Keystone: Data storage adapter ideas

[molomby]

In addition to the initial planned adapter (Mongoose and Knex #297), a number of other ideas have been tabled:

• SQLite -- Actually pretty powerful. Good for small apps and demo.
• In-memory DB -- An adapter without any external dependancies/processes that we can load with data out of a JSON file. This would be _super_ useful for testing. Possibly NeDB? Claims to implement the important parts of the Mongo API.
• OpenCRUD -- As in another, external GraphQL endpoint. This could be another KS 5 instance or any other system that conformed with OpenCRUD.
• Prisma -- Specifically the new Prisma Client. This would let devs wrap "any database" in a GraphQL layer then interface with it through a consistent data adapter.
• PostGraphile -- "Instant, secure and fast GraphQL API for Postgres"
• ORM adapters (eg. Objection or Bookshelf) -- Like the Knex adapter but also creates model classes and, internally, uses them instead of plain Knex "records". Good for larger systems?
• Write-only flat-file (JSON, CSV, text, etc.) -- A write only adapter that adds (structured?) lines to a file on disk. Could be a nice interface for logging data for later analysis? Either remotely or in-app
• Remote structured data services (Airtable, Google Sheets, etc.) -- Would be pretty cool...
• Other persistence engines -- Redis, Elasticsearch, ..?
• Data integrations -- Stripe, Algolia, Harvest? (Read-only?)
• System data sources -- File system, system metrics/resources, ..? Actually, probably just lists defs for osquery

Speaking of all this, our adapter framework might need to support something akin to "feature flags". That is, a way for a (Keystone or List?) adapter to communicate to Keystone which features it supports, eg. readable, writable, updatable, maybe "findable"? As in, can values be searched for or only referenced by ID?.

[Neo-Zhixing]

Would it be possible to add support for Google Firestore as well?

[ahmadawais]

I'd like to suggest adding the following:

• Google Cloud Firestore (Firestore pay as you go plan is excellent)
• Fauna Serverless Database with first-class GraphQL Support

[yangkghjh]

Flat-file support will be helpful to use keystone with other static website generators.

[molomby]

• Some kind of configurable REST adapter, as per #1573

[VinayaSathyanarayana]

Consider https://hasura.io

[stale[bot]]

It looks like there hasn't been any activity here in over 6 months. Sorry about that! We've flagged this issue for special attention. It wil be manually reviewed by maintainers, not automatically closed. If you have any additional information please leave us a comment. It really helps! Thank you for you contribution. :)

[JumpLink]

Instead of add support for NeDB, Keystone could add support for LinvoDB wich is based on NeDB, but has a Mongoose-like features and interfaces. Which could make it easier to add support for, because the current MongoDB adapter is using Mongoose. On the other hand, I would also be happy with SQLite over Prisma.

There is also a Mongoose Driver for LinvoDB, but this looks outdated.