We're working on a newer, leaner, more modular, and more interoperable solution to the same challenge that the current datalad-catalog aims to address. This new development is taking place within the broader context of making DataLad datasets interoperable with linked and semantic (meta)data. For more background, see this issue. To keep up to date, follow progress at psychoinformatics-de/datalad-concepts, psychoinformatics-de/shacl-vue, and in the new development branch. Because of this redirected focus, datalad-catalog itself will be downscaled by focusing on maintenance and assessing the priority of new features on a case-by-case basis.
DataLad Catalog is a free and open source command line tool, with a Python API, that assists with the automatic generation of user-friendly, browser-based data catalogs from structured metadata. It is an extension to DataLad and forms part of the broader ecosystem of DataLad's distributed metadata handling and (meta)data publishing tools.
This software was developed with support from:
- the German Federal Ministry of Education and Research (BMBF 01GQ1905)
- the US National Science Foundation (NSF 1912266)
- the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under grant SFB 1451 (431549029, INF project).
- the MKW-NRW: Ministerium fΓΌr Kultur und Wissenschaft des Landes Nordrhein-Westfalen under the Kooperationsplattformen 2022 program, grant number: KP22-106A
Navigate to https://datalad-catalog.netlify.app/ to view a live demo of a catalog generated with DataLad Catalog.
This demo site is hosted via Netlify and it builds from the datalad_catalog/catalog directory of the main branch of this repository.
DataLad Catalog can receive commands to create a new catalog, add and remove metadata entries to/from an existing catalog, serve an existing catalog locally, and more. Metadata can be provided to DataLad Catalog from any number of arbitrary metadata sources, as an aggregated set or as individual metadata items. DataLad Catalog has a dedicated schema (using the JSON Schema vocabulary) against which incoming metadata items are validated. This schema allows for standard metadata fields as one would expect for datasets of any kind (such as name, doi, url, description, license, authors, and more), as well as fields that support identification, versioning, dataset context and linkage, and file tree specification.
The process of generating a catalog, after metadata entry validation, involves:
- aggregation of the provided metadata into the catalog filetree, and
- generating the assets required to render the user interface in a browser.
The output is a set of structured metadata files, as well as a Vue.js-based browser interface that understands how to render this metadata in the browser. What is left for the user is to host this content on their platform of choice and to serve it for the world to see.
With your virtual environment manager of choice, create a virtual environment and ensure
you have a recent version of Python installed. Then activate the environment. E.g. with venv:
python -m venv my_catalog_env
source my_catalog_env/bin/activate
Run the following from your command line:
pip install datalad-catalog
If you are a developer and would like to contribute to the code, instead clone the code base from GitHub and install with pip local changes :
git clone https://github.com/datalad/datalad-catalog.git
cd datalad-catalog
pip install -e .
Congratulations! You have now installed datalad-catalog.
Because this is an extension to datalad and builds on metadata handling functionality, the installation process also installs datalad and datalad-metalad as dependencies, although these do not have to be used as the only sources of metadata for a catalog.
While the catalog generation process does not expect data to be structured as DataLad datasets, it can still be very useful to do so when building a full (meta)data management pipeline from raw data to catalog publishing. For complete instructions on how to install datalad and git-annex, please refer to the DataLad Handbook.
Similarly, the metadata input to datalad-catalog can come from any source as long as it conforms to the catalog schema. While the catalog does not expect metadata originating only from datalad-metalad's extractors, this tool has advanced metadata handling capabilities that will integrate seamlessly with DataLad datasets and the catalog generation process.
The overall catalog generation process actually starts several steps before the involvement of datalad-catalog. Steps include:
- curating data into datasets (a group of files in an hierarchical tree)
- adding metadata to datasets and files (the process for this and the resulting metadata formats and content vary widely depending on domain, file types, data availability, and more)
- extracting the metadata using an automated tool to output metadata items into a standardized and queryable set
- in the current context: translating the metadata into the catalog schema
- in the current context: using
datalad-catalogto generate a catalog from the schema-conforming metadata
The first four steps in this list can follow any arbitrarily specified procedures and can use any arbitrarily specified tools to get the job done. If these steps are completed, correctly formatted data can be input, together with some configuration details, to datalad-catalog. This tool then provides several basic commands for catalog generation and customization. For example:
# CREATE a new catalog from scratch:
datalad catalog-create -c /tmp/my-cat
#ADD metadata to an existing catalog:
datalad catalog-add -c /tmp/my-cat -m path/to/metadata.jsonl
# SET a property of an existing catalog, such as the home page of an existing catalog - i.e. the first dataset displayed when navigating to the root URL of the catalog:
datalad catalog-set -c /tmp/my-cat -i abcd -v 1234 home
# SERVE the content of the catalog via a local HTTP server at http://localhost:8001:
datalad catalog-serve -c /tmp/my-cat -p 8001
# VALIDATE metadata against a catalog schema without adding it to the catalog::
datalad catalog-validate -c /tmp/my-cat/-m path/to/metadata.jsonl
# GET a property of an existing catalog, such as the catalog configuration:
datalad catalog-get -c /tmp/my-cat/ config
# REMOVE a specific metadata record from an existing catalog:
datalad catalog-remove -c /tmp/my-cat -i efgh -v 5678
# TRANSLATE a metalad-extracted metadata item from a particular source structure into the catalog schema. A dedicated translator should be provided and exposed as an entry point (e.g. via a DataLad extension) as part of the 'datalad.metadata.translators' group:
datalad catalog-translate -c /tmp/my-cat -m path/to/metadata.jsonl
# RUN A WORKFLOW for recursive metadata extraction (using datalad-metalad), translating metadata to the catalog schema, and adding the translated metadata to a new catalog:
datalad catalog-workflow -t new -c /tmp/my-cat -d path/to/superdataset -e metalad_core
# RUN A WORKFLOW for updating a catalog after registering a subdataset to the superdataset which the catalog represents. This workflow includes extraction (using datalad-metalad), translating metadata to the catalog schema, and adding the translated metadata to the existing catalog:
datalad catalog-workflow -t new -c /tmp/my-cat -d path/to/superdataset -s path/to/subdataset -e metalad_coreTo explore the basic functionality of datalad-catalog, please refer to the tutorial in the DataLad Handbook.
The DataLad ecosystem provides a complete set of free and open source tools that, together, provide full control over dataset/file access and distribution, version control, provenance tracking, metadata addition/extraction/aggregation, and catalog generation.
DataLad itself can be used for decentralised management of data as lightweight, portable and extensible representations. DataLad MetaLad can extract structured high- and low-level metadata and associate it with these datasets or with individual files. And at the end of the workflow, DataLad Catalog can turn the structured metadata into a user-friendly data browser.
Importantly, DataLad Catalog can operate independently as well. Since it provides its own schema in a standard vocabulary, any metadata that conforms to this schema can be submitted to the tool in order to generate a catalog. Metadata items do not necessarily have to be derived from DataLad datasets, and the metadata extraction does not have to be conducted via DataLad MetaLad.
Even so, the provided set of tools can be particularly powerful when used together in a distributed (meta)data management pipeline.
Please create a new issue if you have any feedback, comments, or requests.
If you'd like to contribute as a developer, you need to install a number of extra dependencies:
cd datalad-catalog
pip install -r requirements-devel.txt
This installs sphinx and related packages for documentation building, coverage for code coverage,
black for linting, and pytest for testing.
To make a contribution to the code or documentation, please:
- create an issue describing the bug/feature
- fork the project repository,
- create a branch from
main, - commit your changes,
- check that linting tests succeed: from the project root directory, run
black . - check that tests succeed: from the project root directory, run
python -m pytest - push your commits to your fork
- create a pull request with a clear description of the changes
- check that all continuous integration tests succeed on the pull request
Thanks goes to these wonderful people (emoji key):
This project follows the all-contributors specification. Contributions of any kind welcome!