Skip to main content

Graph-based data exploration tool

Project description

https://travis-ci.org/ginkgobioworks/curious.svg?branch=master

Curious traverses relationships in a relational database. Curious queries allow users to explore relationships among objects, traverse recursive relationships, and jump between loosely connected databases. Curious also provides a JSON interface to the objects. Users and programmers can use Curious queries in analysis scripts and applications.

Curious favors a data centric model of application construction; Curious queries expose normalized, relational data, reducing UI dependency on UI specific API end-points serving denormalized data. Changing what data an UI needs no longer requires changing the UI specific end-points.

Curious works well with deep data models with many relationships. A Curious query can traverse 10s of foreign key like relationships efficiently. Curious queries always operate on sets of objects, and can connect a small number of objects via a relationship to a large number of objects, then via another relationship from the large number of objects to a smaller set again. For example, Book to Authors to Country of Residence. Unlike GraphQL, Curious outputs relationships between objects, rather than an ever growing tree of JSON representations of the objects.

Example

Book.last(10) Book.author_set Author.country(continent__name="North America")

Query Language

The query language allows traversing models by identfying the relationships between them, through foreign keys in Django models, or arbitrary id-mapping functions. A Curious query is a space-separated set of terms, which connect models together by relationships.

Several kinds of “joins” are possible using these relationship primitives:

  • A traditional inner join Book Book.author_set

  • A left outer join: Book.last(10) ?(Book.author_set)

  • A recusrive join: Parent.children_*

Furthermore, at each stage in a join, filtering can happen:

  • Filtering by Django field lookups: Book Book.author_set(id__in=[2,3,4])

  • Filtering by subquery: Book +(Book.author_set(id__in=[2,3,4]))

  • Filtering by exclusive subquery Book -(Book.author_set(id__in=[2,3,4]))

Finally, relationships can generate counts:

  • Counting Book Book.author_set__count

Configuring Curious

import myapp.models
from curious import model_registry

def register():
  model_registry.register(myapp.models)

Then include register when your Django app boots up.

Using Curious

Turn off CSRF. Deploy it as a Django app.

Writing Customized Relationships

Use filter and deferred to real functions.

Development

Requires Docker. Spin up your container using the provided docker-compose.yml file and Makefile by running make image. This creates an image with a correct git configuration for your user, which makes it easy to release. All of the commands you should need to run are defined the Makefile as targets. All of the targets except for image, are meant to be run inside the Docker container, but can be run from the host machine by having -ext appended to them. For example, to run tests, you could either call make test from inside the container, or make test-ext from the host.

If you are modifying the static assets during development, they can be recompiled with the build_assets make task, or by calling python setup.py build_assets.

./make test-ext

Deployment

Deployment of tagged commits happens to PyPI automatically via Travis CI. To bump and deploy a new version, run make bump/[foo]-ext, where [foo] is major, minor, or patch. Then git push origin --tags master.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

curious-1.2.0.tar.gz (1.8 MB view details)

Uploaded Source

Built Distribution

curious-1.2.0-py2-none-any.whl (1.9 MB view details)

Uploaded Python 2

File details

Details for the file curious-1.2.0.tar.gz.

File metadata

  • Download URL: curious-1.2.0.tar.gz
  • Upload date:
  • Size: 1.8 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No

File hashes

Hashes for curious-1.2.0.tar.gz
Algorithm Hash digest
SHA256 9c02a3b5fdf3172b1bb3d3f8c59b5fd2d93776da99efc26f79d95f05e532e472
MD5 ec50dd1c6a472ffbf34513c636afb61d
BLAKE2b-256 190cb15807f2a98560fe9b728aa146ad545d95e225db2d7ed7872f399dc9f694

See more details on using hashes here.

File details

Details for the file curious-1.2.0-py2-none-any.whl.

File metadata

File hashes

Hashes for curious-1.2.0-py2-none-any.whl
Algorithm Hash digest
SHA256 dbf314c2d998bba3cdb375c7b4af732d57fbda95798bf32f9a6be3442dd2a8bf
MD5 99df7d74c2691a6d2a6316423aec48a6
BLAKE2b-256 8281f0bb678a9a399c0b54edda5d10cd507508031965b51f4e3dd964b11a4e18

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page