Update services/sentiment.py

services/sentiment.py

@@ -1,26 +1,102 @@
 from transformers import pipeline
+import os
 
-classifier = pipeline(
-    "sentiment-analysis",
-    model="w11wo/indonesian-roberta-base-sentiment-classifier"
-)
+# 🔥 PATH MODEL (hasil fine-tuning)
+LOCAL_MODEL_PATH = "model/final_model"
 
+# 🔄 fallback model (pretrained)
+FALLBACK_MODEL = "w11wo/indonesian-roberta-base-sentiment-classifier"
+
+
+# 🚀 INIT MODEL
+def load_model():
+    try:
+        # 👉 cek apakah model fine-tuning ada
+        if os.path.exists(LOCAL_MODEL_PATH):
+            print("✅ Load fine-tuned model")
+            return pipeline("sentiment-analysis", model=LOCAL_MODEL_PATH)
+
+        else:
+            print("⚠️ Load fallback model (RoBERTa)")
+            return pipeline("sentiment-analysis", model=FALLBACK_MODEL)
+
+    except Exception as e:
+        print("❌ Gagal load model:", e)
+        return None
+
+
+# 🔥 LOAD SEKALI SAJA (BIAR CEPAT)
+classifier = load_model()
+
+
+# 🧠 NORMALISASI LABEL
+def normalize_label(label):
+    label = label.lower()
+
+    # 👉 untuk model huggingface (positive/negative)
+    if "positive" in label:
+        return "Positive"
+    elif "negative" in label:
+        return "Negative"
+    elif "neutral" in label:
+        return "Neutral"
+
+    # 👉 untuk model fine-tuned (LABEL_0,1,2)
+    if label == "label_0":
+        return "Negative"
+    elif label == "label_1":
+        return "Neutral"
+    elif label == "label_2":
+        return "Positive"
+
+    return "Neutral"
+
+
+# 🔍 PREDICT UTAMA
 def predict(texts):
     results = []
 
+    if classifier is None:
+        print("⚠️ Model tidak tersedia")
+        return ["Neutral"] * len(texts)
+
     try:
-        outputs = classifier(texts)
-        for o in outputs:
-            label = o['label'].lower()
+        # 🔥 batched prediction (lebih cepat)
+        outputs = classifier(texts, batch_size=8, truncation=True)
 
-            if "positive" in label:
-                results.append("Positive")
-            elif "negative" in label:
-                results.append("Negative")
-            else:
-                results.append("Neutral")
+        for o in outputs:
+            label = normalize_label(o['label'])
+            results.append(label)
 
-    except:
+    except Exception as e:
+        print("❌ Error saat prediksi:", e)
         results = ["Neutral"] * len(texts)
 
+    return results
+
+
+# 🔎 PREDICT SINGLE (opsional)
+def predict_single(text):
+    return predict([text])[0]
+
+
+# 📊 PREDICT + SCORE (opsional untuk analisis lebih lanjut)
+def predict_with_score(texts):
+    results = []
+
+    if classifier is None:
+        return [{"label": "Neutral", "score": 0}] * len(texts)
+
+    try:
+        outputs = classifier(texts, batch_size=8, truncation=True)
+
+        for o in outputs:
+            results.append({
+                "label": normalize_label(o['label']),
+                "score": round(o['score'], 4)
+            })
+
+    except Exception as e:
+        print("❌ Error:", e)
+
     return results