can't init dbsync when table has an index on id column?

Nobody has claimed this yet.

I can't seem to start dbsync , initialised from the database when my table has a btree index on the id column.

```sql
CREATE TABLE my_polygon_table (
id INT GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
comment text,
date date,
recorded_by text,
validated boolean,
validated_by text,
geom geometry(Polygon,4326)
);

CREATE INDEX IF NOT EXISTS my_polygon_table_geom_gist
ON my_polygon_table USING gist (geom);

CREATE INDEX IF NOT EXISTS my_polygon_table_id_btree
ON my_polygon_table USING btree (id);
```

I get this error in dbsync :

```
Error: postgres cmd error: ERROR: column "id" specified more than once
|
|
| SQL:
| CREATE TABLE "z_for_mergin_use_only_db_sync"."my_polygon_table" ("id" integer NOT NULL, "id" integer NOT NULL, "comment" text, "date" date, "recorded_by" text, "validated" boolean, "validated_by" text, "geom" geometry(Polygon,4326), PRIMARY KEY ("id"));
```

However if I do the same without creating an index it works fine.

Reproduce the issue using the SQL schema in the report, with and without the btree index, then trace dbsync's database-initialization path that generates the CREATE TABLE statement. Done means initialization succeeds when the id column has the additional btree index and no duplicate id column appears in the generated SQL.

Written by the indexing model from the issue text.