Improve sample() method

[npatki]

Problem Description

Let's make sampling more user friendly. We can create multiple methods for different user needs.

The sample() method should only be used to sample entire rows (no conditions).

Expected behavior

Parameters:

• num_rows is a required param
• randomize_samples will determine whether or not there should be a fixed seed (default: True)
• Remove all existing params

# works with any tabular model
from sdv.tabular import CTGAN
model = CTGAN()
model.fit(data)

# creates 1000 rows. Different rows created every time
model.sample(num_rows=1000)

# Fix a seed so that this call always creates rows in the same order
model.sample(num_rows=1000, randomize_samples=False)

Error Handling

Guidance if the user does not specify num_rows.

>>> model.sample()
Error: You must specify the number of rows to sample (eg. num_rows=100).

Guidance if the user passes in the conditions argument.

>>> model.sample(num_rows=100, conditions={'age': 45})
Error: This method does not support the conditions parameter. Please create Condition objects
and pass them into the sample_conditions method. See User Guide or API for more details.