Hugging Face's logo

Datasets:
PierreMesure
/
open-sfs

Modalities:
Text
Formats:
parquet
Languages:
Swedish
Size:
10K - 100K
Tags:
sweden
legislation
rättsdatabas
författningar
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files Community
1
open-sfs / README.md
PierreMesure's picture
PierreMesure
Update README
3e463a5
preview code
|
raw
history blame contribute delete
2.61 kB
metadata
license: apache-2.0
language:
  - sv
tags:
  - sweden
  - legislation
  - rättsdatabas
  - författningar
size_categories:
  - 10K<n<100K
configs:
  - config_name: default
    data_files:
      - split: train
        path: data.parquet

Open SFS

Open SFS is a dataset gathering the content of the Swedish code of statutes (svensk författningssamling). It consists of all the legal Swedish texts in their text version with metadata, as well as the official PDF version when available.

This data is gathered from various sources using a script, also available on Github.

What data is in the dataset

The goal of this dataset is to gather as much high-quality data as possible from the Swedish code of statutes in a format that is easy to reuse.

Today, the Swedish code of statutes is available in various forms on several different websites:

  1. All the laws (lag) and regulations (förordning) are available in text version with basic metadata on a portal of the government chancellery rkrattsbaser.gov.se. The data in the dataset is extracted from the beta version since it uses a modern API (Elasticsearch's proxy API). This portal only provides the most recent version of the texts, it does not have the original versions nor the content of the laws and regulations that modify existing texts (ändringsförfattningar).
  2. Original versions of the texts are available in PDF format (both main texts and modifying regulations): a. from 1998:306 to 2018:159 at rkrattsdb.gov.se b. from 2018:160 to now at svenskforfattningssamling.se

In addition to these official sources, more data can be accessed from unofficial sources:

  1. Lagen.nu has a sparsely documented API to acess the laws and regulations in both JSON and XHTML. The texts have been slightly reformated so they have a better structure than the versions found in (1). Modifications are not available.
  2. Lagboken.se has no API but contains all the final versions of the texts as well as PDF versions of all texts and modifying texts. Before 1993, PDF versions are scanned from the yearly book versions of the legislation. From 1993 to 1998, they are digital PDFs with the same format as those found at (2a).

For now, the content in the dataset consists of texts and metadata from (1) and PDF documents from 1993 when available from (2a), (2b) and (4).