Apteryx

Kiwi are flightless birds endemic to New Zealand of the genus Apteryx — Wikipedia

The backend server for apteryx, a grocery store inventory management tool. Implemented with Express, Apollo GraphQL, and Postgres.

Features

Apteryx currently features a:

• Product catalog with fields for price information and distributor data
• List of expiring products with dates and pull information

Planned features (in no particular order):

• Topstock counting for inventory management — open an aisle, work the topstock, and scan in the remaining
• Product scan and search for finding products
• Sale price programmer to automatically time price changes in the database
• Order system, generating and entering order information

Installing

Included is a docker compose file, which should run everything — server, migrations, and database. Add a .env file with the args outlined within docker-compose.yml, and run.

Managing GraphQL Queries

The root index.ts file in /src/graphql manages configuration of schema, resolvers, and data sources for Apollo.

Schema

Schema definitions are namespaced to their purpose. Each will have definition, extending the base query and mutation types.

Schema mutations are abstracted further, nesting the resolvers into the namespaced entry for better organization of repeated query names. This means queries normally shaped like this:

# Incorrect query usage for the defined schema.
mutation Mutation($product: ProductInput!) {
    createProduct(product: $product) {
        id
    }
}

... turn into a query with the function nested one level further, like this:

# Correct.
mutation Mutation($product: ProductInput!) {
    # Namespaced "product"
    product {
        # Create function is nested one level in.
        create(product: $product) {
            id
        }
    }
}

Resolvers

Resolvers pass logic to the respective data source provider. When modifying mutations, keep in mind the extra query layer, as demonstrated above. As of this, they're in their namespaced folders, with files for query and mutation objects.

Both resolvers are exported and added to the resolvers field in the graphql index file.

Exported resolvers, again because of the layered mutation schema, export an object. This is to pass the query to a default resolver, which handles the request. You can find an example of this in the product mutation resolver.

Data Sources

Data sources reference provider classes, stored in /src/graphql/providers. These just interact with the services layer (/src/services) to perform database transactions based on GraphQL types.

When adding new data sources, update the interface in the graphql index file to include the new types.