Splitgraph is a tool for building, versioning and querying reproducible datasets. It's inspired by Docker and Git, so it feels familiar. And it's powered by PostgreSQL, so it works seamlessly with existing tools in the Postgres ecosystem. Use Splitgraph to package your data into self-contained data images that you can share with other Splitgraph instances.
Splitgraph.com, or Splitgraph Cloud, is a public Splitgraph instance where you can share and discover data. It's a Splitgraph peer powered by the Splitgraph Core code in this repository, adding proprietary features like a data catalog, multitenancy, and a distributed SQL proxy.
sgr (the command line client) or a local Splitgraph Engine, see the Installation section of this readme.
Splitfiles give you a declarative language, inspired by Dockerfiles, for expressing data transformations in ordinary SQL familiar to any researcher or business analyst. You can reference other images, or even other databases, with a simple JOIN.
When you build data with Splitfiles, you get provenance tracking of the resulting data: it's possible to find out what sources went into every dataset and know when to rebuild it if the sources ever change. You can easily integrate Splitgraph into your existing CI pipelines, to keep your data up-to-date and stay on top of changes to upstream sources.
Splitgraph images are also version-controlled, and you can manipulate them with Git-like operations through a CLI. You can check out any image into a PostgreSQL schema and interact with it using any PostgreSQL client. Splitgraph will capture your changes to the data, and then you can commit them as delta-compressed changesets that you can package into new images.
Splitgraph supports PostgreSQL foreign data wrappers. We call this feature mounting. With mounting, you can query other databases (like PostgreSQL/MongoDB/MySQL) or open data providers (like Socrata) from your Splitgraph instance with plain SQL. You can even snapshot the results or use them in Splitfiles.
Splitgraph isn't opinionated and doesn't break existing abstractions. To any existing PostgreSQL application, Splitgraph images are just another database. We have carefully designed Splitgraph to not break the abstraction of a PostgreSQL table and wire protocol, because doing otherwise would mean throwing away a vast existing ecosystem of applications, users, libraries and extensions. This means that a lot of tools that work with PostgreSQL work with Splitgraph out of the box.
The code in this repository, known as Splitgraph Core, contains:
sgrcommand line client:
sgris the main command line tool used to work with Splitgraph "images" (data snapshots). Use it to ingest data, work with splitfiles, and push data to Splitgraph.com.
Documentation is available at https://www.splitgraph.com/docs, specifically:
We also recommend reading our Blog, including some of our favorite posts:
dbtwith Splitgraph: versioning, sharing, cross-DB joins
sgrmust have access to Docker. You either need to install Docker locally or have access to a remote Docker socket.
For Linux and OSX, once Docker is running, install Splitgraph with a single script:
bash -c "$(curl -sL https://github.com/splitgraph/splitgraph/releases/latest/download/install.sh)"
This will download the
sgr binary and set up the Splitgraph Engine Docker container.
See the installation guide for more installation methods.
You can follow the quick start guide that will guide you through the basics of using Splitgraph with public and private data.
Alternatively, Splitgraph comes with plenty of examples to get you started.
curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | pythonto manage dependencies
git clone --recurse-submodules https://github.com/splitgraph/splitgraph.git
cd engine && make
The test suite requires docker-compose. You will also
need to add these lines to your
/etc/hosts or equivalent:
127.0.0.1 local_engine 127.0.0.1 remote_engine 127.0.0.1 objectstorage
To run the core test suite, do
docker-compose -f test/architecture/docker-compose.core.yml up -d poetry run pytest -m "not mounting and not example"
To run the test suite related to "mounting" and importing data from other databases (PostgreSQL, MySQL, Mongo), do
docker-compose -f test/architecture/docker-compose.core.yml -f test/architecture/docker-compose.mounting.yml up -d poetry run pytest -m mounting
Finally, to test the example projects, do
# Example projects spin up their own engines docker-compose -f test/architecture/docker-compose.core.yml -f test/architecture/docker-compose.core.yml down -v poetry run pytest -m example
All of these tests run in CI.