Add Google Vision OCR block to workflows

[brunopicinin]

Description

Adds a new workflow OCR block, based on Google Vision API. The block outputs the text for the whole image as well as detected language and sv.Detections(...) for text blocks with proper labels.

Type of change

Please delete options that are not relevant.

• New feature (non-breaking change which adds functionality)

How has this change been tested, please provide a testcase or example of how you tested the change?

1. Tested locally with an image of pure text, using ocr_text_detection mode:

2. Tested locally with a picture containing some text, using text_detection mode:

3. Tested locally with an image without text:

[CLAassistant]

All committers have signed the CLA.

[brunopicinin]

I ended up using the fullTextAnnotation for the blocks. Doing some tests, it seems to give "more natural" results than the textAnnotations for the text blocks. Also, it comes with confidence score on DOCUMENT_TEXT_DETECTION mode.

As an exemple, the difference in the Google Vision API output for the following image:

[PawelPeczek-Roboflow]

Great contribution 💪
I am very positively surprised on the speed and quality. The comments are minor - I am willing to accept the contribution and apply myself or you can do it if you wanted.

Do you have any comments / suggestions / feature requests regarding Workflows ecosystem as a result of this contribution?