string_type_benchmark_plotting.py

import hvplot
import polars as pl


def plot_combined_csv(
    filter_dataset_file: str, dataset_size: str, title: str, normalized: bool = False
):
    df = pl.read_csv(f"output/filter/filter_{dataset_size}_*.csv").with_columns(
        (pl.col("Mask") * 100).cast(int)
    )

    plot = df.plot.bar(
        x="Mask",
        y="Time",
        by=["Polars_Version"],
        color=["#73bfb8", "#0075FF"],
        xlabel="Selectivity [%]",
        ylabel="Time [s]",
        title=f"{title} by Polars version",
    )

    plot = plot.redim.values(Polars_Version=["0.20.6", "0.20.5"])

    hvplot.save(plot, filename=f"{filter_dataset_file}.html")


plot_combined_csv(
    "filter_small16", "small16", "Filter on 16x1M rows with <= 12 byte strings"
)
plot_combined_csv(
    "filter_medium16", "medium16", "Filter on 16x1M rows between 1 and 200 byte strings"
)
plot_combined_csv(
    "filter_large16", "large16", "Filter on 16x1M rows with 500 byte strings"
)