Quelle  memory-embedding-adapter.ts

import {
  sanitizeEmbeddingCacheHeaders,
  type MemoryEmbeddingProviderAdapter,
} from "openclaw/plugin-sdk/memory-core-host-engine-embeddings";
import {
  createLmstudioEmbeddingProvider,
  DEFAULT_LMSTUDIO_EMBEDDING_MODEL,
} from "./src/embedding-provider.js";

export const lmstudioMemoryEmbeddingProviderAdapter: MemoryEmbeddingProviderAdapter = {
  id: "lmstudio",
  defaultModel: DEFAULT_LMSTUDIO_EMBEDDING_MODEL,
  transport: "remote",
  authProviderId: "lmstudio",
  allowExplicitWhenConfiguredAuto: true,
  create: async (options) => {
    const { provider, client } = await createLmstudioEmbeddingProvider({
      ...options,
      provider: "lmstudio",
      fallback: "none",
    });
    return {
      provider,
      runtime: {
        id: "lmstudio",
        cacheKeyData: {
          provider: "lmstudio",
          baseUrl: client.baseUrl,
          model: client.model,
          headers: sanitizeEmbeddingCacheHeaders(client.headers, ["authorization"]),
        },
      },
    };
  },
};