projecte_ionic/node_modules/@angular-devkit/architect/src/api.d.ts

268 lines
12 KiB
TypeScript
Executable File

/**
* @license
* Copyright Google LLC All Rights Reserved.
*
* Use of this source code is governed by an MIT-style license that can be
* found in the LICENSE file at https://angular.io/license
*/
import { analytics, experimental, json, logging } from '@angular-devkit/core';
import { Observable, SubscribableOrPromise } from 'rxjs';
import { Schema as RealBuilderInput, Target as RealTarget } from './input-schema';
import { Schema as RealBuilderOutput } from './output-schema';
import { State as BuilderProgressState, Schema as RealBuilderProgress } from './progress-schema';
export declare type Target = json.JsonObject & RealTarget;
export { BuilderProgressState };
export declare type BuilderRegistry = experimental.jobs.Registry<json.JsonObject, BuilderInput, BuilderOutput>;
/**
* An API typed BuilderProgress. The interface generated from the schema is too permissive,
* so this API is the one we show in our API. Please note that not all fields are in there; this
* is in addition to fields in the schema.
*/
export declare type TypedBuilderProgress = {
state: BuilderProgressState.Stopped;
} | {
state: BuilderProgressState.Error;
error: json.JsonValue;
} | {
state: BuilderProgressState.Waiting;
status?: string;
} | {
state: BuilderProgressState.Running;
status?: string;
current: number;
total?: number;
};
/**
* Declaration of those types as JsonObject compatible. JsonObject is not compatible with
* optional members, so those wouldn't be directly assignable to our internal Json typings.
* Forcing the type to be both a JsonObject and the type from the Schema tells Typescript they
* are compatible (which they are).
* These types should be used everywhere.
*/
export declare type BuilderInput = json.JsonObject & RealBuilderInput;
export declare type BuilderOutput = json.JsonObject & RealBuilderOutput;
export declare type BuilderProgress = json.JsonObject & RealBuilderProgress & TypedBuilderProgress;
/**
* A progress report is what the tooling will receive. It contains the builder info and the target.
* Although these are serializable, they are only exposed through the tooling interface, not the
* builder interface. The watch dog sends BuilderProgress and the Builder has a set of functions
* to manage the state.
*/
export declare type BuilderProgressReport = BuilderProgress & {
target?: Target;
builder: BuilderInfo;
};
/**
* A Run, which is what is returned by scheduleBuilder or scheduleTarget functions. This should
* be reconstructed across memory boundaries (it's not serializable but all internal information
* are).
*/
export interface BuilderRun {
/**
* Unique amongst runs. This is the same ID as the context generated for the run. It can be
* used to identify multiple unique runs. There is no guarantee that a run is a single output;
* a builder can rebuild on its own and will generate multiple outputs.
*/
id: number;
/**
* The builder information.
*/
info: BuilderInfo;
/**
* The next output from a builder. This is recommended when scheduling a builder and only being
* interested in the result of that single run, not of a watch-mode builder.
*/
result: Promise<BuilderOutput>;
/**
* The output(s) from the builder. A builder can have multiple outputs.
* This always replay the last output when subscribed.
*/
output: Observable<BuilderOutput>;
/**
* The progress report. A progress also contains an ID, which can be different than this run's
* ID (if the builder calls scheduleBuilder or scheduleTarget).
* This will always replay the last progress on new subscriptions.
*/
progress: Observable<BuilderProgressReport>;
/**
* Stop the builder from running. Returns a promise that resolves when the builder is stopped.
* Some builders might not handle stopping properly and should have a timeout here.
*/
stop(): Promise<void>;
}
/**
* Additional optional scheduling options.
*/
export interface ScheduleOptions {
/**
* Logger to pass to the builder. Note that messages will stop being forwarded, and if you want
* to log a builder scheduled from your builder you should forward log events yourself.
*/
logger?: logging.Logger;
/**
* Target to pass to the builder.
*/
target?: Target;
}
/**
* The context received as a second argument in your builder.
*/
export interface BuilderContext {
/**
* Unique amongst contexts. Contexts instances are not guaranteed to be the same (but it could
* be the same context), and all the fields in a context could be the same, yet the builder's
* context could be different. This is the same ID as the corresponding run.
*/
id: number;
/**
* The builder info that called your function. Since the builder info is from the builder.json
* (or the host), it could contain information that is different than expected.
*/
builder: BuilderInfo;
/**
* A logger that appends messages to a log. This could be a separate interface or completely
* ignored. `console.log` could also be completely ignored.
*/
logger: logging.LoggerApi;
/**
* The absolute workspace root of this run. This is a system path and will not be normalized;
* ie. on Windows it will starts with `C:\\` (or whatever drive).
*/
workspaceRoot: string;
/**
* The current directory the user is in. This could be outside the workspace root. This is a
* system path and will not be normalized; ie. on Windows it will starts with `C:\\` (or
* whatever drive).
*/
currentDirectory: string;
/**
* The target that was used to run this builder.
* Target is optional if a builder was ran using `scheduleBuilder()`.
*/
target?: Target;
/**
* Schedule a target in the same workspace. This can be the same target that is being executed
* right now, but targets of the same name are serialized.
* Running the same target and waiting for it to end will result in a deadlocking scenario.
* Targets are considered the same if the project, the target AND the configuration are the same.
* @param target The target to schedule.
* @param overrides A set of options to override the workspace set of options.
* @param scheduleOptions Additional optional scheduling options.
* @return A promise of a run. It will resolve when all the members of the run are available.
*/
scheduleTarget(target: Target, overrides?: json.JsonObject, scheduleOptions?: ScheduleOptions): Promise<BuilderRun>;
/**
* Schedule a builder by its name. This can be the same builder that is being executed.
* @param builderName The name of the builder, ie. its `packageName:builderName` tuple.
* @param options All options to use for the builder (by default empty object). There is no
* additional options added, e.g. from the workspace.
* @param scheduleOptions Additional optional scheduling options.
* @return A promise of a run. It will resolve when all the members of the run are available.
*/
scheduleBuilder(builderName: string, options?: json.JsonObject, scheduleOptions?: ScheduleOptions): Promise<BuilderRun>;
/**
* Resolve and return options for a specified target. If the target isn't defined in the
* workspace this will reject the promise. This object will be read directly from the workspace
* but not validated against the builder of the target.
* @param target The target to resolve the options of.
* @return A non-validated object resolved from the workspace.
*/
getTargetOptions(target: Target): Promise<json.JsonObject>;
getProjectMetadata(projectName: string): Promise<json.JsonObject>;
getProjectMetadata(target: Target): Promise<json.JsonObject>;
/**
* Resolves and return a builder name. The exact format of the name is up to the host,
* so it should not be parsed to gather information (it's free form). This string can be
* used to validate options or schedule a builder directly.
* @param target The target to resolve the builder name.
*/
getBuilderNameForTarget(target: Target): Promise<string>;
/**
* Validates the options against a builder schema. This uses the same methods as the
* scheduleTarget and scheduleBrowser methods to validate and apply defaults to the options.
* It can be generically typed, if you know which interface it is supposed to validate against.
* @param options A generic option object to validate.
* @param builderName The name of a builder to use. This can be gotten for a target by using the
* getBuilderForTarget() method on the context.
*/
validateOptions<T extends json.JsonObject = json.JsonObject>(options: json.JsonObject, builderName: string): Promise<T>;
/**
* Set the builder to running. This should be used if an external event triggered a re-run,
* e.g. a file watched was changed.
*/
reportRunning(): void;
/**
* Update the status string shown on the interface.
* @param status The status to set it to. An empty string can be used to remove the status.
*/
reportStatus(status: string): void;
/**
* Update the progress for this builder run.
* @param current The current progress. This will be between 0 and total.
* @param total A new total to set. By default at the start of a run this is 1. If omitted it
* will use the same value as the last total.
* @param status Update the status string. If omitted the status string is not modified.
*/
reportProgress(current: number, total?: number, status?: string): void;
/**
* API to report analytics. This might be undefined if the feature is unsupported. This might
* not be undefined, but the backend could also not report anything.
*/
readonly analytics: analytics.Analytics;
/**
* Add teardown logic to this Context, so that when it's being stopped it will execute teardown.
*/
addTeardown(teardown: () => Promise<void> | void): void;
}
/**
* An accepted return value from a builder. Can be either an Observable, a Promise or a vector.
*/
export declare type BuilderOutputLike = AsyncIterable<BuilderOutput> | SubscribableOrPromise<BuilderOutput> | BuilderOutput;
export declare function isBuilderOutput(obj: any): obj is BuilderOutput;
export declare function fromAsyncIterable<T>(iterable: AsyncIterable<T>): Observable<T>;
/**
* A builder handler function. The function signature passed to `createBuilder()`.
*/
export interface BuilderHandlerFn<A> {
/**
* Builders are defined by users to perform any kind of task, like building, testing or linting,
* and should use this interface.
* @param input The options (a JsonObject), validated by the schema and received by the
* builder. This can include resolved options from the CLI or the workspace.
* @param context A context that can be used to interact with the Architect framework.
* @return One or many builder output.
*/
(input: A, context: BuilderContext): BuilderOutputLike;
}
/**
* A Builder general information. This is generated by the host and is expanded by the host, but
* the public API contains those fields.
*/
export declare type BuilderInfo = json.JsonObject & {
builderName: string;
description: string;
optionSchema: json.schema.JsonSchema;
};
/**
* Returns a string of "project:target[:configuration]" for the target object.
*/
export declare function targetStringFromTarget({ project, target, configuration }: Target): string;
/**
* Return a Target tuple from a string.
*/
export declare function targetFromTargetString(str: string): Target;
/**
* Schedule a target, and forget about its run. This will return an observable of outputs, that
* as a a teardown will stop the target from running. This means that the Run object this returns
* should not be shared.
*
* The reason this is not part of the Context interface is to keep the Context as normal form as
* possible. This is really an utility that people would implement in their project.
*
* @param context The context of your current execution.
* @param target The target to schedule.
* @param overrides Overrides that are used in the target.
* @param scheduleOptions Additional scheduling options.
*/
export declare function scheduleTargetAndForget(context: BuilderContext, target: Target, overrides?: json.JsonObject, scheduleOptions?: ScheduleOptions): Observable<BuilderOutput>;