Stocks Scrape

scrape-finance.js

const request = require("request-promise");
const cheerio = require("cheerio");
const NUMBER_REGEX = "\\-?\\+?[0-9]{1,3}(,[0-9]{3})*(.[0-9]+)?";
const stockPriceRegex = () =>
  new RegExp(
    `Stock Price(?<price>${NUMBER_REGEX})\\s(?<day_change>${NUMBER_REGEX})\\s(?<day_percent>\\(?\\d+.\\d+%)`,
    "gm"
  );
const afterHoursStockPriceRegex = () =>
  new RegExp(
    `After hours:\\s(?<price>${NUMBER_REGEX})\\s(?<day_change>${NUMBER_REGEX})\\s(?<day_percent>\\(\\d+.\\d+%\\))`,
    "gm"
  );

const stockCard = {
  priceRegex: stockPriceRegex,
  ahRegex: afterHoursStockPriceRegex,
  name: /Stock Price \| (?<name>[^.]+)/gm
};

class StocksController {
  async get(ticker) {
    try {
      const stockPrice = await this.scrape(ticker);
      return stockPrice;
    } catch (e) {
      res.json({
        name: "N/A",
        price: "N/A",
        day_change: "N/A",
        day_percent_change: "N/A"
      });
    }
  }

  async scrape(ticker) {
    if (!ticker) {
      return {};
    }

    const body = await this.getHtml(ticker);
    return this.parseHtml(body, ticker);
  }

  async parseHtml(html, ticker) {
    const { priceRegex } = stockCard;

    const $ = cheerio.load(html);
    const text = $.text();

    // const name = text.getNameFromTicker(ticker);
    const name = "N/A";
    const match = priceRegex().exec(text);

    if (!match) {
      throw new Error(text);
    }

    const { price, day_change, day_percent } = match.groups;
    return {
      name: name,
      price: price,
      day_change: day_change,
      day_percent_change: day_percent,
      after_hours: this.getAfterHours(text)
    };
  }

  async getHtml(stock) {
    const options = {
      userAgent:
        "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36",
      method: "GET",
      url: `https://www.google.com/search?q=${stock} stock price`
    };

    return await request(options);
  }

  getAfterHours(text) {
    const { ahRegex } = stockCard;
    const match = ahRegex().exec(text);
    if (!match) {
      return {};
    }

    const { price, day_change, day_percent } = match.groups;
    return {
      price: price,
      day_change: day_change,
      day_percent_change: day_percent
    };
  }
};


(new StocksController()).get("tsla")