CLabel is a terminal-based cluster labeling tool that allows you to explore text data interactively and label clusters based on reviewing that data.
Install & Quickstart
pip install clabel
clabel to run. Everything should happen in the terminal from there.
clabel can only import CSV files. It expects two columns to be in your csv: a column of text (
string) and a column of cluster labels (
int). You’ll identify these the first time you import a dataset.
The workflow is:
- Pick a cluster to view examples. You’ll view this through a pager so you can page through examples.
- Come up with a name for that cluster (
- Repeat 1 & 2 until all your clusters have names.
You can persist any cluster labels to a
json file when you exit, so you don’t have to complete labeling in one session. Then, you can load those labels in the next time you start
clabel by selecting that
json file and continue labeling.