How to log custom metadata in a table format

A Vectice Table wraps a pandas dataframe into a table format, which you can then log to a Vectice iteration. This enables you to log custom metadata to Vectice in a tabular format.

You should not use Table to store important or sensitive information. Table is for storing low sensitive information that may support your documentation efforts.

For example, if you wish to log a few rows of additional low sensitive information not captured in the dataset, you can utilize Table to log and document them into Vectice. See below for a code example:

import pandas as pd
from vectice import Table

table_dict = {
    "inputs": ["What is Vectice?", "Describe a dog to me"],
    "outputs": ["Vectice is an auto-documentation platform", "A dog is an animal with four legs"],
    "toxicity": [0.0, 0.0],
}

prompt_data_input_output = pd.DataFrame(table_dict)

table = Table(prompt_data_input_output, name="prompt_data")
iteration.log(table)

Table()has a maximum capacity of 100 rows and 20 columns.