Ablation.drop_steps()

Add a variant that removes the rows failing any of the given steps.

Usage

Source

Ablation.drop_steps(
    steps,
    name=None,
)

This is the most flexible strategy: it adds one variant, with the rows that failed any of steps= removed. Use it to test a particular combination of checks (e.g., all the checks on one column), or to give a variant a name of its own.

Parameters

steps: int | list[int]

One or more step numbers. Each must be a row-level step that the ablation can use (active, and without a pre= function unless that function is marked with row_preserving()).

name: str | None = None

A name for the variant, shown in reports and used as its id. By default, it’s named after the steps (e.g., "Steps 1, 3", with the id "steps_1_3").

Returns

Ablation

The Ablation object, with the variant added.

Examples

Here’s a table of temperature readings with some out-of-range values, checked against two different ranges. The ‘model’ estimates the average temperature and is evaluated by its error against the true average of 20.

import pointblank as pb
import polars as pl

temps = [20 + (k * 37 % 11 - 5) * 0.6 for k in range(300)]
for k in range(0, 300, 25):
    temps[k] *= 10
readings = pl.DataFrame({"temp": temps})

def evaluate(train, *, seed, variant):
    sample = train.sample(fraction=1.0, with_replacement=True, seed=seed)
    return abs(sample["temp"].mean() - 20)

validation = (
    pb.Validate(data=readings)
    .col_vals_between(columns="temp", left=-10, right=45)
    .col_vals_lt(columns="temp", value=100)
    .interrogate()
)

ablation = pb.Ablation(
    validation, evaluate=evaluate, metric="error", higher_is_better=False
).drop_steps([1, 2], name="Range checks")

ablation.plan()

Ablation plan
300 rows in the training table. Each version is evaluated 5 times, so the run will make 15 evaluations.
Version	Removed
Baseline
Full table Nothing removed	0 0%
Variants
Range checks 1: Expect that values in temp should be between -10 and 45. 2: Expect that values in temp should be < 100.	12 4.0%
Controls
12 random rows Rows removed at random, drawn anew for each repeat	12 4.0%

The plan has the full table, the one variant (removing the rows that failed either step), and its control, which removes the same number of rows at random.