Premier Workflow

Créez et exécutez votre premier workflow Genesis AI.

---

🎯 Objectif

À la fin de ce guide, vous aurez créé et exécuté un workflow simple qui :

1. Récupère des données depuis une API
2. Les transforme avec une fonction simple
3. Stocke le résultat

---

📝 Étape 1 : Définir le Workflow

Créez un fichier hello-workflow.ts :

// hello-workflow.ts
import { WorkflowBuilder } from "@igon7/core";
import { TemporalExecutor } from "@igon7/executor";

// 1. Définir le workflow
const workflow = new WorkflowBuilder("hello-world")
  
  // Noeud 1 : Fetch data
  .addNode("fetch-data", async () => {
    console.log("📡 Fetching data from API...");
    const response = await fetch("https://jsonplaceholder.typicode.com/posts/1");
    const data = await response.json();
    console.log("✅ Data fetched:", data.title);
    return data;
  })
  
  // Noeud 2 : Transform data
  .addNode("transform-data", async (inputs) => {
    console.log("🔄 Transforming data...");
    const fetchData = inputs["fetch-data"];
    const transformed = {
      id: fetchData.id,
      title: fetchData.title.toUpperCase(),
      transformedAt: new Date().toISOString(),
    };
    console.log("✅ Data transformed:", transformed);
    return transformed;
  }, {
    dependsOn: ["fetch-data"],
  })
  
  // Noeud 3 : Store result
  .addNode("store-result", async (inputs) => {
    console.log("💾 Storing result...");
    const transformed = inputs["transform-data"];
    
    // Simulation de stockage
    await new Promise(resolve => setTimeout(resolve, 1000));
    
    const result = {
      success: true,
      storedData: transformed,
      storedAt: new Date().toISOString(),
    };
    console.log("✅ Result stored:", result);
    return result;
  }, {
    dependsOn: ["transform-data"],
  })
  
  .build();

// 2. Exécuter le workflow
async function main() {
  console.log("🚀 Starting workflow execution...\n");
  
  const executor = new TemporalExecutor({
    temporalAddress: "localhost:7233",
    namespace: "genesis",
    taskQueue: "igon7-workflows",
  });

  try {
    const result = await executor.execute(workflow, {
      workflowId: `hello-world-${Date.now()}`,
    });

    console.log("\n🎉 Workflow completed successfully!");
    console.log("Final result:", result);
  } catch (error) {
    console.error("\n❌ Workflow failed:", error);
    throw error;
  }
}

// Lancer l'exécution
if (import.meta.main) {
  await main();
}

---

▶️ Étape 2 : Exécuter

# Assurez-vous que Temporal tourne
docker-compose up -d  # Dans genesis-temporal

# Assurez-vous que igon7 est installé
cd igon7_deno
pnpm install

# Exécuter le workflow
deno run --allow-all --node-modules-dir hello-workflow.ts

Sortie Attendue

🚀 Starting workflow execution...

📡 Fetching data from API...
✅ Data fetched: sunt aut facere repellat provident occaecati excepturi optio reprehenderit
🔄 Transforming data...
✅ Data transformed: { id: 1, title: 'SUNT AUT FACERE...', transformedAt: '...' }
💾 Storing result...
✅ Result stored: { success: true, storedData: {...}, storedAt: '...' }

🎉 Workflow completed successfully!
Final result: { success: true, storedData: {...}, storedAt: '...' }

---

🔍 Étape 3 : Monitorer

Via Temporal UI

Ouvrez http://localhost:8080 et recherchez votre workflow par ID.

Via igon7 Monitor

import { WorkflowMonitor } from "@igon7/monitor";

const monitor = new WorkflowMonitor({
  cloudApiEndpoint: "http://localhost:3000",
  apiKey: process.env.MONITOR_API_KEY,
});

// S'abonner aux événements
monitor.on("workflow.started", (event) => {
  console.log(`Workflow ${event.workflowId} started`);
});

monitor.on("node.completed", (event) => {
  console.log(`Node ${event.nodeId} completed in ${event.duration}ms`);
});

monitor.on("workflow.completed", (event) => {
  console.log(`Workflow completed with result:`, event.result);
});

// Récupérer les métriques
const metrics = await monitor.getMetrics(workflowId);
console.log(metrics);

---

🧪 Étape 4 : Tester

// hello-workflow.test.ts
import { assertEquals } from "@std/assert";
import { WorkflowBuilder } from "@igon7/core";
import { TestExecutor } from "@igon7/testing";

Deno.test("Hello Workflow - should execute all nodes", async () => {
  const workflow = new WorkflowBuilder("test-hello")
    .addNode("fetch", async () => ({ value: 42 }))
    .addNode("transform", async (inputs) => ({
      result: inputs["fetch"].value * 2
    }), { dependsOn: ["fetch"] })
    .build();

  const executor = new TestExecutor();
  const result = await executor.execute(workflow);

  assertEquals(result.success, true);
  assertEquals(result.result.result, 84);
});

---

🎓 Concepts Clés

Concept	Description
Node	Une tâche individuelle dans le workflow
Edge	Dépendance entre deux noeuds
DAG	Graphe orienté acyclique du workflow
Executor	Moteur qui exécute le workflow
Monitor	Outil de suivi d'exécution

---

📚 Suite

• Architecture igon7
• Workflow Patterns
• API Reference

---

Temps estimé : 20 minutes
Difficulté : Débutant