rivo / duplo Goto Github PK

Hello, apologies if this is not the best way to reach out, was just wondering if there is a suggested way to retrieve a Hash that had been added to a Store.

The use case is for comparing images in 2 separate directories. I hope to create a store of all images in dir1 as store1 and images in dir2 as store2. The stores are then compared by iterating over all the IDs of one store1 , retrieving the Hash for an ID and Querying it against store2. And then doing it the other way around. Hope this makes sense. Any input is greatly appreciated.

Hi, thanks for this nifty package! I have found that it works quite well in a few spot checks so far. My tests with perceptual hashes were disappointing, with results appearing to be mostly random. I see you've discovered this, too. I'll keep testing this lib but so far it's better than other algorithms I've tried and doesn't require ML models! (It is quite slow, I've noticed. Depending on how heavily I use this, I may investigate to see if we can speed it up.)

Ok, question:

Is there a recommended way to serialize a hash so it can be stored and later reused? I am used to hashes simply being a number (albeit, a very large number, 128 to 512 bits in length), so I'm not sure how to handle a struct with various fields constituting a 'hash'.

I see there is an EncodeGob() method on a Store, implying that gob-encoding the hash might be the "standard" way.

If I wanted to store these hashes in a database, what do you recommend?

python 3.3 and above

https://github.com/rif/imgdup2go