Data contracts

Formal Agreements for Scientific Data Exchange

Standardize the structure of scientific exchange to enable seamless reuse across research domains.

See contract taxonomy
Explore the toolkit

Standard

Format agnostic

Defines content/structure, not the file type—YAML is common, JSON also possible.

Cheks

Automated validation

Supports compliance checks prior to exchange or publication using a Python library.

Bilateral

Reusable beyond CROSS

Not CROSS-specific—usable for bilateral exchange as well as via the platform.

Standardize the structure of scientific exchange to enable seamless reuse across research domains.

See contract taxonomy
Explore the toolkit

About

Data contracts

Why contracts matter?

When contracts are shared, teams can exchange scenario inputs/outputs using their agreed upon semantics: field names, types, constraints, and references are explicit and machine-checkable.

Exchange workflows enabled

Data contracts support both bilateral exchange between trams and exchange via CROSSHub, with validation either local or platform enforced

Bilateral exchange

Local validation against the agreed contract; basic quality checks; exchange via common tabular formats.

Via CROSSHub (centralized)

Automatic validation against the associated contract rules on the platform; support ingestion/export in common tabular formats.

Works with any data model

Contracts aren't limited to the CROSS data model. Any model can exchange data as long as it complies with the contract.

What is inside a data contract?

Contracts have two main parts: metadata and schema. Metadata documents wha the dataset represents, ownership, and provenance. The schema defined the logical table structure and any existing data constraints.

Metadata	Schema
Descriptive information Title and description of the dataset	Fields Column names + data types, optionally description
Ownership and licensing Provenance of the data and rules for reusing it	Constraints Required fields, ranges and other rules for validation
Keys and references Primary keys for uniqueness and foreign keys linking to other contracts

Metadata

Schema

Descriptive information

Title and description of the dataset

Fields

Column names + data types, optionally description

Ownership and licensing

Provenance of the data and rules for reusing it

Constraints

Required fields, ranges and other rules for validation

Keys and references

Primary keys for uniqueness and foreign keys linking to other contracts

Python toolkit

CROSSContract + CROSSClient

To make data contracts usable, CROSS provided a Python package with two parts: CrossContract for creating/validating contracts locally and CrossClient for interactive with the platform (retrieve contracts, create contracts, automate submissions)

CROSSContract

Create contracts and validate them

CROSSClient

Interact with the CROSS platform: retrieve/create contracts.

Go to documentation