Use generic schema definitions

[rkistner]

This changes the schema provided over the API to primarily provide a sqlite type, rather than a postgres type. It's more than just renaming the field - the conversion to the SQLite type now happens in the service (in the specific module), rather than the client.

Each column now primarily has:

1. sqlite_type: The type as it will behave in sync rules.
2. internal_type: The original type from the source database. This should not be used to check the type in code, but can be used for descriptions in auto-complete or generated schemas.

The old pg_type and type fields are still kept as is, to preserve backwards-compatibility. However, these do not need to be populated by other modules, and can be removed later.

This now also refactors internals to keep the original type (internal_type) available in the output schema after going through sync rules, allowing us to annotate the generated schema with comments like this:

const assets1 = new Table(
  {
    // id column (text) is automatically included
    name: column.text, // text
    count: column.integer, // int4
    owner_id: column.text // uuid
  },
  { indexes: {} }
);

[changeset-bot]

🦋 Changeset detected

Latest commit: 4ecaee2

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 6 packages

Name	Type
@powersync/service-sync-rules	Minor
@powersync/service-jpgwire	Patch
@powersync/service-core	Patch
@powersync/service-module-postgres	Patch
@powersync/service-image	Patch
test-client	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR