After a few days of trying to get it right, I figure this is more than likely the best place for me to ask.

Does anybody have any suggestions on achieving the task of automating image analysis and trait matching, in order to build out 8006 nft metadata files automatically and accurately?

Details below.. 👇🏼

I am assisting with an NFT collection.

The original dev bailed and tanked a lot of the data. The individual traits are gone, but the generated images remain. 

Of 10k, I have 1294 complete metadata files, I need to identify the traits and build out the remaining 8006 files following the identical schema.👇🏼

I have been able to accomplish much of the automation necessary, but the image analysis and visually memorizing the traits based off of the images and their respective metadata, building a dictionary to use with automatically determining the remaining NFTs traits in order to accurately create the metadata. 👇🏼

I have had zero luck with this function. 

So a short recap more to the point..

I need to..

- Identify the traits of 8006 NFTs and create their metadata files accurately.

What I need..

- Advice on any method to accomplish automating this task.

If anyone has any insight it would be appreciated. Thank You All!