split Term into Term and IndexingTerm

[PSeitz]

• split term and indexing term
• add JsonTermSerializer
• cleanup
• clippy

This splits Term into Term and IndexingTerm. Term is used for querying, IndexingTerm during indexing. This allows to reduce the complexity of Term and a more compact format for IndexingTerm by dropping the type marker at byte 5.

Part of: https://github.com/quickwit-oss/tantivy/issues/1795

[fulmicoton]

As discussed offline, this looks good but let's not merge not progress on #1795 at the moment.