@architect/inventory

@architect/inventory

Architect project resource enumeration utility

Inventory is responsible for interpreting the configuration and shape of a given Architect project, validating its settings, and representing it in a consistent intermediate format.

Install

npm i @architect/inventory

Usage

inventory(options[, callback]) → [Promise]

Runs an inventory against your project. Must be passed an options object that may containing the following parameters:

• cwd - String - Absolute file path of the project being inventoried
• env - Boolean - Queries AWS infra to acquire environment variables for testing, staging, and production environments
• region - String - Sets default AWS region; overrides default, but is overridden by AWS_REGION env var
• rawArc - String - Raw Architect project manifest string, intended for testing; providing this will ignore your local manifest

Returns results via callback, or returns a promise if callback is falsy, and resolves with results.

Inventory object

Inventory returns an object containing two parameters: inv (the project inventory object) and get (a getter helper for querying resources).

inv

The inventory object contains the entirety of a project's data, including Architect defaults, project defaults, inferred resources, userland settings layered from the project and function levels, local preferences, etc. An inventory object should be considered the source of truth about the state of your project, and should not be directly mutated.

Top-level inventory parameters that start with an underscore (e.g. _arc, _project) denote project metadata or internal diagnostic data; all other parameters represent userland project resources.

In a project inventory, null values are used as placeholders for known values or options that were not user-defined. The existence of a non-null value can be inferred as a user having specifically defined a setting. For example: arc.http === null can be construed as the user having not defined an @http pragma. This rule has some exceptions:

• A handful of settings that must be backfilled if not supplied
  • Example: inv.aws.region, which is required by the aws-sdk to function, and will be backfilled if not defined
• Pragmas that infer other pragmas
  • Example: while @static can be defined on its own without any other pragmas, the existence of @http infers @static
  • Thus, the act of adding @http will necessarily make inv.static non-null
• Settings that generate related resources
  • Example: DynamoDB streams can be defined in @tables with stream true; Inventory would interpret a table with stream true as a new inv.streams resource and thus make inv.streams non-null

Note: The inv format is primarily designed and intended for internal use within Architect libraries; as such, Inventory may theoretically introduce breaking changes in its behavior, shape, or naming conventions.

get

You do not need to use the get helper to use a project's inventory, but get does make it much easier to check for the existence of resources, or find specific resources.

The get helper works as such: get.{pragma or property}('parameter or name of resource'). (Not including a parameter or resource name will fail in most cases.)

Examples:

@app
my-app

@http
get /

@events
event

get.app()             // Returns my-app; same as accessing `inv.app`

get.http('get /')     // Returns `get /` resource data
get.http('get /foo')  // Returns undefined
get.http()            // Returns undefined

get.static('folder')  // Returns 'public' (default inferred by existence of @http); same as accessing `inv.app.static`

get.events('event')   // Returns `event` resource data

get.tables('data')    // Returns undefined