MangaDB DEV: Difference between revisions

From AniDB
Jump to navigation Jump to search
Line 111: Line 111:
==Database==
==Database==
===Approach 1===
===Approach 1===
[[Image:Realshitdiagram.png]]
BUSTED for various issues in regards to how data can be reused like credits and dates
BUSTED for various issues in regards to how data can be reused like credits and dates
[[Image:Realshitdiagram.png]]


===Approach 2===
===Approach 2===
[[Image:Mangadb-dbspecs2.png]]
[[Image:Mangadb-dbspecs2.png]]
[[Category:Development]]
[[Category:Development]]

Revision as of 19:41, 3 January 2011

General

this is the place to contribute ideas on a possible future addition of manga data to AniDB.

For other areas of active development on AniDB, check: Development.

Directly related: Generic PersonCompany DEV

Vision

Data

First thing we need to be aware of what data we would like to actually store/how the real world looks like.

Content forms we care about

  • (Light) Novels
  • Manga
  • Doujin

Releaseforms

Content comes in various forms made up of different elements.

chapter only releases

used where?

  • official online distribution

content elements

  • chapters

volume only releases

used where?

  • some novels
  • doujins

content elements

  • volumes which consist of basically 1 "chapter"
  • "chapter"; basically 1 per book
  • omake (depending on wanted granularity)
  • illustrations (depending on wanted granularity)
  • other 1 page content (index etc.) (depending on wanted granularity)

mixed content

used where?

  • everything else

content elements

  • volumes which encapsule the content into datasets
  • chapters
  • omake (depending on wanted granularity)
  • illustrations (depending on wanted granularity)
  • other 1 page content (index etc.) (depending on wanted granularity)

Editions

To make things worse the same content is released in multiple editions. Multiple editions usually share the same content (usually chapters) but might add different bonus content (illustrations, omake etc.). Repackaging and rereleasing are further reasons.

Hence an edition may or may not reuse existing content. It's potentially wanted to "transfer" as much data as possible with this. (where content = credits, release dates) While i say transfer what really is wanted is not duplication, but adressing the same content from multiple angles to lessen maintenance.

Credits

  • Author (作者)
  • Publisher/Printing (出版社)
  • ...

Additional ones may contain assistance, editing, colouring, illustration.

Most credits apply on volume level and for regular manga are given in a short summary towards the end of the book. Though certain types of releases may contain content which doesn't fit that rule. Especially illustrations/guest work should receive credits on chapter level. Same goes for magazine releases

Character

  • Chapter appearance

Nothing more to store here.

Situation and Problems

  • packs aren't standardized
  • repacks and re-edits are common

Hence because of that adding hashes or automating the add via those is somewhat impossible. The only way to do this would be if we break it down to the page level which is somewhat silly. Furthermore as viewer are capable of showing the content without extracting many people don't decompress the packs. Because of that going down to that level is pointless.

Implementation

General

Manga entries

  • basically just a name and a summary/description of the entry
  • further data can't be stored at this level
  • categories/tags would be assigned on this level

Editions

  • compareable to an anime entry
  • made up of several elements (chapter, volumes, etc.)
  • picture

Element Edition

  • to make truly reuseable we need a layer between editions and elements
  • release dates
  • volume/chapter "number"
  • controls the order of elements (some internal number to shift content around)

Elements

  • compareable to an episode entry
  • variable element which may be a volume, chapter, omake, illustration, etc.
  • needs nesting capeabilities to build a 2 level hirachy
  • only volume/chapter(/omake?) ones should be addable to mylist; rest is mainly only in for credits and completeness sake

Files

  • generic files with language flag and group assignment
  • possible would be to assign multiple hashes to each of these generic files to a) stop the clutter problem and b) still allow for automatisation without completely depending on filename parsing

Credits/Character

  • credits will be set on element level; we can't add a manga or edition id in this for easy access as the data wouldn't be universally useable anymore
  • to lessen the pain of data retrieval we will use a caching table which duplicates and condenses the data

Database

Approach 1

BUSTED for various issues in regards to how data can be reused like credits and dates

Approach 2