UPDATE: classfication inference and fastapi web

2-model-development/notebooks/classification.ipynb

@@ -1,11 +1,9 @@
 {
  "cells": [
   {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "747b3a43",
+   "cell_type": "markdown",
+   "id": "cc48b0c6",
    "metadata": {},
-   "outputs": [],
    "source": [
     "# Classification Model\n",
     "Program ini akan membuat model klasifikasi yang dapat memprediksi gaya belajar siswa (fast learner, consistent learner, dan reflective learner). Model klasifikasi akan berfokus pada Decision Tree Algorithm."
@@ -2528,10 +2526,18 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 1,
    "id": "e921578b",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "['Tryhard']\n"
+     ]
+    }
+   ],
    "source": [
     "import joblib\n",
     "import pandas as pd\n",

3-model-delivery/app.py

@@ -1,123 +0,0 @@
-from flask import Flask, request, jsonify
-from flask_cors import CORS
-import google.generativeai as genai
-import joblib
-import pandas as pd
-import os
-
-# Konfigurasi Gemini API
-API_KEY = "PASTE_GEMINI_API_KEY_HERE"  # Ganti dengan API Key kita
-
-genai.configure(api_key=API_KEY)
-
-# Pilih model yang digunakan
-# model yang digunakan bisa diganti sesuai kebutuhan
-model_gemini = genai.GenerativeModel('gemini-2.5-flash')
-
-# Load model klasifikasi ML
-MODEL_PATH = 'model_klasifikasi_gaya_belajar.joblib'
-print(f"[INFO] Sedang memuat model dari: {MODEL_PATH} ...")
-
-loaded_model = None
-
-# Cek apakah file model ada sebelum di-load
-if os.path.exists(MODEL_PATH):
-    try:
-        loaded_model = joblib.load(MODEL_PATH)
-        print("✅ Model berhasil dimuat!")
-    except Exception as e:
-        print(f"❌ Error saat load model: {e}")
-else:
-    print("⚠️ PERINGATAN: File model tidak ditemukan. Pastikan nama file benar.")
-
-# Setup Flask App
-app = Flask(__name__)
-CORS(app) # Mengizinkan akses dari Frontend (React)
-
-# Fungsi 1: Prediksi Label Gaya Belajar (klasifikasi ML)
-def predict_label(features):
-    if loaded_model is None:
-        return "Model Error"
-    
-    # DataFrame sesuai format input model saat training
-    df = pd.DataFrame([{
-        'weekly_completed_modules': features['weekly_completed_modules'],
-        'weekly_active_days': features['weekly_active_days'],
-        'weekly_submission_count': features['weekly_submission_count'],
-        'weekly_avg_exam_score': features['weekly_avg_exam_score']
-    }])
-    
-    try:
-        prediction = loaded_model.predict(df)
-        return prediction[0] # Mengambil hasil label pertama
-    except Exception as e:
-        print(f"Error pada prediksi ML: {e}")
-        return "Error Prediksi"
-
-# Fungsi 2: Generate Insight Pembelajaran (Gemini API)
-def generate_learning_insight(user_features, predicted_label):
-    
-    # Prompt yang akan dikirim ke Gemini
-    prompt = f"""
-    Bertindaklah sebagai "Learning Coach" untuk platform belajar coding.
-    
-    Data Siswa:
-    - Gaya Belajar: {predicted_label}
-    - Modul Selesai: {user_features.get('weekly_completed_modules')}
-    - Hari Aktif: {user_features.get('weekly_active_days')}
-    - Rata-rata Skor: {user_features.get('weekly_avg_exam_score')}
-    
-    Instruksi Utama:
-    Buatlah SATU paragraf pendek (maksimal 3-4 kalimat) yang berisi apresiasi, alasan mendapatkan label {predicted_label}, dan saran belajar singkat.
-
-    Aturan Format (PENTING):
-    1. Output harus PLAIN TEXT murni. 
-    2. DILARANG KERAS menggunakan format markdown seperti bold (**teks**), italic, atau bullet points.
-    3. Jangan gunakan tanda bintang (*) sama sekali.
-    4. Gunakan emoji secukupnya agar terkesan ramah (seperti Duolingo).
-    5. Gaya bahasa: Santai, suportif, dan mengalir.
-    6. Jangan sebut saya sebagai AI atau model bahasa.
-
-    Contoh Output yang BENAR:
-    "Wih, performamu minggu ini luar biasa banget! 🚀 Kamu layak disebut Speedrunner karena berhasil melahap 10 modul dalam seminggu dengan skor rata-rata 95 yang nyaris sempurna. Coba deh tantang dirimu bikin mini project dari materi tadi biar ilmunya makin nempel, semangat terus ya! 🔥"
-    """
-
-    try:
-        response = model_gemini.generate_content(prompt)
-        return response.text.strip()
-    except Exception as e:
-        print(f"Error pada Gemini API: {e}")
-        # Fallback response jika API bermasalah
-        return f"Halo {predicted_label}! Pertahankan performa belajarmu ya. 🚀"
-
-# Endpoint API
-@app.route('/api/analyze-student', methods=['POST'])
-def analyze():
-    # 1. Terima data JSON dari request
-    data = request.json
-    
-    # Gunakan .get() untuk menghindari error jika data kosong (default 0)
-    features = {
-        'weekly_completed_modules': data.get('weekly_completed_modules', 0),
-        'weekly_active_days': data.get('weekly_active_days', 0),
-        'weekly_submission_count': data.get('weekly_submission_count', 0),
-        'weekly_avg_exam_score': data.get('weekly_avg_exam_score', 0)
-    }
-    
-    # 2. Proses Prediksi ML
-    pred_label = predict_label(features)
-    
-    # 3. Generate Insight AI
-    insight = generate_learning_insight(features, pred_label)
-    
-    # 4. Return response JSON
-    return jsonify({
-        "status": "success",
-        "prediction": pred_label,
-        "ai_insight": insight,
-        "features_received": features
-    })
-
-if __name__ == '__main__':
-    print("🚀 Server Flask berjalan di port 5000...")
-    app.run(debug=True, port=5000)

3-model-delivery/main.py

@@ -0,0 +1,59 @@
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+import google.generativeai as genai
+import joblib
+import pandas as pd
+import os
+
+API_KEY = "AIzaSyD1USyNFVhLM8tpyAZr47C757B_UWZNuOk"
+genai.configure(api_key=API_KEY)
+model_gemini = genai.GenerativeModel("gemini-2.5-flash")
+
+MODEL_PATH = 'model_klasifikasi_gaya_belajar.joblib'
+loaded_model = joblib.load(MODEL_PATH) if os.path.exists(MODEL_PATH) else None
+
+app = FastAPI()
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+class StudentFeatures(BaseModel):
+    weekly_completed_modules: int
+    weekly_active_days: int
+    weekly_submission_count: int
+    weekly_avg_exam_score: float
+
+def predict_label(features):
+    df = pd.DataFrame([features.dict()])
+    return loaded_model.predict(df)[0] if loaded_model else "Model Error"
+
+def generate_learning_insight(features, predicted_label):
+    prompt = f"""
+    Bertindaklah sebagai Learning Coach.
+
+    Gaya Belajar: {predicted_label}
+    Modul Selesai: {features.weekly_completed_modules}
+    Hari Aktif: {features.weekly_active_days}
+    Rata-rata Skor: {features.weekly_avg_exam_score}
+
+    Buat paragraf singkat tanpa markdown, tanpa simbol bintang, bahasa santai dan emotikon.
+    """
+    res = model_gemini.generate_content(prompt)
+    return res.text.strip()
+
+@app.post("/api/analyze-student")
+def analyze(features: StudentFeatures):
+    label = predict_label(features)
+    insight = generate_learning_insight(features, label)
+    return {
+        "status": "success",
+        "prediction": label,
+        "ai_insight": insight,
+        "features_received": features.dict()
+    }

3-model-delivery/model_klasifikasi_gaya_belajar.joblib

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:855bde45aa224f5791f6d346a5a94998474d6295f3d7e76f8df677c99bebb9af
-size 5865
+oid sha256:aab8d3f1688f29abc56136cbce952f7909c7dacba749f68fedc1814191454af7
+size 5737