Implement OCR

[FoxxMD]

Detecting text on an image would be an excellent feature for CM:

• detect offensive phrases/words
• detect spam links
• detect watermarks (potentially)

Some bots that do things similar to this:

• Hedonism Bot
• MAGIC_EYE_BOT
• ocrbot
• tiktok-link

Not sure how strong JS is at OCR...might be time to figure out how to interface with a python library.

• tesseract.js
• textract
• ocrad.js (gocr.js)

[FoxxMD]

openCV implemented in this commit opens the possibility for using it for OCR.

[CuriousLyz]

Any updates? It can be useful for watermark spam