Quote:
Originally Posted by DaltonST
@UserEntityTypes are very similar to @UserCategories, but just not in the Tag Browser. UETs are psuedo-columns, just like UCs. Look in table preferences key field_metadata, and you will see that UCs are next to Custom Columns that are next to Standard Columns. In theory (but not in practice since the book-level values would be missing), I could add UETs to the Tag Browser where all of the Standard/Custom Column ETs already are. That is why I like the "@".
|
I can see how UETs and UCs are similar, but I see no advantage in using the same distinguishing leading character. The main significance of the '@' on a UC is its usage in search terms e.g.
@Favourite Authors:true. Its usage on UETs might lead a user to think they are similarly searchable.
Quote:
Originally Posted by DaltonST
I am not a Template person, do not ever use them, and do not know if they are limited to the GUI Cache, or can read metadata.db directly via APSW. A Template function in general program mode that could read a SQLite database directly would work. EM is outside of the GUI Cache entirely, since book ID is not anywhere within EM. That is its point; it is designed to be "above" book-level. So, the general program mode function would be reading from the physical database instead of the in-memory Cache for every book and every potential ET column at startup. Slow for large Libraries with many ET columns. If Templates can do it, I will write the general program mode function and make it available to copy-and-paste. However, if they cannot access APSW, then I know of no way to do what you ask. I am interested to know if they can.
|
My
rough thinking is running along these lines:
- In a library where I wanted to use the Entities Manager I would have to define a custom column to store a list of columns and values that have a corresponding entity: e.g.#entity_cells/Entity Cells, viz:
I chose that particular definition because I would NOT want it in the tag Browser. A individual book would only have one of these with an id that matched the book table item id. If each lookup_name:value pair had a separate row in the table with its own unique id, then it would probably make maintenance easier, but may make the icon template more difficult - so I'm assuming one per book.
- When an entity is created, the column lookup name and value (e.g. 'publisher:Random House', '#illustrator||Clement Hurd') would be added to the book's #entity_cells list. Maintaining the list for changes and deletes would be bit tricky if its stored a single row with multiple items.
- Then I would need to write an icon template for each column likely to have entities, that interrogated the #entity_cells column for an entry that had lookup_name and lookup_name value - i.e contains "publisher" and contains {publisher). Not sure if it would be possible to write a single template to use on any entity-able column.
The Entity Cells column could be displayed in Book details panels. Be nice if each item appeared a separate line in a list, should be feasible with a single long text column. Should the list be sorted one asks - someone's bound to ask 
I haven't tried to map this scheme to the Entity Manager tables - my guess is that's there is considerable overlap.
BR