`dictionary` with large amounts of data

[andyundso]

Hello!

We use your gem to make copies of organizations within our application. There are many objects tied to an organization, and some things are indirectly circular, so I am using the dictionary option to clone a couple of things manually beforehand before then calling deep_clone over the entire organization. Also, without the dictionary, deep_clone sometimes cloned objects multiple times because of the dependencies.

But, the dictionary makes my system run out of memory, as there is too much stuff in the Hash. I am wondering if you have any tips to somehow reduce the memory footprint. Ideas that are in my head:

• Use the to_s representation of an object as key to map to the cloned object (or a SHA512 hash?).
• Allow to define that only certain types of objects need to be looked up in the dictionary (I somewhat know my problematic dependencies).

Appreciate your work and response.

[moiristo]

Hi,

Both seem to be sensible options. Maybe these optimizations should be optional and can be toggled via a dictionary_options param. Something like:

pirate.deep_clone(dictionary_options: { key: :itself, models: :all }) # default
pirate.deep_clone(dictionary_options: { key: :id, models: %i[mateys treasures] })

Could you maybe create a branch for something like this?