1Base is a project which aims to provide an open source platform for open data. And just just any data. Lots of data. From various sources.
You can think of 1Base as a wiki for data.
What's more, 1Base is a database of the future. In addition to storing primitive
types (INTEGER, BOOLEAN, TEXT, ...), it stores anything from images to
video to the reason why you walk your dog.
The key here is representations. See the section on slots for more information.
Let's say you want to build an app that pulls data from an OpenWeather database and OpenStreetMap database. Currently you have to manage 2 endpoints, and you have to get 2 different API keys.
And this is already assuming you've been lucky enough to find OpenWeather and OpenStreetMap as the data provider you need.
If data is gold, and open data is going to change the world, why is it still fragmented? Shouldn't there be a central place to access it?
If you're wanting to get up and running developing 1Base, please follow the directions in the docs.
A library that's used between onebase_api and onebase_web. Contains common settings and testing utilities.
Currently the aim is to get 100% JSON output with logging. The typical logging
library has this functionality, but unit tests still use the standard unittest
output format.
onebase_api is a project for the model infrastructure (what defines the
Nodes, paths, etc.) and the API views. (e.g. /api/validator/1.0/string).
In the future the model infrastructure may split off to be its own project.
The API uses Flask as its view framework with MongoEngine as the model backend framework.
A pretty (so far, pretty ugly!) front-end to manage user accounts and provide a graphical, non-API method to manage the 1Base documents.
Like onebase_api, it's written in Flask, but uses HTML templates. onebase_web has no internal backend, but uses onebase_api's model backend.
This does need a makeover, so if anyone loves to tinker around with webkits, I'm welcome to a pretty design.
A heirarchical navigation technique to reach a node.
You can think of a path as a directory structure on a filesystem.
You can think of a key as a Column on a table. The difference is that a Key houses a specific Type, and each Type contains a reference to a Validator and Representer.
Keys are unique to a Node.
A Node is the backbone to 1Base. It's equivalent to a MySQL table and contains rows of Slots.
Each Path can contain either 0 or 1 Node, but any number of paths can point to a node.
Since nodes are heirarchical, they can either be reached by the path name or by browsing.
A slot is similar to a MySQL cell (or value), with one uniqure trait: the value initially returned is not the final value intended to be returned to the user. Since 1Base can store any type of value (including images and audio) this value is unrepresented, and must be passed to a microservice to be represented to the user.
For example, assume a slot contains the value https://example.com/image.png
and the type is IMAGE. The value https://example.com/image.png is passed
to an API microservice https://api.example.com/1.0/repr/image/ to grab the
image from the server and write it to the stream.
Before this occurs, however, the slot value needs to be validated. A validator is also--you guessed it--a microservice.
You know how on Wikipedia, you must source articles and claims? You must do the same on 1Base.
A source is simply an acknowledgemnt of the originator of the data. Sources are loaded into the database using a method, which can be one of. The method can be externally generated (a server uploading data to 1Base), internally generated (1Base's own aggregation), or manually input (e.g. from a book text).
Sources can be assigned to a Slot or Node. If a Source is assigned to a Node, then every Slot within the node inherits the Source.
Sometimes a discrepancy will arise with a document. Like Wikipedia, 1Base facilitates Discussions between users. Like Sources, Discussions can be assigned to Nodes and Slots. Discussions can also be assigned to Keys or Sources.
I'm hoping this project will evolve into something helpful. So far the major needs are:
- Investing. FLOSS software is awesome but it comes at a price. Please consider Donating on Patreon.
- Developers. Got mad coding skills? Come contribute!
- Contributed Data. People hate the sound of crickets. Let's fill this sucker up!