Python - Convert TIFF, PDF etc. to JPEG in-memory for input to Google Cloud Vision

Question

I have a number of non-JPEG image files that I want to process using Google Cloud Vision, but the API only accepts certain formats (see question Cloud Vision API - PDF OCR and answer https://cloud.google.com/vision/docs/supported-files).

I can use PIL or some such to convert a TIFF to JPEG to be uploaded, but I'd like to avoid a temp file if possible.

So, in python, how do I convert a TIFF in-memory for upload to GCV? numpy array, base64, string..?

Answer 1

As suggested, you should use StringIO or BytesIO for Python 3.

See this question: Python resize image and send to google vision function

Example making a request string:

output = BytesIO()
image_object.save(output, 'JPEG', quality=90)
contents = output.getvalue()
output.close()
ctxt = b64encode(contents).decode()

request_string = [{
    'image': {'content': ctxt},
    'features': [{
        'type': 'TEXT_DETECTION',
        'maxResults': 1
        }]
}]