LibCrowds/libcrowds: v1.0.0-beta.9

This is a big release that contains a lot of additional functionality and fixes. Some highlights are presented below. This should be our last beta release.

Project generation

One of the major changes in this version is to do with the ways in which projects can now be generated from pre-configured templates. These templates define how each task is presented and how its results are analysed. This should make it relatively straightforward for users to add new types of project, without any coding knowledge. It also means that we now can upgrade part of a template (e.g. the tutorial) and have those changes instantly pushed out to all projects that were built from that template.

Results analysis

Our new results analysis process now works consistently across all project types. In essence, multiple contributions are normalised and compared, with the result stored only if the chosen minimum number of matching answers is achieved. If we do not meet that minimum number we can automatically send the task back and ask for an additional contribution. The normalisation rules are configurable via the admin interface.

Web Annotations

The results for all project types are now stored as Web Annotations. This means that, as well as developing consistent ways of analysing the results, we can develop consistent ways of consuming them, such as performing searches via the API. This should prove especially useful when building additional interfaces that need to act in the same way across a range of collection microsites that may be using different task presenters (e.g. Z39.50 and IIIF).

Dependencies upgraded

LibCrowds has been developed using cutting edge software libraries, many of which have now left their alpha/beta stages (e.g. Nuxt.js and Bootstrap 4). Upgrading to the stable versions of the libraries in turn makes LibCrowds more stable.

Documentation

The documentation has been moved to a separate repository. All major functionality has been documented.

Volumes

Projects are now grouped by volume, which helps with the project generation mentioned above and collation of the results. The available volumes for a collection are configurable via the admin interface. Each can use a different type of importer (e.g. Flickr, IIIF or Google Docs).

Forum integration

LibCrowds now integrates with Flarum forums. Rather than having to create a separate account for the forum, single sign-on means that users can sign in to the forum using their LibCrowds account.

Redesigned interfaces

Tweaking of the many interfaces within the platform has been an ongoing process. However, of note in this version is the addition of a new slide out menu that remains consistent across all pages. The various dashboard pages have also been refactored to improve usability.

Consistent and flexible task presenters

The info, help and share buttons are now handled consistently for all task presenters. Previously, some of these buttons only appeared for the IIIF task presenter but now they will also appear for the Z39.50 and any future task presenters. Task presenters are now defined at the template level, rather than the collection level. so we could potentially have multiple types of project within each collection.

Project tags refactored

To enable the sorting and filtering of projects we previously had to specify a set of available tags for each collection then add a selection of these defined tags to each project. Now, we just define some tag types for each collection (e.g. Task or Location) and either choose from current tags or enter new ones as free text when building a new project. This essentially cuts out a slightly clunky configuration stage.

Annotation exports

LibCrowds now provides downloadable datasets for all final result Annotations for each collection. As with the project downloads, Annotations can be downloaded in JSON or CSV formats and are separated by motivation.

Annotation Server

LibCrowds now implements its own Annotation Server, which enables the reading of annotations via a set of API endpoints. The server cannot be used to create, update or delete endpoints, which happens via other means, but it does comply fully with the Web Annotation protocol to ease interoperability between systems. For example, all annotations associated with a particular IIIF manifest, regardless of how many projects were used to generate them, can be retrieved as a collection from a single API endpoint. The documentation contains more details of the available endpoints.

Note that many of the above updates rely on additional backend code added via the pybossa-lc plugin.

See the closed milestone for more details of the issues solved.