split data into train and test in python

Question

i have list of authors in this form:

list_authors = ['Abbott, Edwin Abbott, 1838-1926',
                'Gatlin, Dana, 1884-1940',
                'Riley, W. (William), 1866-1961',
                ...]

i also have list of books and their authors in this form:

list_books = [(15, ['Melville, Herman, 1819-1891'],['Barrie, J. M. (James Matthew), 1860-1937'],['Cather, Willa, 1873-1947']),(27, ['Hardy, Thomas, 1840-1928']),(32, ['Gilman, Charlotte Perkins, 1860-1935']),...]

list of tuples where each tuple is the book id and list of the book authors(can be one or more from the list_authors)

the goal is to split the books without the same author can be in test and train.

tried to use sklearn to split it like this:

train_authors, test_authors, = train_test_split(list_authors, test_size=0.20, random_state=42)

which is not solving the edge case when a book has multiple authors (say two), with one being in the train, and the other in the test split