@probitas/core
Core scenario definition types and utilities.
This package provides the fundamental type definitions that represent scenario structures, along with utilities for loading scenario files and filtering scenarios using selectors. It serves as the foundation layer that other Probitas packages build upon.
Links
- GitHub Repository
- @probitas/probitas - Main package (recommended for most users)
Related Packages
| Package | Description |
|---|---|
| @probitas/builder | Uses these types to build scenarios |
| @probitas/runner | Executes scenario definitions |
| @probitas/discover | Discovers scenario files to load |
Type Definitions
The type system is designed around immutable data structures:
ScenarioDefinition- Complete scenario with name, options, and entriesStepDefinition- Individual step with function, options, and sourceSetupDefinition- Setup hook with cleanup function supportResourceDefinition- Named resource with fn functionStepContext- Context object passed to all functionsEntry- Discriminated union of step, setup, or resource
Options Types
ScenarioOptions- Scenario-level configuration (tags, default step options)StepOptions- Step execution settings (timeout, retry strategy)- Import
Originfrom@probitas/core/originfor file and line information
Function Types
StepFunction- Signature for step execution functionsSetupFunction- Signature for setup hooks (returns cleanup)ResourceFunction- Signature for resource creation functionsSetupCleanup- Return type of setup functions
Loader Utilities
loadScenarios- Load scenario definitions from file pathsLoadScenariosOptions- Options for the loader
Selector Utilities
Selectors provide powerful filtering capabilities:
applySelectors- Filter scenarios using selector stringsparseSelector- Parse selector string into Selector objectsmatchesSelector- Check if a scenario matches a single selectorSelector- Parsed selector objectSelectorType- Type of selector ("tag" or "name")
Installation
deno add jsr:@probitas/coreInterfaces
#ScenarioDefinition
interface ScenarioDefinitionComplete, immutable definition of a scenario.
This is the core type produced by the builder and consumed by the runner. It contains everything needed to execute a scenario: its name, options, and ordered sequence of entries (steps, resources, setups).
Examples
Typical scenario structure
import type { ScenarioDefinition } from "@probitas/core";
// Created by: scenario("Login Flow").step(...).build()
const definition: ScenarioDefinition = {
name: "Login Flow",
tags: ["auth", "smoke"],
steps: [
{
kind: "resource",
name: "api",
fn: () => fetch,
timeout: 30000,
retry: { maxAttempts: 1, backoff: "linear" },
},
{
kind: "step",
name: "Login",
fn: () => {},
timeout: 30000,
retry: { maxAttempts: 1, backoff: "linear" },
},
{
kind: "step",
name: "Verify",
fn: () => {},
timeout: 30000,
retry: { maxAttempts: 1, backoff: "linear" },
},
],
origin: { path: "/tests/auth.probitas.ts", line: 5 },
};
console.log(definition);
| Name | Description |
|---|---|
name | Human-readable scenario name (displayed in reports and CLI) |
tags | Tags for filtering and organizing scenarios. |
steps | Ordered sequence of entries (resources → setups → steps) |
origin | Origin where the scenario was defined |
Properties
- readonly
namestringHuman-readable scenario name (displayed in reports and CLI)
Ordered sequence of entries (resources → setups → steps)
- readonly
origin?OriginOrigin where the scenario was defined
#ScenarioMetadata
interface ScenarioMetadataSerializable scenario metadata (without executable functions).
Used by the JSON reporter and tooling to output scenario information without including non-serializable function references.
Examples
JSON reporter output
{
"name": "Login Flow",
"options": { "tags": ["auth"], "stepOptions": { ... } },
"entries": [
{ "kind": "step", "value": { "name": "Login", "options": { ... } } }
],
"origin": { "path": "/tests/auth.probitas.ts", "line": 5 }
}
| Name | Description |
|---|---|
name | Scenario name |
tags | Tags for filtering and organizing scenarios. |
steps | Entry metadata (functions omitted for serialization) |
origin | Origin where the scenario was defined |
Properties
- readonly
namestringScenario name
Entry metadata (functions omitted for serialization)
- readonly
origin?OriginOrigin where the scenario was defined
#ScenarioOptions
interface ScenarioOptionsConfiguration options for scenario execution.
Defines metadata and default behavior for an entire scenario.
Examples
const options: ScenarioOptions = {
tags: ["api", "integration", "slow"],
stepOptions: {
timeout: 60000,
retry: { maxAttempts: 2, backoff: "linear" }
}
};
| Name | Description |
|---|---|
tags | Tags for filtering and organizing scenarios. |
stepOptions | Default options applied to all steps in this scenario. |
Properties
Default options applied to all steps in this scenario.
Individual steps can override these defaults by specifying their own options in the
.step()call.
#StepContext
interface StepContextExecution context provided to steps, resources, and setup hooks.
The context provides access to:
- Previous step results with full type inference
- All accumulated results as a typed tuple
- Named resources registered with
.resource() - Shared storage for cross-step communication
- Abort signal for timeout and cancellation handling
Examples
Accessing previous result
import type { StepContext } from "@probitas/core";
// Steps receive context with the previous step's result
const stepFn = (ctx: StepContext) => {
// ctx.previous contains the result from the previous step
const prev = ctx.previous as { id: number };
console.log(prev.id); // Access typed result
};
console.log(stepFn);
Using shared store
import type { StepContext } from "@probitas/core";
// The store is shared across all steps in a scenario
const setupFn = (ctx: StepContext) => {
ctx.store.set("startTime", Date.now());
};
const stepFn = (ctx: StepContext) => {
const start = ctx.store.get("startTime") as number;
console.log(`Elapsed: ${Date.now() - start}ms`);
};
console.log(setupFn, stepFn);
| Name | Description |
|---|---|
index | Current step index (0-based). |
previous | Result from the previous step. |
results | All accumulated results as a typed tuple. |
store | Shared key-value storage for cross-step communication. |
resources | Named resources registered with `.resource()`. |
signal | Abort signal that fires on timeout or manual cancellation. |
Properties
- readonly
indexnumberCurrent step index (0-based).
Useful for conditional logic based on position in the scenario.
- readonly
previousunknownResult from the previous step.
Fully typed based on what the previous step returned. For the first step, this is
unknown. - readonly
resultsreadonly unknown[]All accumulated results as a typed tuple.
Allows accessing any previous result by index:
import type { StepContext } from "@probitas/core"; const stepFn = (ctx: StepContext) => { ctx.results[0]; // First step's result ctx.results[1]; // Second step's result }; console.log(stepFn); - readonly
storeMap<string, unknown>Shared key-value storage for cross-step communication.
Use this for data that doesn't fit the step result pattern, such as metadata or configuration set during setup.
- readonly
resourcesRecord<string, unknown>Named resources registered with
.resource().Resources are typed based on their registration:
import type { StepContext } from "@probitas/core"; interface DbResource { query(sql: string): unknown; } // Access resources registered with .resource() const stepFn = (ctx: StepContext) => { const db = ctx.resources["db"] as DbResource; return db.query("SELECT 1"); }; console.log(stepFn); - readonly
signal?AbortSignalAbort signal that fires on timeout or manual cancellation.
Pass this to fetch() or other APIs that support AbortSignal for proper timeout handling.
#StepDefinition
interface StepDefinition<T = unknown>Immutable definition of a scenario step.
Contains all information needed to execute a single step: the step function, its options, and debugging metadata.
| Name | Description |
|---|---|
kind | — |
name | Human-readable step name (displayed in reports) |
fn | Step function to execute |
timeout | Maximum execution time in milliseconds. |
retry | Retry configuration for handling transient failures |
origin | Origin where the step was defined (for error messages) |
Properties
- readonly
kind"step" | "resource" | "setup" - readonly
namestringHuman-readable step name (displayed in reports)
Step function to execute
- readonly
timeoutnumberMaximum execution time in milliseconds.
If the step takes longer, a
TimeoutErroris thrown. Default: 30000 (30 seconds) - readonly
retry{ maxAttempts: number; backoff: "linear" | "exponential" }Retry configuration for handling transient failures
- readonly
origin?OriginOrigin where the step was defined (for error messages)
#StepOptions
interface StepOptionsConfiguration options for individual step execution.
Controls timeout and retry behavior for a step. These options can be set at:
- Step level (highest priority)
- Scenario level (applies to all steps in scenario)
- Default values (30s timeout, no retry)
Examples
const options: StepOptions = {
timeout: 60000, // 60 seconds
retry: {
maxAttempts: 3,
backoff: "exponential" // Waits 1s, 2s, 4s between retries
}
};
| Name | Description |
|---|---|
timeout | Maximum execution time in milliseconds. |
retry | Retry configuration for handling transient failures |
Properties
- readonly
timeout?numberMaximum execution time in milliseconds.
If the step takes longer, a
TimeoutErroris thrown. Default: 30000 (30 seconds) - readonly
retry?{ maxAttempts?: number; backoff?: "linear" | "exponential" }Retry configuration for handling transient failures
Types
#SetupCleanup
type SetupCleanup = void | unknown | Disposable | AsyncDisposableCleanup handler returned by setup functions.
Setup functions can return various cleanup mechanisms that are automatically invoked after the scenario completes (regardless of success or failure).
Supported cleanup patterns:
void/undefined: No cleanup needed() => void: Synchronous cleanup function() => Promise<void>: Async cleanup functionDisposable: Object with[Symbol.dispose]()methodAsyncDisposable: Object with[Symbol.asyncDispose]()method
#StepFunction
type StepFunction<T = unknown> = (ctx: StepContext) => unknownFunction signature for step execution.
A step function receives the execution context and returns a value (sync or async) that becomes available to subsequent steps.
#StepMetadata
type StepMetadata = Omit<StepDefinition, "fn">Serializable step metadata (without the function).
Used for JSON output, tooling, and inspection without executing code.
