[FR] Embed OCR natively into ss object for accurate search

[AnsahMohammad]

Currently the search is limited to what is in the note or searched by the AI, but it can be made more accurate by option of using the OCR of what is in the image.

Only works OCR supported lang images.

This can also make OCR copy of images instant

• Add a field in SS object to store OCR metadata
• Include options in search to exlude sepcific fields such as notes/AI summary/tags/OCR

can include this as another service, which goes through the images and can do this in background. And now when the new images are loaded by the filewatcher this service is called and action is run in background to index the data.

[mriddyagrawal]

Hmm, do you think including the OCR data directly into the same field as the AI notes could be better? Or like, creating a new field under the ss. I think creating a copy could be overkill and inefficient, so we will have to make the links as well between ocr and source images.

Guessing Tesseract can work locally? Forgive me, I am close to a noob, but I would love to work on this, it seems like a nuclear enough task for me.

Would this be using the existing service, which works when we click the "T" button on ss pages?