🪐 Weasel Project: Detecting techtools from the Python Bytes podcast.

This project contain a demo of a spaCy pipeline to detect programming tools and Python packages from podcast transcripts.

📋 project.yml

The project.yml defines the data assets required by the project, as well as the available commands and workflows. For details, see the Weasel documentation.

⏯ Commands

The following commands are defined by the project. They can be executed using weasel run [name]. Commands are only re-run if their inputs have changed.

Command	Description
index	Index the raw data so that a Prodigy recipe can easily search through it.
annot-search	Use a searching trick to annotate
annot-model	Use a model trick to annotate
annot-export	Get annotations out of Prodigy
convert	Convert annotations to spaCy format.
config	Create a configuration file for training.
train	Train a named entity recognition model
package	Package the trained model so it can be installed

⏭ Workflows

The following workflows are defined by the project. They can be executed using weasel run [name] and will run the specified commands in order. Commands are only re-run if their inputs have changed.

Workflow	Steps
all	convert → config → train → package