# Introduction

Local Offline Assessor for FAIR (LOAF) is a Python library for checking metadata: the information
that describes a dataset, such as its title, creator and licence. The checks
address FAIR principles: Findable, Accessible, Interoperable and Reusable.

An assessor defines what to check and, where applicable, how to assign points.

## Supported assessors

| Assessor                                                                  | Latest supported version | Checks offline / total  | Metrics with offline checks / total |
| ------------------------------------------------------------------------- | ------------------------ | ----------------------- | ----------------------------------- |
| [F-UJI](/local-offline-assessor-for-fair-loaf/assessors/fuji)             | 3.5.1                    | 24 / 31                 | 14 / 17                             |
| [FAIR Champion](/local-offline-assessor-for-fair-loaf/assessors/champion) | Core Tests 0.5.12        | 13 + 2 conditional / 16 | 12 / 13                             |

Versions are those supported by this library. A metric groups checks; the metric
count includes partial support. F-UJI supports all checks in 13 metrics and some
checks in one more. Champion's two conditional checks can decide only some cases
offline; search indexing is unsupported. These counts describe capability, not
how many checks will reach a decision for every input.

## How it works

1. Choose an assessor and version.
2. Supply the dataset's metadata.
3. The assessor interprets the metadata and applies its own checks.
4. Read the results, messages and any scores the assessor provides.

## What offline means

The library checks the information you supply. It does not download datasets,
open their links or contact external services. Checks that need information
unavailable offline are marked `indeterminate` and receive no score.

You can use it to check metadata before publishing it or in an application
without network access.

## Start here

Use the [quickstart](/local-offline-assessor-for-fair-loaf/quickstart) to run an assessment. The [Python API](/local-offline-assessor-for-fair-loaf/reference)
explains the arguments; [Results](/local-offline-assessor-for-fair-loaf/results) explains the response.
The [F-UJI](/local-offline-assessor-for-fair-loaf/assessors/fuji) and [FAIR Champion](/local-offline-assessor-for-fair-loaf/assessors/champion) pages explain
their requirements and offline limits.

Developed for [EOSC Data Commons](https://www.eosc-data-commons.eu/).