Graphback: Create MongoDB DB layer on top of runtime

[wtrocki]

Currently graphback supports only knex based databases. We should create community supported mongo version

[udith-shalinda]

Do I need need to create an alternative file for KnexDBDataProvider

[wtrocki]

@udith-shalinda Yes. That will be the way forward. We can implement it in the same package for now for simplicity

[udith-shalinda]

@wtrocki Anyway I have created a separate file. I had to pass mongoose models from resolvers and return context.crudService.findBy("note", args.fields); will be changed to return context.crudService.findBy(MongooseModel, args.fields);

Can you give me a reference where I can find how to generate mongoose model?

[wtrocki]

Why not use mongo driver directly? Using mongose in this context makes less sense. Also please make sure you use latest changes on master. Current crud interface looks different

[wtrocki]

I have published new docs that explain the changes on master. Things will be much easier to integrate now - even when using mongoose

[udith-shalinda]

Now I'm using mongo driver instead of mongoose.

[udith-shalinda]

With the new version,I'm little bit confused about how you manage table name in curdservice

[wtrocki]

@udith-shalinda Mapping abstraction brings some complexity as people can specify different collection name on their model. Something like

""" db.name: custom_name """

We do not need to care about enabling this for mongo.
I guess for the initial version you can use the original name from the type or something like graphqlType.name.toLowerCase()

FYI @craicoverflow

[wtrocki]

Thanks to awesome contribution to @udith-shalinda we have now mongo supported (relationships/batching will land soon as well)