Extracting HTML from GMail

gmail-html.utils.ts

  /**
   * search for email with gmailSearchQuery (e.g. "to: someone@gmail.com subject: spam")
   * and get the raw content of the mail
   */
  async searchEmail(gmailSearchQuery: string): Promise<any[]> {
    const auth = await this.authorize();
    const gmail = google.gmail({ version: 'v1', auth });
    return new Promise((resolve, reject) => {
      gmail.users.messages.list(
        {
          userId: 'me',
          q: gmailSearchQuery
        },
        async (err, res) => {
          if (err) reject(err);
          if (res.data.messages === undefined) {
            resolve([]);
          } else {
            resolve(
              await Promise.all(
                res.data.messages.map(m => this.getRawMessage(gmail, m.id))
              )
            );
          }
        }
      );
    });
  }
  
  /**
   *  with the message id, get the raw mail content
   */
  private getRawMessage(gmail: gmail_v1.Gmail, message_id: string) {
    return new Promise((resolve, reject) => {
      gmail.users.messages.get(
        {
          userId: 'me',
          id: message_id,
          format: 'raw'
        },
        (err2, res2) => {
          if (err2) reject(err2);
          resolve({
            id: message_id,
            raw: res2.data.raw
          });
        }
      );
    });
  }

  /**
   * with the raw mail content, extract the HTML and load it to cheerio
   */
  getHtmlFromEmailBody(raw: string) {
    const message = Buffer.from(raw, 'base64')  //convert base64 to text and clean it up
      .toString()
      .split(/(\r\n|\n|\r)/)
      .map(s => s.trim())
      .filter(s => s.length > 0)
      .map(s => s.replace(/=3D/g, '='));
    const formattedMessage: string[] = [];
    let currentLine = '';
    let start = false;
    for (const line of message) {
      if (line.indexOf('DOCTYPE html') !== -1) {
        start = true;
      }
      if (start && line.startsWith('----')) {
        break;
      }
      if (start) {
        if (line.slice(-1) === '=') {
          currentLine += line.slice(0, -1);
        } else {
          currentLine += line;
          formattedMessage.push(currentLine);
          currentLine = '';
        }
      }
    }
    return cheerio.load(formattedMessage.join());
  }