Refactor executable IDs to use strings #262

This commit unifies executable ID structure across categories and
scripts, paving the way for more complex ID solutions for #262.
It also refactors related code to adapt to the changes.

Key changes:

- Change numeric IDs to string IDs for categories
- Use named types for string IDs to improve code clarity
- Add unit tests to verify ID uniqueness

Other supporting changes:

- Separate concerns in entities for data access and executables by using
  separate abstractions (`Identifiable` and `RepositoryEntity`)
- Simplify usage and construction of entities.
- Remove `BaseEntity` for simplicity.
- Move creation of categories/scripts to domain layer
- Refactor CategoryCollection for better validation logic isolation
- Rename some categories to keep the names (used as pseudo-IDs) unique
  on Windows.

tests/unit/shared/Stubs/CategorySelectionStub.ts

@@ -1,11 +1,12 @@
 import type { CategorySelection } from '@/application/Context/State/Selection/Category/CategorySelection';
 import type { CategorySelectionChangeCommand } from '@/application/Context/State/Selection/Category/CategorySelectionChange';
+import type { ExecutableId } from '@/domain/Executables/Identifiable';
 import { StubWithObservableMethodCalls } from './StubWithObservableMethodCalls';
 
 export class CategorySelectionStub
   extends StubWithObservableMethodCalls<CategorySelection>
   implements CategorySelection {
-  public isCategorySelected(categoryId: number, revert: boolean): boolean {
+  public isCategorySelected(categoryId: ExecutableId, revert: boolean): boolean {
     const call = this.callHistory.find(
       (c) => c.methodName === 'processChanges'
       && c.args[0].changes.some((change) => (