fix: Handle non utf-8 characters in OME-XML

[weiji14]

By using String::from_utf8_lossy.

Changes:

• Used another sample OME-TIFF file that has non utf8 characters in the OME-XML
• Increased prefetch bytes to more than 32kb as workaround to avoid hanging when parsing the StripOffsets tag (xref Tokio hangs when opening TIFF #89)

Fixes #101

[kylebarron]

Ok so to clarify the spec says that these characters should always be UTF8? And which spec are we talking about? TIFF or OME-TIFF?

This tag type is named Value:Ascii, so it seems like a mismatch that we're hitting utf8 issues. Should we rename that to Value::Text? or Value::String?

Alternatively we could store a Vec<u8> but if this is string data, then Vec<u8> seems improper