Tutorial on godot engine resources

The normal flow is you only have one remap pointer like res://model.whisper and one data resource in res://.godot/imported/model-ADSADSAGFASBASDSADASDA.res.

Using godot-whisper design. This is also a problem in iree.gd and this sqlite pull request.

Problem 1

In your design, you have no pointers, and the data resource is the string of the path. So you must go to the export dialogue and manually whitelist the whisper model.

Problem 2

I can only ZSTD compress the resource data, a string like "res://model.whisper".

I want Godot Engine to compress the model with ZSTD compression as a .res resource like a packed byte array.

Problem 3

I want to transfer the model over the network. Syncing the data property with a resource is enough to do this.

It MAY be called data or some other property name.

It MUST have a property that stores a transformed database data structure.

https://datatracker.ietf.org/doc/html/rfc2119#section-1

Example workflow 1:

1. copy the database to make it readable and writable to user://
2. load database at user://some_database.sqlite

Example workflow 2:

1. The inspector window can be edited during the development stage
2. On publish, the database is locked unless workflow 1

I don't have enough design sense to review adding an ORM.

void create_table(const String &p_table_name, const TypedArray<SQLiteColumnSchema> &p_columns);
void drop_table(const String &p_table_name);

void insert_row(const String &p_name, const Dictionary &p_row_dict);
void insert_rows(const String &p_name, const TypedArray<Dictionary> &p_row_array);

Array select_rows(const String &p_name, const String &p_conditions, const Array &p_columns_array);
bool update_rows(const String &p_name, const String &p_conditions, const Dictionary &p_updated_row_dict);
bool delete_rows(const String &p_name, const String &p_conditions);
TypedArray<String> get_table_names() const;
TypedArray<SQLiteColumnSchema> get_columns(const String &p_name) const;
Variant execute(const String &p_query_string);

For this pull request, please keep creating a resource database and the create_query code path for that resource because I can't figure out the shape of the mapper.

I would postpone the object relation mapping to a subsequent pull request.

https://martinfowler.com/bliki/OrmHate.html suggests either spending a lot of time writing a better objection relation mapper or avoiding the problem.

Edited:

According to the title, a SQLite resource is what is requested, not an Object Relation Mapper.

What parts did Piet Bronders & Jeroen De Geeter do?

Can you separate the object relation mapping system into a different pr?

I have to focus on other essential tasks, and I want to merge the Resource saver, resource loader, and sqlite node.

Also, I'm curious about who those third-party people are.

What parts did Piet Bronders & Jeroen De Geeter do?

Well, no parts he did, but the High Level api is similar to his one (and I took inspiration from his repo)
https://github.com/2shady4u/godot-sqlite

Can you separate the object relation mapping system into a different pr?

I have to focus on other essential tasks, and I want to merge the Resource saver, resource loader, and sqlite node.

Also, I'm curious about who those third-party people are.

I could try later