This prototype is an attempt to present a user interface on top of existing linked data tooling to help users build CSV-Ws from their existing CSV data.
It will allow users to upload a CSV file along with (optional) accompanying metadata files, provide basic metadata about the overall dataset and then generate a form for them to complete to describe each column of their data. If the optional metadata files are provided, some of this form may be pre-populated with values drawn from the files.
The csvcubed tool will be used to generate the files comprising the CSV-W at the end. This prototype currently uses the Gov.UK Prototype Kit which facilitates rapid prototyping using the Gov.UK Design System. For details of frontend components being used see these resources.
npm install to install dependencies
npm start-api or yarn start-api to run the flask backend
npm start or yarn start to run the site, which should report the website to be available on http://localhost:3000
Changes will likely be needed primarily in the app/ directory, where app/views/ provide the HTML for each page, and app/routes.js the currently used logic behind each page submission.
As this is a prototype, we don't expect it to be the final home of some of this functionality and some problems like file management are being solved in other places. This work should focus on proving the user experience of generating forms specific to the provided CSV and submitting metadata programatically to csvcubed.
Only most basic behaviour in place, but running in a local server manner:
- User can 'upload' a CSV file, which is stored on the local server
- User can provide metadata fields for dataset
- User is presented metadata fields for each column
- This may initially take the shape of one very long form, the UX can be iterated separately
- User can submit metadata form, and csvcubed is called
- Results of csvcubed are stored on the local server
- Links to the resultant files are visible to the user via the interface
The optional qube-config.json file can be provided and the values used to inform the journey:
- User can 'upload' the qube-config.json file
- JSON file can be interpretted, and impact the form shown
Results should be submitted and presented the same
Once these behaviours are in place, the prototype can be used to demonstrate the ability of csvcubed to those who would not be able to install the tool directly and also for user testing to ensure the journey is straightforward for users.
This CSV-W Builder may then evolve into a part of the dd-cms codebase or may find a home as an independent tool. Depending on which direction we take at that point, the file upload components and error handling may need further work or may be impacted by the codebase it moves to.