Update app.py

app.py

@@ -1,9 +1,12 @@
 import os
 import numpy as np
 import pandas as pd
+import torch
+import timm
+import torch.nn as nn
 from flask import Flask, request, jsonify, render_template_string
-from tensorflow.keras.models import load_model
-from tensorflow.keras.utils import load_img, img_to_array
+from torchvision import transforms
+from PIL import Image
 from math import radians, cos, sin, sqrt, atan2
 import joblib
 import requests
@@ -19,30 +22,46 @@ app = Flask(__name__)
 UPLOAD_FOLDER = '/tmp/uploads'
 os.makedirs(UPLOAD_FOLDER, exist_ok=True)
 
-# Load model dan data
-soil_model = load_model("models/best_vit_model.pth")
+# Label klasifikasi tanah
+class_labels = ['Black Soil', 'Cinder Soil', 'Laterite Soil', 'Peat Soil', 'Yellow Soil']
+num_classes = len(class_labels)
+
+# Load model ViT (PyTorch)
+soil_model = timm.create_model("vit_base_patch16_224", pretrained=False, num_classes=num_classes)
+soil_model.head = nn.Linear(soil_model.head.in_features, num_classes)
+soil_model.load_state_dict(torch.load("models/best_vit_model.pth", map_location=torch.device("cpu")))
+soil_model.eval()
+
+# Load model crop recommendation
 crop_model = joblib.load("models/model_random_forest.joblib")
 crop_model_label = joblib.load("models/label_encoder.joblib")
+
+# Load data referensi
 soil_df = pd.read_csv("data/soil_data_with_class.csv")
 agri_df = pd.read_csv("data/tips_menanam_dan_manfaat_tanaman.csv")
-class_labels = ['Black Soil', 'Cinder Soil', 'Laterite Soil', 'Peat Soil', 'Yellow Soil']
 
-# Fungsi preprocessing gambar
+# Fungsi preprocessing gambar (PyTorch style)
 def preprocess_image(img_path):
-    img = load_img(img_path, target_size=(224, 224))
-    img_array = img_to_array(img) / 255.0
-    return np.expand_dims(img_array, axis=0)
+    transform = transforms.Compose([
+        transforms.Resize((224, 224)),
+        transforms.ToTensor(),
+        transforms.Normalize(mean=[0.5]*3, std=[0.5]*3)
+    ])
+    img = Image.open(img_path).convert("RGB")
+    return transform(img).unsqueeze(0)
 
 # Fungsi prediksi jenis tanah
 def predict_soil_type(img_path):
-    img_array = preprocess_image(img_path)
-    prediction = soil_model.predict(img_array)
-    predicted_index = np.argmax(prediction)
+    img_tensor = preprocess_image(img_path)
+    with torch.no_grad():
+        outputs = soil_model(img_tensor)
+        probabilities = torch.softmax(outputs, dim=1).numpy()[0]
+    predicted_index = np.argmax(probabilities)
     predicted_class = class_labels[predicted_index]
-    accuracy = float(prediction[0][predicted_index])
+    accuracy = float(probabilities[predicted_index])
     return predicted_class, accuracy
 
-# Fungsi hitung jarak terdekat
+# Hitung jarak dengan haversine
 def find_nearest_soil_data_weighted(soil_type, lat, lon, n_points=1):
     filtered = soil_df[soil_df['Class_Name'] == soil_type].copy()
     if filtered.empty:
@@ -54,7 +73,7 @@ def find_nearest_soil_data_weighted(soil_type, lat, lon, n_points=1):
         lat2, lon2 = radians(row['Location_Latitude']), radians(row['Location_Longitude'])
         dlat = lat2 - user_lat
         dlon = lon2 - user_lon
-        a = sin(dlat / 2) ** 2 + cos(user_lat) * cos(lat2) * sin(dlon / 2) ** 2
+        a = sin(dlat / 2)**2 + cos(user_lat) * cos(lat2) * sin(dlon / 2)**2
         c = 2 * atan2(sqrt(a), sqrt(1 - a))
         return 6371 * c
 
@@ -72,7 +91,7 @@ def find_nearest_soil_data_weighted(soil_type, lat, lon, n_points=1):
         "distance_km": float(row['Distance_km'])
     }
 
-# Ambil data cuaca
+# Ambil data cuaca dari OpenWeather
 def get_weather_data(lat, lon):
     url = f"https://api.openweathermap.org/data/2.5/weather?lat={lat}&lon={lon}&appid={OPENWEATHER_API_KEY}&units=metric"
     res = requests.get(url)
@@ -84,7 +103,7 @@ def get_weather_data(lat, lon):
         "humidity": float(data['main']['humidity']),
     }
 
-# Fungsi untuk mengambil tips bertani dan manfaat dari agri_df
+# Ambil tips bertani dan manfaat tanaman
 def get_farming_tips(df, crop_name):
     crop_name_str = str(crop_name)
     match = df[df['Nama Tanaman'].str.lower() == crop_name_str.lower()]
@@ -98,7 +117,7 @@ def get_farming_tips(df, crop_name):
         }
     return None
 
-# Fungsi reverse geocoding
+# Reverse geocoding
 def get_location_name(lat, lon):
     try:
         url = f"https://nominatim.openstreetmap.org/reverse?lat={lat}&lon={lon}&format=json"
@@ -136,7 +155,7 @@ def index():
     """
     return render_template_string(html_content)
 
-# Endpoint analisis utama
+# Endpoint analisis
 @app.route('/analyze', methods=['POST'])
 def analyze():
     if 'image' not in request.files:
@@ -222,5 +241,6 @@ def analyze():
 
     return jsonify(result)
 
+# Jalankan server
 if __name__ == '__main__':
     app.run(debug=True)