Simple pipeline | Basic Feature Engineering

[Paulitos]

Hi, I'm trying to do this code in R (scikit-learn code provided):

from sklearn.pipeline import Pipeline
from sklearn.impute import SimpleImputer
from sklearn.preprocessing import StandardScaler
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OrdinalEncoder
from sklearn.preprocessing import OneHotEncoder

num_pipeline = Pipeline([
    ('imputer', SimpleImputer(strategy='median')),
    ('std_scaler', StandardScaler())
])

cat_pipeline = Pipeline([
    ('oh_encoder', OneHotEncoder())
])

full_pipeline = ColumnTransformer([
    ('num', num_pipeline, num_attribs),
    ('cat', cat_pipeline, cat_attribs)
])

# non-selected columns such as 'PassengerId' and 'Cabin' will be
# dropped after passing through the column transformer
X_train_copy_transformed = df_transform(X_train_copy)
print(X_train_copy_transformed.columns)
X_train_copy_prepared = full_pipeline.fit_transform(X_train_copy_transformed)
print(X_train_copy_prepared[0])

What I have so far is this:

library(mlr3)
library(mlr3pipelines)

num_pipeline <- po("imputemedian") %>>% po("scale", center = TRUE, scale = TRUE)
cat_pipeline <- po("encode", method = "one-hot")

full_pipeline <- po("pipe", list(
  po("select", num_attribs) %>>% num_pipeline,
  po("select", cat_attribs) %>>% cat_pipeline
))

X_train_copy_transformed <- df_transform(X_train_copy)
print(colnames(X_train_copy_transformed))
X_train_copy_prepared <- full_pipeline$train(list(data = list(x = X_train_copy_transformed)))$predict(list(x = X_train_copy_transformed))
print(X_train_copy_prepared[1, ])

The error has to do with Error in .__PipeOp__initialize(self = self, private = private, super = super, : Assertion on 'id' failed: Must have length 1. but I don't really understand the select function, I think is because I still have problems with oop. Anyways, if you want to try the notebook I'm working with here the notebook for you to try with the data and deal with the errors. Any help would be appreciated.

https://www.kaggle.com/code/paulitos/spaceship-titanic-in-r-paulos