We want to avoid saving files to disk if we use the file immediately. This is especially interesting for ml pipelines.

We want to keep our files in memory and not on disk. We do this by using the BytesIO module.

from io import BytesIO

def generate_file_in_memory():  
    # Generate file content  
    file_content = "This is a sample text file content."  
  
    # Save to in-memory buffer  
    file_buffer = BytesIO()
    # The encode() part is not always necessary, 
    # savefig from matplotlib does it  automatically for example.
    file_buffer.write(file_content.encode()) 

    return file_buffer

file_buffer = generate_file_in_memory()
print(file_buffer.getvalue())