@preapexis/pi-kit 1.1.3 → 1.2.1

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@preapexis/pi-kit",
-  "version": "1.1.3",
+  "version": "1.2.1",
   "description": "Personal Pi coding-agent kit with safety extensions, status UI, prompt workflows, skills, and themes.",
   "keywords": [
     "pi-package",

package/extensions/litellm-provider.ts

@@ -1,302 +0,0 @@
-// cSpell:words litellm preapexis deepseek
-import type { ExtensionAPI } from "@earendil-works/pi-coding-agent";
-
-type LiteLlmModelInfo = {
-  model_name?: string;
-  litellm_params?: {
-    model?: string;
-  };
-  model_info?: {
-    id?: string;
-    name?: string;
-    display_name?: string;
-    description?: string;
-    max_tokens?: number;
-    max_input_tokens?: number;
-    context_window?: number;
-    input_cost_per_token?: number;
-    output_cost_per_token?: number;
-    supports_vision?: boolean;
-    supports_function_calling?: boolean;
-    mode?: string;
-  };
-};
-
-type OpenAiModel = {
-  id?: string;
-  object?: string;
-  owned_by?: string;
-};
-
-type ModelPayload = {
-  data?: Array<LiteLlmModelInfo | OpenAiModel>;
-};
-
-type PiModel = {
-  id: string;
-  name: string;
-  reasoning: boolean;
-  input: Array<"text" | "image">;
-  cost: {
-    input: number;
-    output: number;
-    cacheRead: number;
-    cacheWrite: number;
-  };
-  contextWindow: number;
-  maxTokens: number;
-  compat?: {
-    supportsDeveloperRole?: boolean;
-    supportsReasoningEffort?: boolean;
-  };
-};
-
-const PROVIDER_ID = "litellm";
-const DEFAULT_BASE_URL = "http://localhost:4000/v1";
-
-export default async function (pi: ExtensionAPI): Promise<void> {
-  const baseUrl = normalizeBaseUrl(
-    process.env.LITELLM_BASE_URL ?? DEFAULT_BASE_URL
-  );
-
-  async function registerLiteLlmProvider(): Promise<void> {
-    const models = await discoverModels(baseUrl);
-
-    pi.registerProvider(PROVIDER_ID, {
-      name: "LiteLLM",
-      baseUrl,
-      apiKey: "$LITELLM_API_KEY",
-      api: "openai-completions",
-      models
-    });
-  }
-
-  await registerLiteLlmProvider();
-
-  pi.registerCommand("litellm-refresh", {
-    description: "Refresh LiteLLM models from the LiteLLM proxy",
-    handler: async (_args, ctx) => {
-      try {
-        await registerLiteLlmProvider();
-
-        ctx.ui.notify(
-          [
-            "LiteLLM models refreshed.",
-            "",
-            `Base URL: ${baseUrl}`,
-            "",
-            "Run /model to select a LiteLLM model."
-          ].join("\n"),
-          "info"
-        );
-      } catch (error) {
-        ctx.ui.notify(
-          [
-            "LiteLLM model refresh failed.",
-            "",
-            error instanceof Error ? error.message : String(error)
-          ].join("\n"),
-          "error"
-        );
-      }
-    }
-  });
-}
-
-function normalizeBaseUrl(value: string): string {
-  return value.replace(/\/+$/, "");
-}
-
-function getProxyRoot(baseUrl: string): string {
-  return baseUrl.endsWith("/v1") ? baseUrl.slice(0, -3) : baseUrl;
-}
-
-async function discoverModels(baseUrl: string): Promise<PiModel[]> {
-  const proxyRoot = getProxyRoot(baseUrl);
-
-  const endpoints = [
-    `${proxyRoot}/v1/model/info`,
-    `${proxyRoot}/model/info`,
-    `${baseUrl}/models`
-  ];
-
-  for (const endpoint of endpoints) {
-    try {
-      const payload = await fetchModelPayload(endpoint);
-      const models = payloadToPiModels(payload);
-
-      if (models.length > 0) {
-        return models;
-      }
-    } catch {
-      // Try the next endpoint.
-    }
-  }
-
-  const fallbackModels = getFallbackModels();
-
-  if (fallbackModels.length > 0) {
-    return fallbackModels;
-  }
-
-  throw new Error(
-    [
-      "Could not discover LiteLLM models.",
-      "",
-      `Tried: ${endpoints.join(", ")}`,
-      "",
-      "Make sure LiteLLM is running and set LITELLM_API_KEY if your proxy requires auth.",
-      "You can also set LITELLM_MODELS as a comma-separated fallback."
-    ].join("\n")
-  );
-}
-
-async function fetchModelPayload(url: string): Promise<ModelPayload> {
-  const headers: Record<string, string> = {
-    accept: "application/json"
-  };
-
-  const apiKey = process.env.LITELLM_API_KEY;
-
-  if (apiKey) {
-    headers.authorization = `Bearer ${apiKey}`;
-  }
-
-  const response = await fetch(url, {
-    method: "GET",
-    headers
-  });
-
-  if (!response.ok) {
-    throw new Error(`${url} returned HTTP ${response.status}`);
-  }
-
-  return (await response.json()) as ModelPayload;
-}
-
-function payloadToPiModels(payload: ModelPayload): PiModel[] {
-  const data = Array.isArray(payload.data) ? payload.data : [];
-  const seen = new Set<string>();
-  const models: PiModel[] = [];
-
-  for (const item of data) {
-    const model = modelFromPayloadItem(item);
-
-    if (!model) continue;
-    if (seen.has(model.id)) continue;
-
-    seen.add(model.id);
-    models.push(model);
-  }
-
-  return models.sort((a, b) => a.id.localeCompare(b.id));
-}
-
-function modelFromPayloadItem(
-  item: LiteLlmModelInfo | OpenAiModel
-): PiModel | undefined {
-  const modelInfo = "model_info" in item ? item.model_info : undefined;
-
-  const id =
-    "model_name" in item && item.model_name
-      ? item.model_name
-      : "id" in item && item.id
-        ? item.id
-        : modelInfo?.id;
-
-  if (!id) return undefined;
-
-  const name = modelInfo?.display_name ?? modelInfo?.name ?? id;
-
-  const contextWindow =
-    modelInfo?.context_window ?? modelInfo?.max_input_tokens ?? 128000;
-
-  const maxTokens = modelInfo?.max_tokens ?? 4096;
-
-  const supportsVision =
-    modelInfo?.supports_vision === true || looksLikeVisionModel(id);
-
-  return {
-    id,
-    name,
-    reasoning: looksLikeReasoningModel(id),
-    input: supportsVision ? ["text", "image"] : ["text"],
-    cost: {
-      input: costPerMillion(modelInfo?.input_cost_per_token),
-      output: costPerMillion(modelInfo?.output_cost_per_token),
-      cacheRead: 0,
-      cacheWrite: 0
-    },
-    contextWindow,
-    maxTokens,
-    compat: {
-      supportsDeveloperRole: false,
-      supportsReasoningEffort: false
-    }
-  };
-}
-
-function costPerMillion(value: number | undefined): number {
-  if (typeof value !== "number" || Number.isNaN(value)) {
-    return 0;
-  }
-
-  return value * 1_000_000;
-}
-
-function looksLikeVisionModel(id: string): boolean {
-  const lower = id.toLowerCase();
-
-  return (
-    lower.includes("vision") ||
-    lower.includes("vl") ||
-    lower.includes("gpt-4o") ||
-    lower.includes("gemini") ||
-    lower.includes("claude")
-  );
-}
-
-function looksLikeReasoningModel(id: string): boolean {
-  const lower = id.toLowerCase();
-
-  return (
-    lower.includes("reason") ||
-    lower.includes("thinking") ||
-    lower.includes("o1") ||
-    lower.includes("o3") ||
-    lower.includes("o4") ||
-    lower.includes("r1") ||
-    lower.includes("deepseek")
-  );
-}
-
-function getFallbackModels(): PiModel[] {
-  const raw = process.env.LITELLM_MODELS;
-
-  if (!raw?.trim()) {
-    return [];
-  }
-
-  return raw
-    .split(",")
-    .map((value) => value.trim())
-    .filter(Boolean)
-    .map((id) => ({
-      id,
-      name: id,
-      reasoning: looksLikeReasoningModel(id),
-      input: looksLikeVisionModel(id) ? ["text", "image"] : ["text"],
-      cost: {
-        input: 0,
-        output: 0,
-        cacheRead: 0,
-        cacheWrite: 0
-      },
-      contextWindow: 128000,
-      maxTokens: 4096,
-      compat: {
-        supportsDeveloperRole: false,
-        supportsReasoningEffort: false
-      }
-    }));
-}