Updating how database object information is getting retrieved (#79)

* Updating how database object information is getting retrieved

I updated the SQLite query for gathering database objects to make use of the JSON1 extension so you can grab tables and views name, their associated columns with types and set it to the schema. This removes the need to work with DDL's. Hints for Tables and Views works since my approach is they are both database objects.

src/lib/database/index.js

@@ -1,4 +1,3 @@
-import stms from './_statements'
 import fu from '@/lib/utils/fileIo'
 // We can import workers like so because of worker-loader:
 // https://webpack.js.org/loaders/worker-loader/
@@ -81,24 +80,14 @@ class Database {
 
   async refreshSchema () {
     const getSchemaSql = `
-      SELECT name, sql
-      FROM sqlite_master
-      WHERE type='table' AND name NOT LIKE 'sqlite_%';
+    WITH columns as (SELECT a.tbl_name, json_group_array(json_object('name', b.name,'type', IIF(b.type = '', 'N/A', b.type))) as column_json
+    FROM sqlite_master a, pragma_table_info(a.name) b
+    WHERE a.type in ('table','view') AND a.name NOT LIKE 'sqlite_%' group by tbl_name
+    )
+    SELECT json_group_array(json_object('name',tbl_name, 'columns', json(column_json))) objects from columns;
     `
     const result = await this.execute(getSchemaSql)
-    // Parse DDL statements to get column names and types
-    const parsedSchema = []
-    if (result && result.values && result.values.name) {
-      result.values.name.forEach((table, index) => {
-        parsedSchema.push({
-          name: table,
-          columns: stms.getColumns(result.values.sql[index])
-        })
-      })
-    }
-
-    // Refresh schema
-    this.schema = parsedSchema
+    this.schema = JSON.parse(result.values.objects[0])
   }
 
   async execute (commands) {