Fix: Encode str to bytes before writing to buffered IO

[mesquita]

If you try something like this:

html_buffer = BytesIO()
report.save_html(html_buffer)

where report is

report = Report(
    metrics=[
        DataDriftPreset(),
    ]
)

after a successful report.run

You'll get:

TypeError: a bytes-like object is required, not 'str'

due to the fact that render in filename.write(render) is a string, not bytes.
In this PR, I'm encoding render to bytes before writing to the buffer.

[mike0sv]

Can't you just use StringIO instead of BytesIO?

[mesquita]

Can't you just use StringIO instead of BytesIO?

Hi @mike0sv, since save_html accepts str or IO, it should work with BytesIO as well. In my case, it'd be very useful.

I can suggest something like this:

          if isinstance(filename, str):
              with open(filename, "w", encoding="utf-8") as out_file:
                  out_file.write(render)
          elif isinstance(filename, io.BytesIO):
              filename.write(render.encode(encoding="utf-8"))
          elif isinstance(filename, io.StringIO):
              filename.write(render)
          else:
              raise ValueError("filename should be a string or a file-like object")

what do you think?

[mike0sv]

This will not work because of duck typing. Those instance checks will return False for objects created with open(...) for example. Same for typing.BinaryIO and typing.TextIO.
If you need to get html payload, you can use get_html instead of save_html.
As for type annotation, probably we should correct annotation to typing.TextIO or use more flexible check. I don't know how to do this from the top of my head though