Sayiir

https://docs.sayiir.dev
Admin
Sayiir is a fast, simple durable workflow engine with graph-based execution, continuation-driven...
Tokens:98,568
Snippets:1,050
Trust Score:2.2
Update:4 months ago
Show doc for...
Tokens:
Raw
### Complete Workflow Example (Python)

Source: https://docs.sayiir.dev/guides/signals-and-events

A full Python example demonstrating a workflow that creates an order, waits for a manager's approval signal (with a 24-hour timeout), and then fulfills the order. It includes task definitions and backend setup.

```python
from datetime import timedelta
from sayiir import (
    task, Flow, run_durable_workflow, send_signal,
    resume_workflow, InMemoryBackend
)


@task
def create_order(order_id: int) -> dict:
    print(f"Creating order {order_id}")
    return {"order_id": order_id, "status": "pending"}


@task
def fulfill(approval: dict) -> str:
    approver = approval.get("by", "unknown")
    print(f"Fulfilling order (approved by {approver})")
    return f"Fulfilled order (approved by {approver})"


workflow = (
    Flow("approval")
    .then(create_order)
    .wait_for_signal("manager_approval", timeout=timedelta(hours=24))
    .then(fulfill)
    .build()
)


backend = InMemoryBackend()


# Start workflow
status = run_durable_workflow(workflow, "order-42", 42, backend=backend)
print(f"Status: {status}")  # WaitingForSignal


# External system sends approval
send_signal("order-42", "manager_approval", {"by": "alice"}, backend=backend)


```

--------------------------------

### PostgreSQL Connection URL Examples

Source: https://docs.sayiir.dev/reference/configuration

Illustrates various examples of PostgreSQL connection URLs, including local development setups, custom ports, and connection parameters like SSL mode and timeouts.

```text
# Local development
postgresql://postgres:password@localhost/sayiir


# With custom port
postgresql://user:pass@db.example.com:5433/workflows


# Connection parameters
postgresql://user:pass@host/db?sslmode=require&connect_timeout=10
```

--------------------------------

### Start, Signal, and Resume Workflow (Rust)

Source: https://docs.sayiir.dev/guides/signals-and-events

Illustrates starting a workflow with `CheckpointingRunner`, sending an event to it, and resuming its execution. This uses the `sayiir_runtime` crate and requires a `backend` implementation.

```rust
use sayiir_runtime::CheckpointingRunner;


let runner = CheckpointingRunner::new(backend.clone());


// Start the workflow
let status = runner.run(&workflow, "order-42", 42u64).await?;
println!("{:?}", status);  // WaitingForSignal


// Signal can be sent immediately or hours later
let payload = json!({"by": "alice"});
backend.send_event("order-42", "manager_approval", payload).await?;


// Resume to continue execution
let status = runner.resume(&workflow, "order-42").await?;
println!("{:?}", status);  // Completed
```

--------------------------------

### Install Sayiir with pnpm

Source: https://docs.sayiir.dev/getting-started/nodejs

Install the Sayiir package using the pnpm package manager.

```bash
pnpm add sayiir
```

--------------------------------

### Rust Task Metadata Examples

Source: https://docs.sayiir.dev/reference/configuration

Configure individual task behavior and documentation using Rust.

```rust
#[task(
    id = "process_payment",
    timeout = "30s",
    retries = 3,
    backoff = "2s",
    tags = "payment",
    tags = "external",
    tags = "critical",
    description = "Processes customer payment via Stripe API"
)]
async fn charge_card(order: Order) -> Result<Receipt, BoxError> {
    Ok(Receipt { receipt_id: "R123", status: "paid" })
}


// Or construct TaskMetadata manually
use sayiir_core::task::{TaskMetadata, RetryPolicy};
use std::time::Duration;


let metadata = TaskMetadata {
    display_name: Some("Process Payment".to_string()),
    description: Some("Processes customer payment via Stripe API".to_string()),
    timeout: Some(Duration::from_secs(30)),
    retry_policy: Some(RetryPolicy {
        max_retries: 3,
        initial_delay: Duration::from_secs(2),
        backoff_multiplier: 2.0,
    }),
    tags: vec!["payment".to_string(), "external".to_string(), "critical".to_string()],
};
```

--------------------------------

### Install Sayiir

Source: https://docs.sayiir.dev/getting-started/python

Install the Sayiir Python package using pip.

```bash
pip install sayiir
```

--------------------------------

### Python Database Connection

Source: https://docs.sayiir.dev/reference/configuration

Python example for configuring the PostgresBackend with connection pooling details.

```python
# Python: connection pooling is built into the Rust backend
backend = PostgresBackend("postgresql://user:pass@host/sayiir")
```

--------------------------------

### Python Task Metadata Example

Source: https://docs.sayiir.dev/reference/configuration

Configure individual task behavior and documentation using Python.

```python
from sayiir import task, RetryPolicy


@task(
    "process_payment",
    timeout="30s",
    retries=RetryPolicy(max_retries=3, initial_delay_secs=2.0),
    tags=["payment", "external", "critical"],
    description="Processes customer payment via Stripe API"
)
def charge_card(order: dict) -> dict:
    return {"receipt_id": "R123", "status": "paid"}
```

--------------------------------

### Node.js Task Metadata Example

Source: https://docs.sayiir.dev/reference/configuration

Configure individual task behavior and documentation using Node.js.

```javascript
import { task } from "sayiir";


const chargeCard = task("process_payment", (order: Record<string, unknown>) => {
    return { receipt_id: "R123", status: "paid" };
}, {
    timeout: "30s",
    retry: {
        maxAttempts: 3,
        initialDelay: "2s",
        backoffMultiplier: 2.0,
    },
    tags: ["payment", "external", "critical"],
    description: "Processes customer payment via Stripe API",
});
```

--------------------------------

### Workflow Definition and Instance Example

Source: https://docs.sayiir.dev/concepts/architecture

Illustrates a simple workflow definition and its corresponding instances, showing their states and identifiers.

```text
Definition: fetch_user → send_email         (definition_hash = "a1b2c3")
Instance 1: instance_id = "order-123"       (InProgress, at send_email)
Instance 2: instance_id = "order-456"       (Completed)
Instance 3: instance_id = "order-789"       (Failed)
```

--------------------------------

### Rust Database Connection

Source: https://docs.sayiir.dev/reference/configuration

Rust example for initializing the PostgresBackend, noting that connection pooling is managed internally.

```rust
// Rust: connection pool is managed internally
let backend = PostgresBackend::new("postgresql://user:pass@host/sayiir").await?;
```

--------------------------------

### Run Cargo Project

Source: https://docs.sayiir.dev/tutorials/background-jobs-rust

Execute the full example cargo project locally. This command initiates the video processing pipeline, including real ffmpeg transcoding.

```bash
cargo run
```

--------------------------------

### TaskMetadata Example

Source: https://docs.sayiir.dev/reference/rust-api

Example of creating TaskMetadata, which includes display name, description, timeout, retry policy, and tags. This metadata is associated with a specific task.

```rust
use sayiir_core::task::TaskMetadata;
use std::time::Duration;


let metadata = TaskMetadata {
    display_name: Some("Process Order".to_string()),
    description: Some("Validates and processes customer orders".to_string()),
    timeout: Some(Duration::from_secs(30)),
    retry_policy: Some(policy),
    tags: vec!["orders".to_string(), "payment".to_string()],
};

```

--------------------------------

### Start, Signal, and Resume Workflow (Node.js)

Source: https://docs.sayiir.dev/guides/signals-and-events

Demonstrates the lifecycle of a workflow in Node.js: starting it with `runDurableWorkflow`, sending a signal using `sendSignal`, and resuming it with `resumeWorkflow`. Ensure the workflow definition and backend are correctly provided.

```javascript
import { runDurableWorkflow, sendSignal, resumeWorkflow } from "sayiir";


// Start the workflow
const status1 = runDurableWorkflow(workflow, "order-42", 42, backend);
console.log(status1.status); // "awaiting_signal"


// Signal can be sent immediately or hours later
sendSignal("order-42", "manager_approval", { by: "alice" }, backend);


// Resume to continue execution
const status2 = resumeWorkflow(workflow, "order-42", backend);
console.log(status2.status); // "completed"
```

--------------------------------

### Start Worker and Get Handle

Source: https://docs.sayiir.dev/reference/python-api

Start a distributed worker using its `start` method, passing a list of workflows it can execute. This returns a `WorkerHandle` for lifecycle control.

```python
workflow = Flow("pipeline").then(my_task).build()
handle = worker.start([workflow])
```

--------------------------------

### Python Distributed Worker Setup

Source: https://docs.sayiir.dev/guides/distributed-workers

Sets up a distributed worker in Python to poll for and execute tasks from a PostgreSQL backend. This example demonstrates defining tasks, building a workflow, and starting the worker.

```Python
import signal
from sayiir import task, Flow, PostgresBackend
from sayiir.worker import Worker


@task
def fetch_order(order_id: int) -> dict:
    return {"order_id": order_id, "total": 99.99}


@task
def charge_payment(order: dict) -> str:
    return f"charged ${order['total']}"


@task
def send_confirmation(receipt: str) -> str:
    return f"confirmed: {receipt}"


workflow = (
    Flow("order-pipeline")
    .then(fetch_order)
    .then(charge_payment)
    .then(send_confirmation)
    .build()
)


backend = PostgresBackend("postgresql://localhost/sayiir")
worker = Worker("worker-1", backend, poll_interval=2.0)
handle = worker.start([workflow])

```

--------------------------------

### Install Sayiir for Cloudflare Workers

Source: https://docs.sayiir.dev/getting-started/cloudflare

Install the Sayiir package for Cloudflare Workers using pnpm.

```bash
pnpm add @sayiir/cloudflare
```

--------------------------------

### Python Workflow Example

Source: https://docs.sayiir.dev/

Defines and runs a simple workflow in Python using Sayiir tasks for fetching user data and sending an email.

```python
from sayiir import task, Flow, run_workflow


@task
def fetch_user(user_id: int) -> dict:
    return {"id": user_id, "name": "Alice"}


@task
def send_email(user: dict) -> str:
    return f"Sent welcome to {user['name']}"


workflow = Flow("welcome").then(fetch_user).then(send_email).build()
result = run_workflow(workflow, 42)

```

--------------------------------

### Rust Workflow Example

Source: https://docs.sayiir.dev/

Defines and runs a simple workflow in Rust using Sayiir tasks with specified timeouts and retries.

```rust
use sayiir_runtime::prelude::*;


#[task(timeout = "30s", retries = 3)]
async fn fetch_user(id: u64) -> Result<User, BoxError> {
    db.get_user(id).await
}


#[task]
async fn send_email(user: User) -> Result<(), BoxError> {
    email_service.send_welcome(&user).await
}


let workflow = workflow! {
    name: "welcome",
    steps: [fetch_user, send_email]
}.unwrap();

```

--------------------------------

### Simple Sayiir Workflow Example

Source: https://docs.sayiir.dev/getting-started/cloudflare

Implement a basic Sayiir workflow with two tasks: fetching user data and sending a welcome email. The workflow is built using `task` and `flow` and executed via an `Engine`.

```typescript
import { task, flow, Engine } from "@sayiir/cloudflare";


const fetchUser = task("fetch-user", async (id: number) => {
  const res = await fetch(`https://api.example.com/users/${id}`);
  return res.json() as Promise<{ id: number; name: string }>;
});


const sendEmail = task("send-email", async (user: { id: number; name: string }) => {
  return `Sent welcome to ${user.name}`;
});


const onboarding = flow<number>("onboarding")
  .then(fetchUser)
  .then(sendEmail)
  .build();


export default {
  async fetch(request: Request, env: Env): Promise<Response> {
    const engine = await Engine.create(env.DB);
    const status = await engine.run(onboarding, "onboard-42", 42);
    return Response.json(status);
  },
};
```

--------------------------------

### Python: Run Workflow and Send Signal

Source: https://docs.sayiir.dev/guides/signals-and-events

Starts a workflow, prints its status (expected to be 'WaitingForSignal'), and then sends the 'manager_approval' signal.

```python
from sayiir import run_durable_workflow, send_signal, resume_workflow


# Start the workflow
status = run_durable_workflow(workflow, "order-42", 42, backend=backend)
print(status)  # WaitingForSignal


# Signal can be sent immediately or hours later
send_signal("order-42", "manager_approval", {"by": "alice"}, backend=backend)

```

--------------------------------

### Node.js/TypeScript Workflow Example

Source: https://docs.sayiir.dev/

Defines and runs a simple workflow in Node.js/TypeScript using Sayiir tasks for fetching user data and sending an email.

```javascript
import { task, flow, runWorkflow } from "sayiir";


const fetchUser = task("fetch-user", (userId: number) => {
  return { id: userId, name: "Alice" };
});


const sendEmail = task("send-email", (user: { id: number; name: string }) => {
  return `Sent welcome to ${user.name}`;
});


const workflow = flow<number>("welcome")
  .then(fetchUser)
  .then(sendEmail)
  .build();


const result = await runWorkflow(workflow, 42);

```

--------------------------------

### .start()

Source: https://docs.sayiir.dev/reference/nodejs-api

Starts the worker and returns a handle for lifecycle control. This spawns a background thread that polls for tasks, claims them, and dispatches them to registered task functions.

```APIDOC
## .start()

### Description
Start the worker and return a handle for lifecycle control. Spawns a background thread that polls for tasks, claims them, and dispatches them to registered task functions.

### Returns:
`WorkerHandle`

### Example
```javascript
const handle = worker.start();
process.on("SIGTERM", () => handle.shutdown());
```
```

--------------------------------

### Local Testing with Sayiir's InMemoryBackend

Source: https://docs.sayiir.dev/comparisons/vs-step-functions

Demonstrates local testing of Sayiir workflows using an in-memory backend, requiring no external services or setup. This approach is suitable for rapid development and testing.

```python
# Test with in-memory backend (no database, no setup)
from sayiir import InMemoryBackend, run_workflow


backend = InMemoryBackend()
result = await run_workflow(workflow, test_input, backend=backend)
assert result == expected_output
```

--------------------------------

### Define and Run a Simple Workflow

Source: https://docs.sayiir.dev/getting-started/nodejs

Create a workflow by defining tasks and chaining them using `flow()`. This example demonstrates a simple user fetching and email sending process.

```typescript
import { task, flow, runWorkflow } from "sayiir";


const fetchUser = task("fetch-user", (userId: number) => {
  return { id: userId, name: "Alice" };
});


const sendEmail = task("send-email", (user: { id: number; name: string }) => {
  return `Sent welcome to ${user.name}`;
});


const workflow = flow<number>("welcome")
  .then(fetchUser)
  .then(sendEmail)
  .build();


const result = await runWorkflow(workflow, 42);
console.log(result); // "Sent welcome to Alice"
```

--------------------------------

### Use PostgreSQL Backend for Durable Workflows

Source: https://docs.sayiir.dev/getting-started/nodejs

Configure `PostgresBackend` for production use with durable workflows. This example assumes `DATABASE_URL` is set in the environment.

```typescript
import { PostgresBackend } from "sayiir";


// Connects and runs migrations automatically
const backend = PostgresBackend.connect(process.env.DATABASE_URL!);
const status = runDurableWorkflow(workflow, "order-123", 42, backend);
```

--------------------------------

### Cloudflare Workers Runtime Example

Source: https://docs.sayiir.dev/changelog/v05

Demonstrates how to run Sayiir workflows natively within Cloudflare Workers, utilizing D1 for persistence and handling scheduled events.

```typescript
import { task, flow, Engine } from "@sayiir/cloudflare";


const onboarding = flow<number>("onboarding")
  .then(fetchUser)
  .then(sendEmail)
  .build();


export default {
  async fetch(request: Request, env: Env): Promise<Response> {
    const engine = await Engine.create(env.DB);
    const status = await engine.run(onboarding, "onboard-42", 42);
    return Response.json(status);
  },
  async scheduled(event: ScheduledEvent, env: Env): Promise<void> {
    const engine = await Engine.create(env.DB);
    await engine.resumeAll(onboarding);
  },
};

```

--------------------------------

### Pluggable Storage Backends in Sayiir

Source: https://docs.sayiir.dev/concepts/architecture

Shows how to initialize different storage backend implementations. Use InMemory for testing and PostgreSQL for production, or implement a custom backend.

```rust
// In-memory (testing)
let backend = InMemoryBackend::new();
```

```rust
// PostgreSQL (production)
let backend = PostgresBackend::new(pool);
```

```rust
// Custom (bring your own)
impl PersistentBackend for MyBackend { ... }
```

--------------------------------

### Define and Run a Workflow with a Loop

Source: https://docs.sayiir.dev/guides/loops-and-iteration

This snippet demonstrates defining a workflow with setup, loop, and finalize tasks, then running it. The loop task 'countdown' reduces a number until it reaches 1, returning a final value.

```typescript
const setup = task("setup", (x: number) => x + 10);


const countdown = task("countdown", (n: number) => {
    if (n <= 1) return LoopResult.done(0);
    return LoopResult.again(n - 1);
});


const finalize = task("finalize", (x: number) => x * 100);


const wf = flow<number>("pipeline")
    .then(setup)
    .loop(countdown, { maxIterations: 20 })
    .then(finalize)
    .build();


await runWorkflow(wf, 5);
// 5 → setup → 15 → countdown (15→…→1→done 0) → finalize → 0
```

--------------------------------

### Initialize PostgresBackend (Python)

Source: https://docs.sayiir.dev/guides/durable-workflows

Use PostgresBackend for production. State survives process restarts. Ensure you have the necessary connection details.

```Python
from sayiir import InMemoryBackend, PostgresBackend


# Production
backend = PostgresBackend(
    host="localhost",
    port=5432,
    user="postgres",
    password="password",
    database="sayiir"
)
```

--------------------------------

### Python Workflow Context Example

Source: https://docs.sayiir.dev/roadmap

Demonstrates how to use a shared workflow context to get and set state within a Python task. The context can be used to pass data between tasks without explicit manual passing.

```python
@task
def search_web(query: dict, ctx: WorkflowContext) -> list[dict]:
    depth = ctx.get("depth", "detailed")
    results = do_search(query["topic"], depth)
    ctx.set("web_results_count", len(results))
    return results

```

--------------------------------

### Create and Configure Database

Source: https://docs.sayiir.dev/reference/configuration

SQL commands to create a dedicated database and user for Sayiir, granting necessary privileges.

```sql
CREATE DATABASE sayiir;
CREATE USER sayiir_user WITH PASSWORD 'secure_password';
GRANT ALL PRIVILEGES ON DATABASE sayiir TO sayiir_user;
```

--------------------------------

### Initialize PostgresBackend (Node.js)

Source: https://docs.sayiir.dev/guides/durable-workflows

Use PostgresBackend for production in Node.js. State survives process restarts. Connect using a PostgreSQL connection string.

```Node.js
import { InMemoryBackend, PostgresBackend } from "sayiir";


// Production
const backend = PostgresBackend.connect(
    "postgresql://postgres:password@localhost:5432/sayiir"
);
```

--------------------------------

### Start a Worker and Handle Shutdown

Source: https://docs.sayiir.dev/reference/nodejs-api

Starts the worker process, which begins polling for and executing tasks. It returns a handle that can be used to gracefully shut down the worker upon receiving a SIGTERM signal.

```typescript
const handle = worker.start();
process.on("SIGTERM", () => handle.shutdown());

```

--------------------------------

### Initialize PostgresBackend (Rust)

Source: https://docs.sayiir.dev/guides/durable-workflows

Use PostgresBackend for production in Rust. State survives process restarts. Requires async context for connection.

```Rust
use sayiir_persistence::InMemoryBackend;
use sayiir_postgres::PostgresBackend;
use sayiir_runtime::prelude::*;


// Production
let backend = PostgresBackend::<JsonCodec>::connect(
    "postgresql://localhost/sayiir"
).await?;
```

--------------------------------

### Start Parallel Execution with .fork()

Source: https://docs.sayiir.dev/reference/nodejs-api

Use .fork() to start parallel execution of multiple branches. Each branch is defined using the branch() function. The results are later combined using .join().

```typescript
import { flow, branch, task } from "sayiir";


const sendEmail = task("send-email", (order) => ({ emailSent: true }));
const shipOrder = task("ship-order", (order) => ({ shipped: true }));


const workflow = flow<Order>("process")
  .then(chargePayment)
  .fork([
    branch("email", sendEmail),
    branch("shipping", shipOrder),
  ])
  .join("finalize", ([email, shipping]) => ({ ...email, ...shipping }))
  .build();
```

--------------------------------

### .fork

Source: https://docs.sayiir.dev/reference/python-api

Starts a fork for parallel execution of workflow branches.

```APIDOC
## .fork()

### Description
Start a fork for parallel execution.

### Returns
ForkBuilder instance
```

--------------------------------

### .route

Source: https://docs.sayiir.dev/reference/python-api

Starts conditional branching based on a routing key determined by a key function.

```APIDOC
## .route(key_fn, *, keys)

### Description
Start conditional branching based on a routing key. The `keys` parameter declares all valid branch keys upfront — Sayiir validates `.branch()` calls against this set immediately and checks exhaustiveness at `.done()`.
Declaring `keys=` is the recommended approach. It catches typos and missing branches at build time. Only omit it when keys are determined at runtime (e.g., dynamically built workflows).

### Parameters
* `key_fn` — Callable that returns a string routing key.
* `keys` (list[str], keyword-only) — All valid routing keys. `.branch()` calls validate against this set, and `.done()` checks exhaustiveness.

### Returns
BranchBuilder instance
```

--------------------------------

### Initialize Python Tracing

Source: https://docs.sayiir.dev/guides/observability

Call this once at startup to initialize Sayiir's tracing in your Python application.

```python
import sayiir


# Initialize logging (call once at startup)
sayiir.init_tracing()
```

--------------------------------

### Sayiir Server Architecture Overview

Source: https://docs.sayiir.dev/server

Illustrates the basic architecture of Sayiir Server, showing its interaction with user applications, PostgreSQL, worker pools, and the dashboard.

```text
Your App → Sayiir Server (gRPC) → PostgreSQL
               ↕
         Worker Pool (auto-scaled)
               ↕
         Dashboard (web UI)
```

--------------------------------

### Initialize InMemoryBackend (Python)

Source: https://docs.sayiir.dev/guides/durable-workflows

Use InMemoryBackend for development and testing. State is lost when the process stops.

```Python
from sayiir import InMemoryBackend, PostgresBackend


# Development/testing
backend = InMemoryBackend()
```

--------------------------------

### Configure PostgreSQL Backend in Python

Source: https://docs.sayiir.dev/reference/configuration

Initialize the PostgreSQL backend for durable storage in Python. The connection URL specifies the database credentials and location.

```python
from sayiir import PostgresBackend


# Connection URL format
backend = PostgresBackend(
    "postgresql://username:password@host:port/database"
)


# Example URLs
backend = PostgresBackend("postgresql://postgres:password@localhost:5432/sayiir")
backend = PostgresBackend("postgresql://user:pass@db.example.com/workflows")
```

--------------------------------

### Node.js Countdown Loop

Source: https://docs.sayiir.dev/guides/loops-and-iteration

A Node.js example of a countdown loop. It uses the `LoopResult.again` to continue and `LoopResult.done` to exit.

```typescript
import { task, flow, runWorkflow, LoopResult } from "sayiir";


const countdown = task("countdown", (n: number) => {
    if (n <= 1) return LoopResult.done(0);
    return LoopResult.again(n - 1);
});


const wf = flow<number>("countdown").loop(countdown).build();


const result = await runWorkflow(wf, 5);
// 5 	4 	3 	2 	1 (done) 	result = 0

```

--------------------------------

### Python: Connect to PostgreSQL Backend

Source: https://docs.sayiir.dev/guides/postgres-production

Initialize a PostgresBackend with a connection string and use it to run a durable workflow. The backend automatically handles migrations on first connection.

```python
from sayiir import PostgresBackend, Flow, run_durable_workflow


backend = PostgresBackend("postgresql://localhost/sayiir")
status = run_durable_workflow(workflow, "order-123", 42, backend=backend)
```

--------------------------------

### Worker Processing Multiple Workflows (Python)

Source: https://docs.sayiir.dev/guides/distributed-workers

Define and start a worker that can process tasks from multiple distinct workflows.

```python
order_wf = Flow("orders").then(process_order).build()
email_wf = Flow("emails").then(send_email).build()


handle = worker.start([order_wf, email_wf])
```

--------------------------------

### Run Jaeger with Docker

Source: https://docs.sayiir.dev/guides/observability

Start a local Jaeger instance using Docker for development and testing of OpenTelemetry traces.

```shell
docker run -d --name jaeger \
  -p 16686:16686 \
  -p 4317:4317 \
  jaegertracing/jaeger:latest
```

--------------------------------

### Import Sayirre Prelude

Source: https://docs.sayiir.dev/getting-started/rust

Import the prelude to bring common types and macros into scope.

```rust
use sayiir_runtime::prelude::*;
// Re-exports: WorkflowBuilder, CheckpointingRunner, PooledWorker,
// WorkerHandle, InMemoryBackend, JsonCodec, TaskRegistry,
// task (macro), workflow (macro), etc.
```

--------------------------------

### Define and Run a Workflow with Tasks

Source: https://docs.sayiir.dev/guides/distributed-workers

Defines tasks and a workflow, then starts a worker to process them. Includes graceful shutdown on SIGTERM.

```typescript
import { task, flow, PostgresBackend, Worker } from "sayiir";


const fetchOrder = task("fetch-order", (orderId: number) => {
  return { orderId, total: 99.99 };
});


const chargePayment = task("charge-payment", (order: { total: number }) => {
  return `charged $${order.total}`;
});


const sendConfirmation = task("send-confirmation", (receipt: string) => {
  return `confirmed: ${receipt}`;
});


const workflow = flow<number>("order-pipeline")
  .then(fetchOrder)
  .then(chargePayment)
  .then(sendConfirmation)
  .build();


const backend = PostgresBackend.connect(process.env.DATABASE_URL!);
const worker = new Worker("worker-1", backend, [workflow], {
  pollInterval: "2s",
});
const handle = worker.start();


// Graceful shutdown on SIGTERM
process.on("SIGTERM", () => handle.shutdown());

```

--------------------------------

### Rust Retry Policy Example

Source: https://docs.sayiir.dev/reference/configuration

Configure automatic retry behavior for tasks that may fail transiently using Rust.

```rust
use std::time::Duration;
use sayiir_core::task::RetryPolicy;


let policy = RetryPolicy {
    max_retries: 3,
    initial_delay: Duration::from_secs(1),
    backoff_multiplier: 2.0,
};


#[task(retries = 3, backoff = "1s", backoff_multiplier = 2.0)]
async fn flaky_task(data: Data) -> Result<Output, BoxError> {
    // May fail and retry with exponential backoff
    Ok(process(data))
}
```

--------------------------------

### Sayiir Checkpoint-Resume Recovery Example

Source: https://docs.sayiir.dev/comparisons/vs-prefect

Illustrates Sayiir's checkpoint-resume durability model. After each task, its output is stored, allowing workflows to resume from the last completed task upon failure, avoiding re-execution of already finished work.

```python
# Sayiir checkpoint-resume example
@task
async def task_a() -> str:
    print("Task A executed")
    return "A done"


@task
async def task_b(a_result: str) -> str:
    print("Task B executed")
    raise Exception("Crash!")


@task
async def task_c(b_result: str) -> str:
    print("Task C executed")
    return "C done"


workflow = Flow("recovery").then(task_a).then(task_b).then(task_c).build()


# First run: A completes, B crashes
try:
    await run_workflow(workflow, backend=postgres_backend)
except Exception:
    pass  # Crash after A completes


# Resume: A is skipped (already completed), B retries, C runs
result = await run_workflow(workflow, backend=postgres_backend)
# Output: Only "Task B executed" and "Task C executed" print
# task_a is not re-executed

```

--------------------------------

### Default Workflow with JsonCodec

Source: https://docs.sayiir.dev/guides/serialization-and-versioning

Demonstrates creating a workflow using the default JsonCodec. No explicit codec configuration is needed.

```rust
use sayiir_runtime::prelude::*


// JsonCodec is the default — no explicit codec needed
let workflow = workflow! {
    name: "order_pipeline",
    registry: registry,
    steps: [validate, charge, ship]
}.unwrap();

```

--------------------------------

### Node.js Retry Policy Example

Source: https://docs.sayiir.dev/reference/configuration

Configure automatic retry behavior for tasks that may fail transiently using Node.js.

```javascript
import { task } from "sayiir";


const flakyTask = task("flaky_task", (data: Record<string, unknown>) => {
    // May fail and retry with exponential backoff
    return process(data);
}, {
    retry: {
        maxAttempts: 3,
        initialDelay: "1s",
        backoffMultiplier: 2.0,
    },
});
```

--------------------------------

### Configure PostgreSQL Backend in Rust

Source: https://docs.sayiir.dev/reference/configuration

Initialize the PostgreSQL backend in Rust for persistent storage. The `new` method takes a connection URL and requires an asynchronous context.

```rust
use sayiir_postgres::PostgresBackend;


// Connection URL format
let backend = PostgresBackend::new(
    "postgresql://username:password@host:port/database"
).await?;


// Example URLs
let backend = PostgresBackend::new(
    "postgresql://postgres:password@localhost:5432/sayiir"
).await?;


let backend = PostgresBackend::new(
    "postgresql://user:pass@db.example.com/workflows"
).await?;
```

--------------------------------

### Python Retry Policy Examples

Source: https://docs.sayiir.dev/reference/configuration

Configure automatic retry behavior for tasks that may fail transiently using Python.

```python
from sayiir import RetryPolicy, task


# Simple: int shorthand (uses 1s initial delay, 2× backoff)
@task(retries=3)
def flaky_task(data: dict) -> dict:
    return process(data)


# Full control: custom backoff timing
@task(retries=RetryPolicy(
    max_retries=3,
    initial_delay_secs=0.5,
    backoff_multiplier=3.0,
))
def critical_task(data: dict) -> dict:
    return process(data)
```

--------------------------------

### Configure PostgreSQL Backend in Node.js

Source: https://docs.sayiir.dev/reference/configuration

Set up the PostgreSQL backend for production workflows in Node.js using a connection URL. The `connect` method is used for initialization.

```javascript
import { PostgresBackend } from "sayiir";


// Connection URL format
const backend = PostgresBackend.connect(
    "postgresql://username:password@host:port/database"
);


// Example URLs
const backend1 = PostgresBackend.connect("postgresql://postgres:password@localhost:5432/sayiir");
const backend2 = PostgresBackend.connect("postgresql://user:pass@db.example.com/workflows");
```

--------------------------------

### WorkflowClient Initialization and Operations (Node.js)

Source: https://docs.sayiir.dev/changelog/v04

Initializes a WorkflowClient with a PostgresBackend and demonstrates submitting, cancelling, pausing, unpausing, sending signals, and checking the status of a workflow.

```javascript
import { WorkflowClient, PostgresBackend } from "sayiir";


const backend = PostgresBackend.connect(process.env.DATABASE_URL!);
const client = new WorkflowClient(backend);


// Submit a workflow — a worker will pick it up
const status = client.submit(workflow, "order-42", { items: [1, 2, 3] });


// Lifecycle operations
client.cancel("order-42", { reason: "Out of stock", cancelledBy: "system" });
client.pause("order-42", { reason: "Maintenance window" });
client.unpause("order-42");
client.sendSignal("order-42", "approval", { approved: true });
const current = client.status("order-42");
```

--------------------------------

### Worker ID Naming Conventions

Source: https://docs.sayiir.dev/reference/configuration

Examples of Rust code snippets illustrating different strategies for generating unique worker IDs.

```rust
// By hostname
format!("worker-{}", hostname)


// By hostname and process ID
format!("worker-{}-{}", hostname, process_id)


// By role and instance
format!("email-worker-{}", instance_number)
```

--------------------------------

### Get Workflow Instance Status

Source: https://docs.sayiir.dev/reference/python-api

Retrieve the current status of a workflow instance. The status object provides information about completion and output.

```python
status = client.status("order-42")
if status.is_completed():
    print(status.output)
```

--------------------------------

### WorkflowClient Initialization and Operations (Python)

Source: https://docs.sayiir.dev/changelog/v04

Initializes a WorkflowClient with a PostgresBackend and demonstrates submitting, cancelling, pausing, unpausing, sending signals, and checking the status of a workflow.

```Python
from sayiir import WorkflowClient, PostgresBackend


backend = PostgresBackend("postgresql://localhost/sayiir")
client = WorkflowClient(backend)


# Submit a workflow — a worker will pick it up
status = client.submit(workflow, "order-42", {"items": [1, 2, 3]})


# Lifecycle operations
client.cancel("order-42", reason="Out of stock", cancelled_by="system")
client.pause("order-42", reason="Maintenance window")
client.unpause("order-42")
client.send_signal("order-42", "approval", {"approved": True})
status = client.status("order-42")
```

--------------------------------

### Resume Durable Workflow

Source: https://docs.sayiir.dev/reference/nodejs-api

Resumes a previously started durable workflow from its last saved checkpoint using the provided instance ID and backend.

```javascript
import { resumeWorkflow } from "sayiir";


const status = resumeWorkflow(workflow, "run-001", backend);
```

--------------------------------

### Initialize InMemoryBackend (Node.js)

Source: https://docs.sayiir.dev/guides/durable-workflows

Use InMemoryBackend for development and testing in Node.js. State is lost when the process stops.

```Node.js
import { InMemoryBackend, PostgresBackend } from "sayiir";


// Development/testing
const backend = new InMemoryBackend();
```

--------------------------------

### Python Task with Timeout and Retries

Source: https://docs.sayiir.dev/guides/retries-and-timeouts

Example of a Python task configured with a 5-second timeout and 3 retries. If the task exceeds the timeout, it will be retried.

```python
from sayiir import task
import time

@task(timeout="5s", retries=3)
def slow_operation(data: str) -> str:
    # If this takes >5 seconds, it will be retried
    time.sleep(10)
    return data

```

--------------------------------

### Pluggable Codecs in Sayiir

Source: https://docs.sayiir.dev/concepts/architecture

Demonstrates how to instantiate different codec implementations for serialization. Choose based on performance (rkyv), readability (JSON), or custom needs.

```rust
// Zero-copy for maximum performance (default)
let codec = RkyvCodec::new();
```

```rust
// Human-readable for debugging
let codec = JsonCodec::new();
```

```rust
// Custom format (implement Codec trait)
let codec = MyCustomCodec::new();
```

--------------------------------

### Sayiir: Unified Task Model for Workflow Logic

Source: https://docs.sayiir.dev/comparisons/vs-temporal

This Sayiir example illustrates its unified task model where I/O operations are directly included within tasks, simplifying workflow development by eliminating the need to split logic between workflows and activities.

```python
# Sayiir: Just write tasks
@task
async def fetch_user(user_id: int) -> dict:
    return await db.get_user(user_id)


@task
async def send_email(user: dict) -> None:
    await email_service.send_welcome(user)


workflow = Flow("welcome").then(fetch_user).then(send_email).build()

```

--------------------------------

### PostgreSQL Connection String Format

Source: https://docs.sayiir.dev/guides/postgres-production

Standard PostgreSQL URL format for establishing connections. Examples show basic, authenticated, and SSL-enabled connections.

```sql
postgresql://[user[:password]@][host][:port][/database][?parameters]
```

--------------------------------

### Sayiir: Define a Durable Task with @task

Source: https://docs.sayiir.dev/comparisons/vs-elsa

Any Python async function can be made durable by decorating it with `@task`. This example shows sending an email.

```python
# Sayiir: Any function becomes a task
@task
async def send_email(to: str, subject: str, body: str):
    async with aiosmtplib.SMTP("smtp.example.com") as smtp:
        msg = MIMEText(body)
        msg["Subject"] = subject
        msg["To"] = to
        await smtp.send_message(msg)

```

--------------------------------

### Resume All Workflow Instances with Options

Source: https://docs.sayiir.dev/getting-started/cloudflare

Resume all eligible workflow instances, including those parked due to delays or waiting for signals, and those at risk of eviction. This example includes options to set a stale threshold and a limit for the number of instances to resume.

```typescript
async scheduled(event: ScheduledEvent, env: Env): Promise<void> {
  const engine = await Engine.create(env.DB);
  await engine.resumeAll(orderApproval, { staleAfter: 60, limit: 20 });
}
```

--------------------------------

### Connect to PostgresBackend

Source: https://docs.sayiir.dev/reference/nodejs-api

Establishes a connection to a PostgreSQL database for production use, ensuring state survives process restarts. Requires PostgreSQL 13 or higher and supports auto-migration.

```javascript
import { PostgresBackend } from "sayiir";


const backend = PostgresBackend.connect("postgresql://user:pass@localhost:5432/sayiir");
```

--------------------------------

### Parallel Execution with Fork and Join

Source: https://docs.sayiir.dev/getting-started/python

Use `.fork()` to start parallel branches and `.join()` to combine their results. This is useful for running independent tasks concurrently.

```python
from sayiir import task, Flow, branch, run_workflow


@task
def validate_payment(order: dict) -> dict:
    return {"payment": "valid"}


@task
def check_inventory(order: dict) -> dict:
    return {"stock": "available"}


workflow = (
    Flow("checkout")
    .fork([
        branch("payment", validate_payment),
        branch("inventory", check_inventory),
    ])
    .join("finalize", lambda results: {**results[0], **results[1]})
    .build()
)


result = run_workflow(workflow, {"id": 1})
```

--------------------------------

### Rust Task with Timeout and Retries

Source: https://docs.sayiir.dev/guides/retries-and-timeouts

An example of a Rust task configured with a 5-second timeout and 3 retries. Exceeding the timeout will trigger a retry for the task.

```rust
#[task(timeout = "5s", retries = 3, backoff = "2s")]
async fn slow_operation(data: String) -> Result<String, BoxError> {
    // If this takes >5 seconds, it will be retried
    tokio::time::sleep(Duration::from_secs(10)).await;
    Ok(data)
}

```

--------------------------------

### WorkflowClient Initialization and Operations (Rust)

Source: https://docs.sayiir.dev/changelog/v04

Initializes a WorkflowClient with a PostgresBackend and demonstrates submitting, cancelling, pausing, unpausing, sending signals, and checking the status of a workflow.

```rust
use sayiir_runtime::WorkflowClient;


let backend = PostgresBackend::<JsonCodec>::connect(url)?;
let client = WorkflowClient::new(backend);


// Submit a workflow — a worker will pick it up
let (status, output) = client.submit(&workflow, "order-42", input).await?;


// Lifecycle operations
client.cancel("order-42", Some("Out of stock".into()), Some("system".into())).await?;
client.pause("order-42", Some("Maintenance window".into()), None).await?;
client.unpause("order-42").await?;
client.send_event("order-42", "approval", payload).await?;
let status = client.status("order-42").await?;
```

--------------------------------

### Add Sayirre Dependencies

Source: https://docs.sayiir.dev/getting-started/rust

Add the core runtime and your chosen backend to your Cargo.toml file.

```toml
[dependencies]
sayiir-runtime = "0.4"       # core runtime + runners + macros


# Optional — pick your backend
sayiir-persistence = "0.4"   # InMemoryBackend (dev/testing)
sayiir-postgres = "0.4"      # PostgreSQL backend (production)
```

--------------------------------

### SQL: Query Active Workflows

Source: https://docs.sayiir.dev/guides/postgres-production

Example SQL query to monitor active workflows by counting instances grouped by their status within the last hour.

```sql
SELECT status, COUNT(*)
FROM sayiir_workflow_instances
WHERE created_at > NOW() - INTERVAL '1 hour'
GROUP BY status;
```

--------------------------------

### Get a Specific Task Result

Source: https://docs.sayiir.dev/reference/python-api

Retrieve the result of a specific task within a workflow instance. Returns the decoded output or None if the task was not executed.

```python
result = client.get_task_result("order-42", "validate_order")
if result is not None:
    print(result)
```

--------------------------------

### Rust: Define and Run a Workflow with Signals

Source: https://docs.sayiir.dev/guides/signals-and-events

This Rust snippet defines a workflow using Sayiir's core and persistence crates. It demonstrates starting a workflow, sending an event to trigger a signal, and resuming the workflow to completion using an in-memory backend.

```rust
use sayiir_core::prelude::*;
use sayiir_persistence::InMemoryBackend;
use sayiir_runtime::CheckpointingRunner;
use serde_json::json;
use std::time::Duration;


#[task]
async fn create_order(id: u64) -> Result<serde_json::Value, BoxError> {
    println!("Creating order {}", id);
    Ok(json!({ "order_id": id, "status": "pending" }))
}


#[task]
async fn fulfill(approval: serde_json::Value) -> Result<String, BoxError> {
    let approver = approval["by"].as_str().unwrap_or("unknown");
    println!("Fulfilling order (approved by {})", approver);
    Ok(format!("Fulfilled order (approved by {})", approver))
}


#[tokio::main]
async fn main() -> Result<(), BoxError> {
    let ctx = WorkflowContext::default();
    let backend = InMemoryBackend::new();
    let runner = CheckpointingRunner::new(backend.clone());


    let workflow = WorkflowBuilder::new(ctx)
        .with_registry()
        .then_fn(create_order)
        .wait_for_signal(
            "approval_wait",
            "manager_approval",
            Some(Duration::from_secs(86400))
        )
        .then_fn(fulfill)
        .build()?;


    // Start workflow
    let status = runner.run(&workflow, "order-42", 42u64).await?;
    println!("Status: {:?}", status);  // WaitingForSignal


    // External system sends approval
    let payload = json!({"by": "alice"});
    backend.send_event("order-42", "manager_approval", payload).await?;


    // Resume workflow
    let status = runner.resume(&workflow, "order-42").await?;
    println!("Final status: {:?}", status);  // Completed


    Ok(())
}

```

--------------------------------

### Node.js Durable Workflow Execution and Resumption

Source: https://docs.sayiir.dev/guides/retries-and-timeouts

Illustrates starting a durable workflow in Node.js and resuming it. Retry state persistence ensures continuity after process interruptions.

```javascript
import { runDurableWorkflow, resumeWorkflow } from "sayiir";


// Start workflow with retries
const status = runDurableWorkflow(workflow, "job-1", "https://flaky-api.com", backend);


// Process crashes during backoff...


// Resume later - retry state is preserved
const status2 = resumeWorkflow(workflow, "job-1", backend);

```

--------------------------------

### Task Metadata Access within Task

Source: https://docs.sayiir.dev/reference/python-api

Access task metadata like timeout, retries, and tags from the TaskExecutionContext. This example shows how to print these values if available.

```python
@task(timeout="30s", retries=3, tags=["io"])
def fetch_data(url: str) -> dict:
    ctx = get_task_context()
    if ctx:
        print(f"Timeout: {ctx.metadata.timeout_secs}s")
        print(f"Tags: {ctx.metadata.tags}")
    return do_fetch(url)
```

--------------------------------

### Sayiir Distributed Workers Example

Source: https://docs.sayiir.dev/comparisons/vs-prefect

Demonstrates how to set up and use Sayiir's optional distributed worker pool for executing tasks in a separate process. This model simplifies distributed execution without requiring complex deployment configurations.

```python
# Sayiir distributed workers (optional)
from sayiir import Flow, WorkerPool, task


@task
async def expensive_task(data: str) -> str:
    # Long-running computation
    return f"Processed {data}"


workflow = Flow("distributed").then(expensive_task).build()


# Start worker pool (separate process)
pool = WorkerPool(backend=postgres_backend, concurrency=10)
pool.start()


# Submit from main application
await run_workflow(workflow, "data", backend=postgres_backend)
# Worker pool automatically picks up and executes

```

--------------------------------

### Use PostgreSQL Backend for Durability

Source: https://docs.sayiir.dev/getting-started/python

Swap to `PostgresBackend` for production environments. This backend connects and runs migrations automatically.

```python
from sayiir import PostgresBackend


# Connects and runs migrations automatically
backend = PostgresBackend("postgresql://localhost/sayiir")
status = run_durable_workflow(workflow, "order-123", 42, backend=backend)
```

--------------------------------

### Integrate Zod for Input Validation

Source: https://docs.sayiir.dev/getting-started/nodejs

Use Zod schemas to validate workflow inputs and task inputs. This example defines an `OrderSchema` for validating order details.

```typescript
import { z } from "zod";
import { task, flow, runWorkflow } from "sayiir";


const OrderSchema = z.object({
  orderId: z.string(),
  amount: z.number().positive(),
});


const processOrder = task("process-order", (order) => {
  // order is validated and typed as { orderId: string; amount: number }
  return { status: "ok", message: `Processed $${order.amount}` };
}, {
  input: OrderSchema,
});


const workflow = flow("typed").then(processOrder).build();
const result = await runWorkflow(workflow, { orderId: "1", amount: 99.99 });
```

--------------------------------

### Run Workflow with Persistence (Production)

Source: https://docs.sayiir.dev/reference/python-api

For production, use `run_workflow` with an `instance_id` and a `backend` (like `PostgresBackend`) for full checkpointing. The output is still returned directly.

```python
from sayiir import run_workflow, PostgresBackend


# Production — same function, just add params
backend = PostgresBackend("postgresql://localhost/sayiir")
result = run_workflow(workflow, 21, instance_id="run-001", backend=backend)
```

--------------------------------

### Get Task Execution Context

Source: https://docs.sayiir.dev/reference/nodejs-api

Retrieve the current task execution context from within a running task. Returns null if called outside of a task execution.

```typescript
import { task, getTaskContext } from "sayiir";


const myTask = task("my-task", (data: { value: number }) => {
  const ctx = getTaskContext();
  if (ctx) {
    console.log(`Running task ${ctx.taskId} in workflow ${ctx.workflowId}`);
    console.log(`Instance: ${ctx.instanceId}`);
  }
  return { result: data.value * 2 };
});
```

--------------------------------

### Initialize InMemoryBackend (Rust)

Source: https://docs.sayiir.dev/guides/durable-workflows

Use InMemoryBackend for development and testing in Rust. State is lost when the process stops.

```Rust
use sayiir_persistence::InMemoryBackend;
use sayiir_postgres::PostgresBackend;
use sayiir_runtime::prelude::*;


// Development/testing
let backend = InMemoryBackend::new();
```
Tokens:
Raw