convector 0.1.0

A tool for transforming conversational data to a unified format

Convector: Conversational Data Transformation Simplified

Convector facilitates the standardization of conversational datasets, making conversational data preparation for NLP tasks efficient and straightforward.

Installation

Install with pip:

pip install convector

Or install from source:

git clone https://github.com/teilomillet/convector && cd convector && pip install .

First Use

Running Convector initiates config.yaml in your directory, storing your transformation profiles.

Using Convector

Transform data files with:

convector process <file_path> [OPTIONS]

Use --help for more information and deep dive in the possibilities.

Profiles

Use -p with a profile name to apply specific transformations. New or modified profiles auto-save to config.yaml.

Default Schema

Out-of-the-box schema structure:

{"instruction":"", "input":"", "output":""}

Examples

Process nested data, assuming conversational format, to JSON:

convector process data.csv -c -f output.json

Automatically create/update a profile with additional columns you want to keep:

convector process data.csv -p new_profile --add-cols 'col1,col2'

Capabilities

• Standardize datasets for model training.
• Convert varied data formats to a unified structure.
• Automate data preparation processes.

Convector serves as a practical tool for data format standardization, offering a suite of options for custom data transformation needs.