Update the response to JavaScript object instead of string

[ayungavis]

Currently, we used the original response from the Google Cloud Vision API and it's a string with a line break (\n) character for each word. So, we want to make it dynamic. It should have a pattern to choose which text you want to extract from the image.

For example, if we extract Indonesia's resident ID (KTP), we only want to get the name and the date of birth. So, we need to add a pattern to the extract function. It should look like this:

public async extract<T>(image: string, { pattern: string }) {
  // code here
}

and if we call it based on the example above, it should look like this:

interface Response {
  name: string;
  birthdate: string;
}

const extractor = new ImageTextExtraction({ apiKey: 'your-key' });
const response = await extractor<Response>(image, { 
  pattern: 'nama|tanggal lahir' // pattern text based on image
}); 

then, the response should be like this:

{
  name: 'John Doe',
  birthdate: '1 January 2023'
}