Add repr for Morpheme/MorphemeList

[polm]

This makes it easier to check values when developing interactively. Probably should have been included with #124.

[kazuma-t]

For object.__repr__(), the Python language reference has the following description.

If at all possible, this should look like a valid Python expression that could be used to recreate an object with the same value (given an appropriate environment).

(from 3.3.1. Basic customization)

This implementation does not seem to follow that purpose. Even if it can't, it is undesirable that it is the same as __str__(), I think.

[polm]

The current output looks like this:

<sudachipy.morphemelist.MorphemeList at 0x7f9afedf7940>
<sudachipy.morpheme.Morpheme at 0x7f9afdc2c850>

That doesn't follow the purpose in the language reference and also isn't useful for anything as far as I can tell.

[kazuma-t]

At least, begin, end, dictionary_id, word_id, and is_oov are needed to identify morpheme .

[eiennohito]

It is impossible to instantiate Morpheme/MorphemeList directly, so I think we need to decide whether their __repr__ should be more useful to developers of SudachiPy or users of SudachiPy.

[kazuma-t]

How about a format like this,

<Morpheme (猫, 0:3, 0, 571365)>
<MorphemeList [(猫, 0:3, 0, 571365), (が, 3:6, 0, 45393), (ぴらる, 6:15, -1, -1)]>

( {surface}, {begin}:{end}, {dict_id}, {word_id}) (dict_id and word_id are -1 in OOV)

[eiennohito]

As the reasoning for the format, being able to detect whether the word comes from user dictionaries/system dictionaries or OOV can help to debug and resolve problems with user dictionaries.