You can clone with
HTTPS or Subversion.
Hi, I'm working on https://github.com/ifesdjeen/entrepot which is in many ways similar to Curator, but the primary store (for now) is Mongodb, and https://github.com/solnic/virtus is used for property bags, but we're planning to support ElasticSearch and Riak quite soon.
I enjoy the idea of Repository pattern implementations a lot, and want to see more good libraries that are aiming to make writing big apps easier.
My question is mostly wether you guys are planning to support foreign (or even embedded, that'd work for mongo, elasticserach, riak and others, too) relationships, and use something that has more features for property bags.
Curator seems to get some attention, but unfortunately for now I can't use it in production because of the mentioned things, and some others (like - set record ID whenever record got inserted to the database, or allowing usage of batch operations on any data store, and atomic modifiers wherever they're supported).
https://github.com/ifesdjeen/entrepot looks very interesting. There is definitely a lot of overlap with Curator.
We haven't needed explicit relationship support yet, but we're not against adding it. Right now, we do it by having Repository classes use other Repositories in the serialize/deserialize methods. Or we have the serialize method flatten an object hierarchy into a nested hash (and the deserialize does the opposite).
I've started playing with a way to specify repository mappings in a declarative fashion. I'd like to expand this to cover simple data types such as dates and times. You can check out my playground at 0012233. I will take a closer look at https://github.com/solnic/virtus and see if it's a good fit.
Curator does set the id (key) when new objects are stored.
Thank you Paul,
Sorry, probably I've done something wrong so the ID haven't been setting. It was an important thing for me, so I'll try again during the week.
It's also good to see an effort to bring repository mappers live! Will keep following your progress.