libcalibre: rewrite with simplified DDD

[phildenhoff]

I'd like to use a subset of DDD's language in libcalibre that should make it easier to extend & maintain. With a better understanding of how libcalibre works (now that I've written (part of) it) comes the ability to write a better version.

• repos: how to write to the database. Entities that also have a table get a repo. Other operations are in the aggregates repo? I guess? Includes ways to store Book Files.
• dtos: requests to create new / update existing entities and aggregates.
• services: specifically for Book, Author, LibraryFile (any file within the lib. root)

Stretch goals

• Remove Arc (Rc is a better choice, but not ref. counting is even better)
• Remove Mutex
• Make setup easier in citadel-rs – creating (and destroying) 3-4 repos & 3-4 services is wild and mostly annoying. It 100% doesn't not need to do that, it can store the services.
• Remove mut self's for services. It seems unnecessary
• Minimize .clone() calls
• Prefer &str to String
• Create struct/types for BookId & AuthorId with parse_from and to_i32 impls (book_${id} => id)
• High perf versions of read functions, e.g. only return IDs or support windows/pagination
• Filtering (search(phrase, fields: &str[]) or similar pattern)

Book

Book aggregate (? or "entity"?) allows for CRUD-ing ("managing")

• title
• authors associated to a Book (update_links which diffs, adds & removes)
• which series this book belongs to (add if missing; replace otherwise)
• series position
• external identifiers
• tags
• publication date
• language
• publisher
• HTML description
• book cover image (replace by file @ path), and
• different format files (e.g. .epub, .kepub, and .mobi).

as well as

find_by_id/1
find_by_author/1
list_all/0
create/1, which accepts a minimal DTO

Author

Author entity supports CRUD operations for

• name
• link
• ...associated book? that seems like a mistake