QRI uses an internal datastore, so interacting with your datasets requires using Qri Desktop or using the command line interface (CLI).
To make interacting with the components of a dataset easier, Qri has added working directories, which map the components of a Qri dataset to files in a designated directory.
For example, the
body component becomes
|component||working directory filenames|
The working directory acts as a staging area for new versions of a dataset. Furthermore, anything that can write to a file can be used to make changes to the dataset!
Once a dataset has been checked out, any third-party programs can be used to interact with the various files:
- A text editor could be used to open
body.csvto find and replace values
- A python or node script could be used programmatically overwrite
meta.jsonwith new metadata.
- A text editor could be used to edit markdown for
Qri uses the term checkout to describe the creation of a working directory for a dataset. A dataset checkout can be accomplished in Qri Desktop using the checkout button in the Dataset pane, or in Qri CLI using
If a dataset is linked to a working directory, the dataset is "checked out".
Qri watches working directories, and is aware of when the files no longer match the most recent version of the dataset. In Qri Desktop, the changes are displayed using colored indicators in the Dataset pane's status tab.
qri status on the command line will report the status of changes to the files in the working directory.
A typical workflow for using working directories to manage changes to a dataset looks like this:
- Create or Add a Qri Dataset
- Checkout the dataset (create a working directory)
- Use third party tools to modify the files (Qri will track the changes)
- Commit the changes to create a new version!
- Repeat steps 3 and 4 as needed