feat: redis buat sendiri

app.py

@@ -1,18 +1,19 @@
 from fastapi import FastAPI, UploadFile, File
 from fastapi.middleware.cors import CORSMiddleware
-from transformers import pipeline
-from PIL import Image
+from transformers import pipeline, AutoTokenizer
 from PIL import Image
 import numpy as np
 import uvicorn
 from tensorflow.keras.models import load_model
-from transformers import (AutoTokenizer)
 import os
+import hashlib
+import diskcache
+import json
 
 # === Inisialisasi FastAPI ===
 app = FastAPI()
 
-# === CORS (opsional) ===
+# === CORS ===
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
@@ -21,14 +22,17 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-# === Load Model Klasifikasi Gambar ===
-model_path = os.path.join(os.path.dirname(__file__), "saved_model", "multidisease_model.h5")
+# === Inisialisasi Cache ===
+cache = diskcache.Cache("/tmp/cache")
+
+# === Load Model Gambar ===
+model_path = os.path.join(
+    os.path.dirname(__file__), "saved_model", "multidisease_model.h5"
+)
 if not os.path.exists(model_path):
     raise FileNotFoundError(f"Model tidak ditemukan di path: {model_path}")
 image_model = load_model(model_path)
 
-
-# Label (ubah sesuai model Anda)
 label_map = {
     0: "BacterialBlight",
     1: "Blast",
@@ -46,7 +50,7 @@ label_descriptions = {
     "Tungro": "Penyakit virus yang membuat daun menguning dan pertumbuhan terhambat.",
 }
 
-# === Load Chatbot Pipeline ===
+# === Load Chatbot ===
 chatbot = pipeline(
     "text-generation",
     model="ARusDian/AgroLens-Chatbot",
@@ -55,29 +59,46 @@ chatbot = pipeline(
 
 
 def preprocess_image(image: Image.Image):
-    image = image.resize((224, 224))  # sesuaikan ukuran input model
-    img_array = np.array(image) / 255.0  # normalisasi
-    img_array = np.expand_dims(img_array, axis=0)  # tambahkan batch dimensi
+    image = image.resize((224, 224))
+    img_array = np.array(image) / 255.0
+    img_array = np.expand_dims(img_array, axis=0)
     return img_array
 
 
-# === Endpoint: Klasifikasi Gambar ===
+def hash_image(image: Image.Image) -> str:
+    """Generate hash for image content."""
+    image_bytes = image.tobytes()
+    return hashlib.md5(image_bytes).hexdigest()
+
+
+# === Endpoint: Prediksi Gambar ===
 @app.post("/predict-image")
 async def predict_image(file: UploadFile = File(...)):
     image = Image.open(file.file).convert("RGB")
-    input_tensor = preprocess_image(image)
+    image_hash = hash_image(image)
+
+    if image_hash in cache:
+        return cache[image_hash]
 
+    input_tensor = preprocess_image(image)
     pred = np.argmax(image_model.predict(input_tensor), axis=1)[0]
     label = label_map.get(pred, "Tidak dikenal")
     description = label_descriptions.get(label, "-")
 
-    return {"prediction": label, "description": description}
+    result = {"prediction": label, "description": description}
+    cache[image_hash] = result
+    return result
 
 
-# === Endpoint: Chatbot Deskripsi Penyakit ===
+# === Endpoint: Chatbot ===
 @app.post("/chatbot")
 async def describe(prompt: dict):
     text = prompt["prompt"]
+    key = hashlib.md5(text.encode()).hexdigest()
+
+    if key in cache:
+        return {"response": cache[key]}
+
     tokenizer = AutoTokenizer.from_pretrained("ARusDian/AgroLens-Chatbot")
     result = chatbot(
         text,
@@ -89,9 +110,11 @@ async def describe(prompt: dict):
         do_sample=True,
         eos_token_id=tokenizer.eos_token_id,
     )[0]["generated_text"]
+
+    cache[key] = result
     return {"response": result}
 
 
-# === Run (opsional) ===
+# === Run lokal (tidak dipakai di Spaces) ===
 if __name__ == "__main__":
     uvicorn.run("app:app", host="0.0.0.0", port=8000)