Group appfolio url

index.ts

const fs = require("fs");
const url = require("url");

type UrlInfo = {
  fullUrl: string;
  hostname: string;
  platform: string | undefined;
  path: string;
  count: number;
};

function isSameUrlInfo(a: UrlInfo, b: UrlInfo): boolean {
  return a.platform === b.platform && a.path === b.path;
}

function getPlatform(hostname: string): string | undefined {
  if (/appfolio/.test(hostname)) {
    return "appfolio";
  }
  return undefined;
}

function getPathPattern(pathname: string): string {
  return pathname
    .replace(/\w+-\w+-\w+-\w+-\w+/g, ":id")
    .replace(/\d+/g, ":number");
}

function parseUrl(urlStr: string): UrlInfo | undefined {
  try {
    const u = url.parse(urlStr);

    return {
      fullUrl: urlStr,
      hostname: u.hostname,
      platform: getPlatform(u.hostname),
      path: getPathPattern(u.pathname),
      count: 1,
    };
  } catch (e) {
    return undefined;
  }
}

function processUrls(urls: string[]): UrlInfo[] {
  const dict: Record<string, UrlInfo> = {};

  urls.forEach((url) => {
    const info = parseUrl(url);

    if (!info?.platform) {
      return;
    }

    if (/help/.test(info.hostname)) {
      return;
    }

    const key = `${info.platform}+${info.path}`;

    if (!dict[key]) {
      dict[key] = info;
    } else {
      dict[key].count++;
    }
  });

  const list = Object.values(dict);

  list.sort((a, b) => {
    return Math.sign(b.count - a.count) || (a.path < b.path ? -1 : 1);
  });

  return list;
}

function run(): void {
  fs.readFile("./urls.txt", "utf-8", (err: Error, data: string) => {
    if (err) {
      console.error("Failed to read file", err);
      return;
    }

    const txt = data.toString();
    const lines = txt.split("\n");
    const list = processUrls(lines);
    const result = list
      .map((item) => {
        const editOrAdd = /edit|new/.test(item.path);
        return `${item.platform}, ${item.path}, ${item.count}, ${editOrAdd}, ${item.fullUrl}`;
      })
      .join("\n");

    console.log(result);
  });
}

run();