Account for readerless formats

[hinerm]

Currently we assume that if a checker exists for a format, then a parser/reader exist as well.

This is actually incorrect as we can have formats that are output-only (e.g., https://github.com/scifio/scifio/blob/master/src/main/java/io/scif/formats/JavaFormat.java)

So we need to rethink either or both:
a) the API, providing a way to search for formats that have their own reader implementations (similar to how we look for writer-enabled formats)
b) use of the API, ensuring we only indicate support for a dataset if we can perform the desired operation (reading or writing)

[ctrueden]

I suggest expanding the FormatService API. Instead of just getFormat(String) we really need a way of asking for a Format for the purpose of parsing, reading, and/or writing separately.

Of course, right now you can call getFormatList(String) and then iterate over the list calling getReaderClass etc. and check for nulls. (We should probably also have Format#canRead(), Format#canParse() and Format#canWrite() to make that more straightforward.) But such iteration is not DRY. The FormatService having built-in API for this would be better.