js[minor]: Add clonePublicDataset and listSharedExamples

js/package.json

@@ -276,4 +276,4 @@
     },
     "./package.json": "./package.json"
   }
-}
+}

js/src/client.ts

@@ -1289,7 +1289,7 @@
     treeFilter?: string;
     isRoot?: boolean;
     dataSourceType?: string;
   }): Promise<any> {
     let projectIds_ = projectIds || [];
     if (projectNames) {
       projectIds_ = [
@@ -1528,6 +1528,61 @@
     return dataset as Dataset;
   }
 
+  /**
+   * Get shared examples.
+   *
+   * @param {string} shareToken The share token to get examples for. A share token is the UUID (or LangSmith URL, including UUID) generated when explicitly marking an example as public.
+   * @param {Object} [options] Additional options for listing the examples.
+   * @param {string[] | undefined} [options.exampleIds] A list of example IDs to filter by.
+   * @returns {Promise<Example[]>} The shared examples.
+   */
+  public async listSharedExamples(
+    shareToken: string,
+    options?: { exampleIds?: string[] }
+  ): Promise<Example[]> {
+    const params: Record<string, string | string[]> = {};
+    if (options?.exampleIds) {
+      params.id = options.exampleIds;
+    }
+
+    const urlParams = new URLSearchParams();
+    Object.entries(params).forEach(([key, value]) => {
+      if (Array.isArray(value)) {
+        value.forEach((v) => urlParams.append(key, v));
+      } else {
+        urlParams.append(key, value);
+      }
+    });
+
+    const response = await this.caller.call(
+      fetch,
+      `${this.apiUrl}/public/${shareToken}/examples?${urlParams.toString()}`,
+      {
+        method: "GET",
+        headers: this.headers,
+        signal: AbortSignal.timeout(this.timeout_ms),
+        ...this.fetchOptions,
+      }
+    );
+    const result = await response.json();
+    if (!response.ok) {
+      if ("detail" in result) {
+        throw new Error(
+          `Failed to list shared examples.\nStatus: ${
+            response.status
+          }\nMessage: ${result.detail.join("\n")}`
+        );
+      }
+      throw new Error(
+        `Failed to list shared examples: ${response.status} ${response.statusText}`
+      );
+    }
+    return result.map((example: any) => ({
+      ...example,
+      _hostUrl: this.getHostUrl(),
+    }));
+  }
+
   public async createProject({
     projectName,
     description = null,
@@ -2681,7 +2736,7 @@
     }
 
     const feedbackResult = await evaluator.evaluateRun(run_, referenceExample);
     const [_, feedbacks] = await this._logEvaluationFeedback(
       feedbackResult,
       run_,
       sourceInfo
@@ -3015,7 +3070,7 @@
   async _logEvaluationFeedback(
     evaluatorResponse: EvaluationResult | EvaluationResults,
     run?: Run,
     sourceInfo?: { [key: string]: any }
   ): Promise<[results: EvaluationResult[], feedbacks: Feedback[]]> {
     const evalResults: Array<EvaluationResult> =
       this._selectEvalResults(evaluatorResponse);
@@ -3054,7 +3109,7 @@
   public async logEvaluationFeedback(
     evaluatorResponse: EvaluationResult | EvaluationResults,
     run?: Run,
     sourceInfo?: { [key: string]: any }
   ): Promise<EvaluationResult[]> {
     const [results] = await this._logEvaluationFeedback(
       evaluatorResponse,
@@ -3120,7 +3175,7 @@
     promptIdentifier: string,
     like: boolean
   ): Promise<LikePromptResponse> {
     const [owner, promptName, _] = parsePromptIdentifier(promptIdentifier);
     const response = await this.caller.call(
       fetch,
       `${this.apiUrl}/likes/${owner}/${promptName}`,
@@ -3232,7 +3287,7 @@
   }
 
   public async getPrompt(promptIdentifier: string): Promise<Prompt | null> {
     const [owner, promptName, _] = parsePromptIdentifier(promptIdentifier);
     const response = await this.caller.call(
       fetch,
       `${this.apiUrl}/repos/${owner}/${promptName}`,
@@ -3281,7 +3336,7 @@
       );
     }
 
     const [owner, promptName, _] = parsePromptIdentifier(promptIdentifier);
     if (!(await this._currentTenantIsOwner(owner))) {
       throw await this._ownerConflictError("create a prompt", owner);
     }
@@ -3314,7 +3369,7 @@
 
   public async createCommit(
     promptIdentifier: string,
     object: any,
     options?: {
       parentCommitHash?: string;
     }
@@ -3323,7 +3378,7 @@
       throw new Error("Prompt does not exist, you must create it first.");
     }
 
     const [owner, promptName, _] = parsePromptIdentifier(promptIdentifier);
     const resolvedParentCommitHash =
       options?.parentCommitHash === "latest" || !options?.parentCommitHash
         ? await this._getLatestCommitHash(`${owner}/${promptName}`)
@@ -3561,4 +3616,105 @@
     });
     return url;
   }
+
+  /**
+   * Clone a public dataset to your own langsmith tenant. 
+   * This operation is idempotent. If you already have a dataset with the given name, 
+   * this function will do nothing.
+
+   * @param {string} tokenOrUrl The token of the public dataset to clone.
+   * @param {Object} [options] Additional options for cloning the dataset.
+   * @param {string} [options.sourceApiUrl] The URL of the langsmith server where the data is hosted. Defaults to the API URL of your current client.
+   * @param {string} [options.datasetName] The name of the dataset to create in your tenant. Defaults to the name of the public dataset.
+   * @returns {Promise<void>}
+   */
+  async clonePublicDataset(
+    tokenOrUrl: string,
+    options: {
+      sourceApiUrl?: string;
+      datasetName?: string;
+    } = {}
+  ): Promise<void> {
+    const { sourceApiUrl = this.apiUrl, datasetName } = options;
+    const [parsedApiUrl, tokenUuid] = this.parseTokenOrUrl(
+      tokenOrUrl,
+      sourceApiUrl
+    );
+    const sourceClient = new Client({
+      apiUrl: parsedApiUrl,
+      // Placeholder API key not needed anymore in most cases, but
+      // some private deployments may have API key-based rate limiting
+      // that would cause this to fail if we provide no value.
+      apiKey: "placeholder",
+    });
+
+    const ds = await sourceClient.readSharedDataset(tokenUuid);
+    const finalDatasetName = datasetName || ds.name;
+
+    try {
+      if (await this.hasDataset({ datasetId: finalDatasetName })) {
+        console.log(
+          `Dataset ${finalDatasetName} already exists in your tenant. Skipping.`
+        );
+        return;
+      }
+    } catch (_) {
+      // `.hasDataset` will throw an error if the dataset does not exist.
+      // no-op in that case
+    }
+
+    // Fetch examples first, then create the dataset
+    const examples = await sourceClient.listSharedExamples(tokenUuid);
+    const dataset = await this.createDataset(finalDatasetName, {
+      description: ds.description,
+      dataType: ds.data_type || "kv",
+      inputsSchema: ds.inputs_schema_definition ?? undefined,
+      outputsSchema: ds.outputs_schema_definition ?? undefined,
+    });
+    try {
+      await this.createExamples({
+        inputs: examples.map((e) => e.inputs),
+        outputs: examples.flatMap((e) => (e.outputs ? [e.outputs] : [])),
+        datasetId: dataset.id,
+      });
+    } catch (e) {
+      console.error(
+        `An error occurred while creating dataset ${finalDatasetName}. ` +
+          "You should delete it manually."
+      );
+      throw e;
+    }
+  }
+
+  private parseTokenOrUrl(
+    urlOrToken: string,
+    apiUrl: string,
+    numParts = 2,
+    kind = "dataset"
+  ): [string, string] {
+    // Try parsing as UUID
+    try {
+      assertUuid(urlOrToken); // Will throw if it's not a UUID.
+      return [apiUrl, urlOrToken];
+    } catch (_) {
+      // no-op if it's not a uuid
+    }
+
+    // Parse as URL
+    try {
+      const parsedUrl = new URL(urlOrToken);
+      const pathParts = parsedUrl.pathname
+        .split("/")
+        .filter((part) => part !== "");
+
+      if (pathParts.length >= numParts) {
+        const tokenUuid = pathParts[pathParts.length - numParts];
+        return [apiUrl, tokenUuid];
+      } else {
+        throw new Error(`Invalid public ${kind} URL: ${urlOrToken}`);
+      }
+    } catch (error) {
+      throw new Error(`Invalid public ${kind} URL or token: ${urlOrToken}`);
+    }
+  }
 }

js/src/schemas.ts

@@ -261,6 +261,8 @@ export interface BaseDataset {
   description: string;
   tenant_id: string;
   data_type?: DataType;
+  inputs_schema_definition?: KVMap;
+  outputs_schema_definition?: KVMap;
 }
 
 export interface Dataset extends BaseDataset {

js/src/tests/client.int.test.ts

@@ -1,4 +1,4 @@
-import { Dataset, Run, TracerSession } from "../schemas.js";
+import { Dataset, Example, Run, TracerSession } from "../schemas.js";
 import {
   FunctionMessage,
   HumanMessage,
@@ -1074,3 +1074,42 @@ test("Test pull prompt include model", async () => {
 
   await client.deletePrompt(promptName);
 });
+
+test("list shared examples can list shared examples", async () => {
+  const client = new Client();
+  const multiverseMathPublicDatasetShareToken =
+    "620596ee-570b-4d2b-8c8f-f828adbe5242";
+  const sharedExamples = await client.listSharedExamples(
+    multiverseMathPublicDatasetShareToken
+  );
+  expect(sharedExamples.length).toBeGreaterThan(0);
+});
+
+test("clonePublicDataset method can clone a dataset", async () => {
+  const client = new Client();
+  const datasetName = "multiverse_math_public_testing";
+  const multiverseMathPublicDatasetURL =
+    "https://smith.langchain.com/public/620596ee-570b-4d2b-8c8f-f828adbe5242/d";
+
+  try {
+    await client.clonePublicDataset(multiverseMathPublicDatasetURL, {
+      datasetName,
+    });
+
+    const clonedDataset = await client.hasDataset({ datasetName });
+    expect(clonedDataset).toBe(true);
+
+    const examples: Example[] = [];
+    for await (const ex of client.listExamples({ datasetName })) {
+      examples.push(ex);
+    }
+    expect(examples.length).toBeGreaterThan(0);
+  } finally {
+    try {
+      // Attempt to remove the newly created dataset if successful.
+      await client.deleteDataset({ datasetName });
+    } catch (_) {
+      // no-op if failure
+    }
+  }
+});