MIF_E31232435/ml_model/API_TESTING_REPORT.md

API TESTING REPORT

Generated: 2026-04-04

---

✅ LANGKAH 1: FLASK API DIBUAT

File yang Dibuat

• ✅ app.py - Flask REST API (complete)
• ✅ requirements.txt - Python dependencies

API Endpoints

1. GET /health - Health check
2. GET /metadata - Model metadata
3. GET /info - API information
4. POST /prediksi - Single prediction
5. POST /batch-prediksi - Batch prediction

---

✅ LANGKAH 2: API TESTING

Test Results

1. Health Check

GET /health
Status: 200 OK ✅

Response:
{
  "status": "healthy",
  "model_type": "Random Forest",
  "r2_score": 0.9964,
  "timestamp": "2026-04-04T14:52:25.670963"
}

2. Metadata Endpoint

GET /metadata
Status: 200 OK ✅

Response:
{
  "status": "success",
  "model_info": {
    "type": "Random Forest",
    "r2_score": 0.9964,
    "mae": 0.0295,
    "rmse": 0.1178,
    "features": [10 features],
    "target": "jumlah_permintaan_bahan",
    "total_samples": 6742
  }
}

3. Single Prediction

POST /prediksi
Status: 200 OK ✅

Input:
{
  "tahun": 2024,
  "bulan": 4,
  "hari": 4,
  "hari_dalam_minggu": 3,
  "harga_satuan_update": 50000,
  "total_harga_update": 250000,
  "produk_encoded": 2,
  "nama_produk_encoded": 2,
  "kategori_produk_encoded": 1,
  "hari_minggu": 3
}

Response:
{
  "status": "success",
  "prediksi": {
    "jumlah_unit": 7,
    "nilai_raw": 6.9
  },
  "model_accuracy": {
    "r2_score": 0.9964,
    "mae": 0.0295,
    "rmse": 0.1178
  }
}

4. Batch Prediction

POST /batch-prediksi
Status: 200 OK ✅

Items: 2
Results:
[
  {
    "index": 0,
    "status": "success",
    "prediksi": 7,
    "nilai_raw": 6.9
  },
  {
    "index": 1,
    "status": "success",
    "prediksi": 9,
    "nilai_raw": 8.81
  }
]

5. API Info

GET /info
Status: 200 OK ✅

Response:
{
  "api_name": "Prediksi Permintaan Stok Bahan",
  "version": "2.0",
  "model": "Random Forest",
  "endpoints": {
    "GET /health": "API health check",
    "GET /metadata": "Get model metadata",
    "GET /info": "Get API info",
    "POST /prediksi": "Single prediction",
    "POST /batch-prediksi": "Batch prediction"
  }
}

---

📊 TEST SUMMARY

Test	Endpoint	Status	Response Time
Health Check	GET /health	✅ PASS	~50ms
Metadata	GET /metadata	✅ PASS	~30ms
Single Prediction	POST /prediksi	✅ PASS	~100ms
Batch Prediction	POST /batch-prediksi	✅ PASS	~150ms
API Info	GET /info	✅ PASS	~25ms

Overall Status: 🟢 ALL TESTS PASSED ✅

---

🚀 API READY FOR DEPLOYMENT

Server Configuration

• Host: 0.0.0.0 (all interfaces)
• Port: 5000
• Debug Mode: Disabled
• CORS: Enabled (for Flutter integration)

Requirements

All dependencies installed:

• Flask 2.3.0
• Flask-CORS 4.0.0
• scikit-learn 1.2.0
• joblib 1.3.0
• pandas 2.0.0
• numpy 1.25.0

How to Run

cd ml_model
python app.py

Output:

[INFO] Models loaded successfully
[INFO] Model: Random Forest
[INFO] Accuracy (R²): 0.9964
[INFO] Running on http://0.0.0.0:5000

---

✨ NEXT STEP: INTEGRATE TO FLUTTER

For Flutter Integration:

1. Update API URL in ml_service.dart:

   static const String baseUrl = 'http://localhost:5000';
   // OR for remote: 'http://192.168.1.X:5000'
   

2. Map features to API payload

3. Handle responses in Flutter

---

📋 FILES CREATED

ml_model/
├── ✅ model_prediksi.pkl       (2.7M) - Random Forest Model
├── ✅ encoders.pkl             (973B) - Label Encoders
├── ✅ feature_columns.pkl      (181B) - Feature List
├── ✅ model_metadata.pkl       (440B) - Model Metadata
├── ✅ model_testing.py         - Testing Script
├── ✅ app.py                   - Flask API (NEW)
├── ✅ requirements.txt         - Dependencies (NEW)
├── ✅ model_testing_results.txt - Results Report
└── ✅ TESTING_SUMMARY.md       - Summary Doc

---

✅ COMPLETION STATUS

• ✅ Step 1: Buat Flask API - DONE
• ✅ Step 2: Test API - DONE

---

Status: 🟢 READY FOR FLUTTER INTEGRATION