Added Google Scholar Integration

docs/core_docs/docs/integrations/tools/google_scholar.mdx

@@ -0,0 +1,32 @@
+---
+hide_table_of_contents: true
+---
+
+import CodeBlock from "@theme/CodeBlock";
+
+# Google Scholar Tool
+
+The Google Scholar Tool enables your agent to access scholarly articles and publications data through the Google Scholar API. Use it to retrieve article titles, abstracts, authors, and publication information directly from Google Scholar.
+
+## Setup
+
+You will need to get an API key from [SerpAPI](https://serpapi.com/), which offers a Google Scholar API as part of its services. Sign up on SerpAPI, retrieve your API key from your account settings, and set it as process.env.SERPAPI_API_KEY or pass it in as an apiKey constructor argument.
+
+## Usage
+
+import IntegrationInstallTooltip from "@mdx_components/integration_install_tooltip.mdx";
+
+<IntegrationInstallTooltip></IntegrationInstallTooltip>
+
+```bash npm2yarn
+npm install @langchain/openai @langchain/community @langchain/core
+```
+
+import ToolExample from "@examples/tools/google_scholar.ts";
+
+<CodeBlock language="typescript">{ToolExample}</CodeBlock>
+
+## Related
+
+- Tool [conceptual guide](/docs/concepts/tools)
+- Tool [how-to guides](/docs/how_to/#tools)

examples/src/tools/google_scholar.ts

@@ -0,0 +1,15 @@
+import { GoogleScholarAPI } from "@langchain/community/tools/google_scholar";
+
+const scholar = new GoogleScholarAPI({
+  apiKey: process.env.SERPAPI_API_KEY
+});
+
+(async () => {
+  try {
+    const query = "Attention is all we need";
+    const results = await scholar._call(query);
+    console.log("Search Results:", results);
+  } catch (error) {
+    console.error("Error querying Google Scholar via SerpApi:", error);
+  }
+})();

libs/langchain-community/src/tools/google_scholar.ts

@@ -0,0 +1,122 @@
+import { Tool } from "@langchain/core/tools";
+import { getEnvironmentVariable } from "@langchain/core/utils/env";
+import fetch from "node-fetch"; // For making HTTP requests
+
+/**
+ * Interface for parameters required by the GoogleScholarAPI class.
+ */
+export interface GoogleScholarAPIParams {
+  /**
+   * Optional API key for accessing the SerpApi service.
+   */
+  apiKey?: string;
+}
+
+/**
+ * Tool for querying Google Scholar using the SerpApi service.
+ */
+export class GoogleScholarAPI extends Tool {
+  /**
+   * Specifies the name of the tool, used internally by LangChain.
+   */
+  static lc_name() {
+    return "GoogleScholarAPI";
+  }
+
+  /**
+   * Returns a mapping of secret environment variable names to their usage in the tool.
+   * @returns {object} Mapping of secret names to their environment variable counterparts.
+   */
+  get lc_secrets(): { [key: string]: string } | undefined {
+    return {
+      apiKey: "SERPAPI_API_KEY",
+    };
+  }
+
+  // Name of the tool, used for logging or identification within LangChain.
+  name = "google_scholar";
+
+  // The API key used for making requests to SerpApi.
+  protected apiKey: string;
+
+  /**
+   * Description of the tool for usage documentation.
+   */
+  description = `A wrapper around Google Scholar API via SerpApi. Useful for querying academic 
+  articles and papers by keywords or authors. Input should be a search query string.`;
+
+  /**
+   * Constructs a new instance of GoogleScholarAPI.
+   * @param fields - Optional parameters including an API key.
+   */
+  constructor(fields?: GoogleScholarAPIParams) {
+    super(...arguments);
+
+    // Retrieve API key from fields or environment variables.
+    const apiKey =
+      fields?.apiKey ?? getEnvironmentVariable("SERPAPI_API_KEY");
+
+    // Throw an error if no API key is found.
+    if (!apiKey) {
+      throw new Error(
+        `SerpApi key not set. You can set it as "SERPAPI_API_KEY" in your environment variables.`
+      );
+    }
+    this.apiKey = apiKey;
+  }
+
+  /**
+   * Makes a request to SerpApi for Google Scholar results.
+   * @param input - Search query string.
+   * @returns A JSON string containing the search results.
+   * @throws Error if the API request fails or returns an error.
+   */
+  async _call(input: string): Promise<string> {
+    // Construct the URL for the API request.
+    const url = `https://serpapi.com/search.json?q=${encodeURIComponent(
+      input
+    )}&engine=google_scholar&api_key=${this.apiKey}`;
+
+    // Make an HTTP GET request to the SerpApi service.
+    const response = await fetch(url);
+
+    // Handle non-OK responses by extracting the error message.
+    if (!response.ok) {
+      let message;
+      try {
+        const json = await response.json(); // Attempt to parse the error response.
+        message = json.error; // Extract the error message from the response.
+      } catch (error) {
+        // Handle cases where the response isn't valid JSON.
+        message =
+          "Unable to parse error message: SerpApi did not return a JSON response.";
+      }
+      // Throw an error with detailed information about the failure.
+      throw new Error(
+        `Got ${response.status}: ${response.statusText} error from SerpApi: ${message}`
+      );
+    }
+
+    // Parse the JSON response from SerpApi.
+    const json = await response.json();
+
+    // Transform the raw response into a structured format.
+    const results =
+      json.organic_results?.map((item: any) => ({
+        title: item.title, // Title of the article or paper.
+        link: item.link, // Direct link to the article or paper.
+        snippet: item.snippet, // Brief snippet or description.
+        publication_info: item.publication_info?.summary
+          ?.split(" - ") // Split the summary at hyphens.
+          .slice(1) // Remove the authors from the start of the string.
+          .join(" - ") ?? "", // Rejoin remaining parts as publication info.
+        authors: item.publication_info?.authors
+          ?.map((author: any) => author.name) // Extract the list of author names.
+          .join(", ") ?? "", // Join author names with a comma.
+        total_citations: item.inline_links?.cited_by?.total ?? "", // Total number of citations.
+      })) ?? `No results found for ${input} on Google Scholar.`;
+
+    // Return the results as a formatted JSON string.
+    return JSON.stringify(results, null, 2);
+  }
+}

libs/langchain-community/src/tools/tests/google_scholar.int.test.ts

@@ -0,0 +1,36 @@
+import { test, expect, describe } from "@jest/globals";
+import { GoogleScholarAPI } from "../google_scholar.js";
+
+describe("GoogleScholarAPI", () => {
+  test("should be setup with correct parameters", async () => {
+    const instance = new GoogleScholarAPI({
+      apiKey: process.env.SERPAPI_API_KEY
+    });
+    expect(instance.name).toBe("google_scholar");
+  });
+
+  test("GoogleScholarAPI returns a string for valid query", async () => {
+    const tool = new GoogleScholarAPI({
+      apiKey: process.env.SERPAPI_API_KEY
+    });
+    const result = await tool.invoke("Artificial Intelligence");
+    expect(typeof result).toBe("string");
+  });
+
+  test("GoogleScholarAPI returns non-empty string for valid query", async () => {
+    const tool = new GoogleScholarAPI({
+      apiKey: process.env.SERPAPI_API_KEY
+    });
+    const result = await tool.invoke("Artificial Intelligence");
+    expect(result.length).toBeGreaterThan(0);
+  });
+
+  test("GoogleScholarAPI returns 'No results found' for bad query", async () => {
+    const tool = new GoogleScholarAPI({
+      apiKey: process.env.SERPAPI_API_KEY
+    });
+    const result = await tool.invoke("dsalkfjsdlfjasdflasdl");
+    expect(result).toBe("\"No results found for dsalkfjsdlfjasdflasdl on Google Scholar.\"");
+  });
+
+});