import type OpenAI from "openai";
import { type PolyModel, modelKey } from "@/lib/ai/core/types";
import { createOpenAICompatAdapter, listOpenAIModels } from "../shared/openai-compat/factory";
import { CEREBRAS_CATALOG } from "./catalog";

/**
 * Cerebras Inference — OpenAI-compatible at https://api.cerebras.ai/v1 (open models, very fast).
 * `GET /v1/models` exposes ids only; the catalog (docs/provider-research/cerebras.md) provides
 * limits, pricing and which reasoning knob each model accepts.
 */
export function normalizeCerebrasModel(m: { id: string }): PolyModel | null {
  const cat = CEREBRAS_CATALOG.get(m.id);
  return {
    key: modelKey("cerebras", m.id),
    id: m.id,
    provider: "cerebras",
    displayName: cat?.displayName ?? prettify(m.id),
    family: cat?.family ?? (m.id.startsWith("gpt-oss") ? "GPT-OSS" : m.id.startsWith("qwen") ? "Qwen" : m.id.startsWith("gemma") ? "Gemma" : m.id.startsWith("llama") ? "Llama" : "Open models"),
    capabilities: cat?.capabilities ?? { text: true, vision: !m.id.startsWith("gpt-oss"), audioInput: false, audioOutput: false, imageGeneration: false, video: false, reasoning: true, tools: true, structuredOutput: true, streaming: true, files: false, webSearch: false },
    limits: cat?.limits ?? {},
    parameters: cat?.parameters ?? { temperature: true, topP: true, maxTokens: true, stop: true, seed: true, frequencyPenalty: true, presencePenalty: true, topK: false, reasoningEffort: true, reasoningEffortLevels: ["low", "medium", "high"], temperatureRange: { min: 0, max: 2 } },
    status: cat?.status ?? "active",
    pricing: cat?.pricing ?? null,
    metadata: { ...(cat?.metadata ?? {}), sortWeight: cat?.sortWeight ?? (m.id.startsWith("gpt-oss") ? 100 : m.id.startsWith("qwen") ? 90 : 80) },
  };
}

function prettify(id: string) {
  return id.replace(/gpt-oss/, "GPT-OSS").replace(/qwen/, "Qwen").replace(/gemma/, "Gemma").replace(/llama/, "Llama").replace(/-(\d+)b\b/, " $1B").replace(/-/g, " ");
}

async function list(client: OpenAI, _k: string, signal?: AbortSignal): Promise<PolyModel[]> {
  return (await listOpenAIModels(client, signal)).map(normalizeCerebrasModel).filter((m): m is PolyModel => Boolean(m));
}

export const cerebrasAdapter = createOpenAICompatAdapter({
  id: "cerebras",
  name: "Cerebras",
  baseURL: "https://api.cerebras.ai/v1",
  keyDocsUrl: "https://cloud.cerebras.ai/",
  keyPrefixHint: "csk-",
  listModels: list,
  messageOptions: { inlineFiles: true },
  tweakParams: (params, settings, req) => {
    const p = params as unknown as Record<string, unknown>;
    const meta = (req.modelInfo?.metadata ?? {}) as Record<string, unknown>;
    // Cerebras validates strictly: only `reasoning_effort` exists (no disable_reasoning / budget).
    if (req.modelInfo?.capabilities.reasoning && settings.reasoningEffort) {
      const levels = req.modelInfo.parameters.reasoningEffortLevels ?? ["low", "medium", "high"];
      const wanted = settings.reasoningEffort === "minimal" ? "low" : settings.reasoningEffort === "xhigh" || settings.reasoningEffort === "max" ? "high" : settings.reasoningEffort;
      if (levels.includes(wanted)) p.reasoning_effort = wanted;
    }
    // gpt-oss rejects tools + response_format together — tools win, JSON is requested in the prompt instead.
    if (meta.toolsWithResponseFormat === false && params.tools?.length && params.response_format) delete p.response_format;
    // Reasoning tokens consume the output cap; never leave a reasoning model with a tiny cap.
    if (req.modelInfo?.capabilities.reasoning && typeof params.max_completion_tokens === "number" && params.max_completion_tokens < Number(meta.minOutputForReasoning ?? 1000)) params.max_completion_tokens = Number(meta.minOutputForReasoning ?? 1000);
  },
  refineError: (status, code, message) => {
    if (status === 401) return "INVALID_API_KEY";
    if (status === 404) return "MODEL_NOT_FOUND";
    if (status === 429 || code === "token_quota_exceeded" || /quota/i.test(message)) return "RATE_LIMITED";
    if (code === "wrong_api_key" || code === "invalid_api_key") return "INVALID_API_KEY";
    if (code === "model_not_found") return "MODEL_NOT_FOUND";
    if (code === "context_length_exceeded") return "CONTEXT_TOO_LONG";
    return undefined;
  },
});