pymer4.expressions

Custom polars expressions useable inside of any polars context and compatible with other polars selectors.

Examples

Center mpg and zscore disp by passing in their names as you would to pl.col()

from pymer4.expressions import center, zscore, scale

mtcars.with_columns(
    center('mpg').alias('mpg_centered'), 
    zscore('disp').alias('disp_zscore')
    ).head()

shape: (5, 5)

mpg	disp	cyl	mpg_centered	disp_zscore
f64	f64	i64	f64	f64
21.0	160.0	6	0.909375	-0.57062
21.0	160.0	6	0.909375	-0.57062
22.8	108.0	4	2.709375	-0.990182
21.4	258.0	6	1.309375	0.220094
18.7	360.0	8	-1.390625	1.043081

Normalize all columns by dividing by their standard deviations using a polars selector like pl.all()

mtcars.with_columns(
    # Normalize all columns
    scale(pl.all()).name.suffix('_scaled')
    ).head()

shape: (5, 6)

mpg	disp	cyl	mpg_scaled	disp_scaled	cyl_scaled
f64	f64	i64	f64	f64	f64
21.0	160.0	6	3.484351	1.290961	3.35961
21.0	160.0	6	3.484351	1.290961	3.35961
22.8	108.0	4	3.783009	0.871399	2.23974
21.4	258.0	6	3.550719	2.081674	3.35961
18.7	360.0	8	3.102731	2.904662	4.47948

Expression are extensible .over('column_name')expressions.

For example, we can z-score mpg over all rows or by sub-groups of cyl

mtcars.with_columns(
    zscore('mpg').alias('mpg_zscore'),
    zscore('mpg').over('cyl').alias('mpg_zscore_by_cyl')
    ).head()

shape: (5, 5)

mpg	disp	cyl	mpg_zscore	mpg_zscore_by_cyl
f64	f64	i64	f64	f64
21.0	160.0	6	0.150885	0.864867
21.0	160.0	6	0.150885	0.864867
22.8	108.0	4	0.449543	-0.856715
21.4	258.0	6	0.217253	1.140053
18.7	360.0	8	-0.230735	1.406224

---

pymer4.expressions.center(expr: Expr | str) → Expr

Mean center the values in the expression

Parameters:

expr (Expr | str) – The expression to center

pymer4.expressions.logit2odds(expr: Expr | str) → Expr

Convert logits to log-odds

Parameters:

expr (Expr | str) – The expression to convert

pymer4.expressions.logit2prob(expr: Expr | str) → Expr

Convert logits to probabilities

Parameters:

expr (Expr | str) – The expression to convert

pymer4.expressions.rank(expr: Expr | str, method='average', descending=False) → Expr

Rank the values in the expression

Parameters:

• expr (Expr | str) – The expression to rank

• method (str, optional) – The method to use for ranking. Defaults to “average”.

• descending (bool, optional) – Whether to rank in descending order. Defaults to False.

pymer4.expressions.scale(expr: Expr | str) → Expr

Scale the values in the expression by their standard deviation

Parameters:

expr (Expr | str) – The expression to scale

pymer4.expressions.zscore(expr: Expr | str) → Expr

Z-score the values in the expression

Parameters:

expr (Expr | str) – The expression to z-score