Create a GraphQL Resolver with Data Loading and Authorization

Write a complete GraphQL resolver for a [ENTITY_NAME] type using [GRAPHQL_FRAMEWORK] (e.g., Apollo Server, GraphQL Yoga, Strawberry, Ariadne) in [PROGRAMMING_LANGUAGE]. Schema context: - Parent type: [PARENT_TYPE] - The [ENTITY_NAME] has these fields: [FIELD_LIST_WITH_TYPES] - Relationships: [ENTITY_NAME] has [RELATIONSHIP_DESCRIPTION] (e.g., "a User has many Posts and belongs to an Organization") Please implement the following resolvers and patterns: 1. **Type Definition**: Write the GraphQL SDL (Schema Definition Language) for [ENTITY_NAME] including input types for mutations. 2. **Query Resolvers**: - `[ENTITY_NAME_LOWERCASE](id: ID!)` — fetch single entity - `[ENTITY_NAME_LOWERCASE_PLURAL](filter: [FILTER_INPUT], pagination: PaginationInput)` — fetch list with cursor-based pagination 3. **Mutation Resolvers**: - `create[ENTITY_NAME]` with input validation - `update[ENTITY_NAME]` with partial update support - `delete[ENTITY_NAME]` with soft delete option 4. **Field Resolvers**: Implement resolvers for relationship fields ([RELATIONSHIP_DESCRIPTION]) using DataLoader pattern to solve the N+1 query problem. Show the DataLoader factory function and batch loading logic. 5. **Authorization**: Add middleware/directive-based authorization checking [ROLE_OR_PERMISSION] before resolver execution. Handle unauthorized access gracefully. 6. **Error Handling**: Use custom GraphQL errors with extensions for error codes: NOT_FOUND, UNAUTHORIZED, VALIDATION_ERROR, and INTERNAL_ERROR. 7. **Context Setup**: Show how the resolver accesses database layer, authenticated user, and DataLoader instances from context. Include TypeScript/type annotations where applicable and add comments explaining the DataLoader batching strategy.