(() => API_KEY_PLACEHOLDER),
     resolveReactToModelSelect: vi.fn(async () => REACT_TO_MODEL_SELECT),
     resolveAutoloadOnMessage: vi.fn(async () => AUTOLOAD_ON_MESSAGE),

package/tests/{legacyModel.test.ts → models/legacyModel.test.ts}

@@ -1,9 +1,9 @@
 import { beforeEach, describe, expect, it } from "vitest";
-import { FALLBACK_CTX } from "../src/constants";
-import { Mode } from "../src/enums/mode";
-import { DataProperty } from "../src/interfaces/endpoints/models";
-import { LegacyModel } from "../src/models/legacyModel";
-import { createMockServer, mockRpc } from "./mocks";
+import { FALLBACK_CTX } from "../../src/constants";
+import { Mode } from "../../src/enums/mode";
+import { DataProperty } from "../../src/interfaces/endpoints/models";
+import { LegacyModel } from "../../src/models/legacyModel";
+import { createMockServer, mockRpc } from "../mocks";
 
 beforeEach(() => {
   mockRpc.mockReset();

package/tests/{routerModel.test.ts → models/routerModel.test.ts}

@@ -1,9 +1,9 @@
 import { beforeEach, describe, expect, it } from "vitest";
-import { FALLBACK_CTX } from "../src/constants";
-import { Mode } from "../src/enums/mode";
-import { DataProperty } from "../src/interfaces/endpoints/models";
-import { RouterModel } from "../src/models/routerModel";
-import { createMockServer, mockRpc } from "./mocks";
+import { FALLBACK_CTX } from "../../src/constants";
+import { Mode } from "../../src/enums/mode";
+import { DataProperty } from "../../src/interfaces/endpoints/models";
+import { RouterModel } from "../../src/models/routerModel";
+import { createMockServer, mockRpc } from "../mocks";
 
 // Helper to create a mock DataProperty
 const createModel = (overrides: Partial<DataProperty> = {}): DataProperty => ({

package/tests/{singleModel.test.ts → models/singleModel.test.ts}

@@ -1,9 +1,9 @@
 import { beforeEach, describe, expect, it } from "vitest";
-import { Mode } from "../src/enums/mode";
-import { Status } from "../src/enums/status";
-import { DataProperty } from "../src/interfaces/endpoints/models";
-import { SingleModel } from "../src/models/singleModel";
-import { createMockServer, mockRpc } from "./mocks";
+import { Mode } from "../../src/enums/mode";
+import { Status } from "../../src/enums/status";
+import { DataProperty } from "../../src/interfaces/endpoints/models";
+import { SingleModel } from "../../src/models/singleModel";
+import { createMockServer, mockRpc } from "../mocks";
 
 beforeEach(() => {
   mockRpc.mockReset();

package/tests/{health.test.ts → server/health.test.ts}

@@ -1,6 +1,6 @@
 import { afterEach, describe, expect, it, vi } from "vitest";
-import { ServerStatus } from "../src/enums/serverStatus";
-import { checkServerHealth } from "../src/utils/health";
+import { ServerStatus } from "../../src/enums/serverStatus";
+import { checkServerHealth } from "../../src/utils/health";
 
 /**
  * Stubs `global.fetch` for one test. The probe only reads the parsed body,

package/tests/{server.test.ts → server/server.test.ts}

@@ -1,9 +1,9 @@
 import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
-import { POLLING_TIMEOUT, SERVER_TIMEOUT } from "../src/constants";
-import { ServerStatus } from "../src/enums/serverStatus";
-import type { LlamaSettingsManager } from "../src/managers/settings";
-import { Server } from "../src/server";
-import { createMockServer, makeSettingsStub, mockRpc } from "./mocks";
+import { POLLING_TIMEOUT, SERVER_TIMEOUT } from "../../src/constants";
+import { ServerStatus } from "../../src/enums/serverStatus";
+import type { LlamaSettingsManager } from "../../src/managers/settings";
+import { Server } from "../../src/server";
+import { createMockServer, makeSettingsStub, mockRpc } from "../mocks";
 
 // Injected into every real Server below; fresh per test so per-case
 // overrides never leak. The Server constructor resolves the API key eagerly

package/tests/{serverManager.test.ts → server/serverManager.test.ts}

@@ -1,13 +1,13 @@
 import { beforeEach, describe, expect, it, vi } from "vitest";
-import { ServerManager } from "../src/managers/server";
-import { BaseModel } from "../src/models/baseModel";
-import { Server } from "../src/server";
+import { ServerManager } from "../../src/managers/server";
+import { BaseModel } from "../../src/models/baseModel";
+import { Server } from "../../src/server";
 import {
   createMockModel,
   createMockServer,
   makeSettingsStub,
   mockRpc,
-} from "./mocks";
+} from "../mocks";
 
 // Injected settings stub — the single instance passed to ServerManager and
 // to every real `new Server(...)` construction below (Server's eager