Some neat little utilities relied upon by Regreso clients. Right now, these can be used to:

• Get metadata info about websites (for prefilling destination details automatically)
• AI-generate a list of practically-usable, relevant tags to describe a location based on metadata (supports provided, self-hostable AI instance URL)

Let's say you wanted to use this module in your own regreso client (or something entirely different), here's how:

• SiteTagger(config) - A class used to control AI-based website tagging behaviors.

  import { SiteTagger } from "@regreso/utils";
  
  async function demo() {
    const tagger = new SiteTagger({
      maxTags: 3,
      aiInstance: "https://ai.hackclub.com",
      maxRetries: 2,
      requestDelay: 1000,
    });
  
    const demoSites = [
      {
        url: "https://github.com/microsoft/vscode",
        headline: "Visual Studio Code",
        description: "A lightweight but powerful source code editor",
      },
      {
        url: "https://css-tricks.com/snippets/css/a-guide-to-flexbox/",
        headline: "A Complete Guide to Flexbox",
        description: "CSS Flexbox layout guide with examples",
      },
    ];
  
    const result = await tagger.generateTags(demoSites[0]);
  
    if (result.success) {
      console.log(
        `tags: ${result.tags.join(", ")} in ${result.metadata?.processingTime} ms.`,
      );
    } else {
      console.log(`error: ${result.error}`);
    }
  
    const batchResults = await tagger.generateTagsBatch(demoSites);
  
    console.log(
      batchResults
        .filter((r) => r.result.success)
        .map((r) => `${r.headline}: [${r.result.tags.join(", ")}]`)
        .join("\n"),
    );
  }
  
  demo();

• getWebDetails(url: String) - Gets the web details scraped from meta tags of specified website's URL. Returns multiple versions if relevant (title, og:title, twitter:title to be specific)

  import { getWebDetails } from "@regreso/utils";
  
  async function demo() {
    const webDetailsResult = await getWebDetails("fbi.gov");
  
    console.log({
      url: webDetailsResult.url, // with protocol appended if missing
      title:
        webDetailsResult.title[0] ??
        webDetailsResult.title[1] ??
        webDetailsResult.title[2],
      description:
        webDetailsResult.description[0] ??
        webDetailsResult.description[1] ??
        webDetailsResult.description[2],
    });
  }
  
  demo();

• Add site metadata scraping
• Add AI-tagging
• Document package
  • Basic Usage Demo
• Support scraping image metadata