Spaces:
Runtime error
Runtime error
File size: 7,695 Bytes
aa3fc00 51bece2 aa3fc00 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 | from __future__ import annotations
from functools import lru_cache
from pathlib import Path
from typing import Any
import cv2
import numpy as np
import skfuzzy as fuzz
import tensorflow as tf
from skfuzzy import control as ctrl
IMAGE_SIZE = (256, 256)
LABEL_TOKENS = ("neg", "neutral", "pos")
LABEL_MAP = {label: index for index, label in enumerate(LABEL_TOKENS)}
LABEL_NAMES = {
LABEL_MAP["neg"]: "Negative",
LABEL_MAP["neutral"]: "Neutral",
LABEL_MAP["pos"]: "Positive",
}
NUM_CLASSES = len(LABEL_TOKENS)
MODEL_PATH = (
Path(__file__).resolve().parent / "artifacts" / "cnn_fis_classifier.keras"
)
def get_model_path() -> Path:
return MODEL_PATH.resolve()
@lru_cache(maxsize=1)
def load_model():
model_path = get_model_path()
if not model_path.exists():
raise FileNotFoundError(
f"Keras model tidak ditemukan di: {model_path}. "
"Pastikan file artefak ada di folder ai_engine/artifacts."
)
return tf.keras.models.load_model(model_path)
def preprocess_image_bytes(image_bytes: bytes) -> np.ndarray:
image_array = np.frombuffer(image_bytes, dtype=np.uint8)
image = cv2.imdecode(image_array, cv2.IMREAD_COLOR)
if image is None:
raise ValueError("Gambar tidak bisa dibaca. Pastikan file upload adalah gambar yang valid.")
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image = cv2.resize(image, IMAGE_SIZE, interpolation=cv2.INTER_AREA)
image = image.astype(np.float32) / 255.0
return image
def normalize_probability_vector(model_output: np.ndarray) -> np.ndarray:
probabilities = np.asarray(model_output, dtype=np.float32).reshape(-1)
if probabilities.shape[0] == NUM_CLASSES:
return probabilities
raise ValueError(
f"Expected {NUM_CLASSES} class probabilities for "
f"{', '.join(LABEL_TOKENS)}, received shape {probabilities.shape}."
)
def probability_dict_from_vector(probability_vector: np.ndarray) -> dict[str, float]:
probabilities = normalize_probability_vector(probability_vector)
if probabilities.shape[0] != NUM_CLASSES:
raise ValueError(
f"Expected {NUM_CLASSES} class probabilities, received shape {probabilities.shape}."
)
return {
"negative": float(probabilities[LABEL_MAP["neg"]]),
"neutral": float(probabilities[LABEL_MAP["neutral"]]),
"positive": float(probabilities[LABEL_MAP["pos"]]),
}
def build_fuzzy_system() -> ctrl.ControlSystemSimulation:
positive_confidence = ctrl.Antecedent(np.linspace(0, 1, 101), "positive_confidence")
neutral_confidence = ctrl.Antecedent(np.linspace(0, 1, 101), "neutral_confidence")
negative_confidence = ctrl.Antecedent(np.linspace(0, 1, 101), "negative_confidence")
final_decision = ctrl.Consequent(np.linspace(0, 100, 101), "final_decision")
for variable in (positive_confidence, neutral_confidence, negative_confidence):
variable["low"] = fuzz.trimf(variable.universe, [0.0, 0.0, 0.45])
variable["medium"] = fuzz.trimf(variable.universe, [0.2, 0.5, 0.8])
variable["high"] = fuzz.trimf(variable.universe, [0.55, 1.0, 1.0])
final_decision["negative"] = fuzz.trimf(final_decision.universe, [0, 0, 35])
final_decision["neutral"] = fuzz.trimf(final_decision.universe, [30, 50, 70])
final_decision["positive"] = fuzz.trimf(final_decision.universe, [65, 100, 100])
rules = [
ctrl.Rule(
positive_confidence["high"]
& neutral_confidence["low"]
& negative_confidence["low"],
final_decision["positive"],
),
ctrl.Rule(
positive_confidence["high"]
& neutral_confidence["medium"]
& negative_confidence["low"],
final_decision["positive"],
),
ctrl.Rule(
positive_confidence["medium"]
& neutral_confidence["low"]
& negative_confidence["low"],
final_decision["positive"],
),
ctrl.Rule(
negative_confidence["high"]
& positive_confidence["low"]
& neutral_confidence["low"],
final_decision["negative"],
),
ctrl.Rule(
negative_confidence["high"]
& positive_confidence["low"]
& neutral_confidence["medium"],
final_decision["negative"],
),
ctrl.Rule(
negative_confidence["medium"]
& positive_confidence["low"]
& neutral_confidence["low"],
final_decision["negative"],
),
ctrl.Rule(
neutral_confidence["high"]
& positive_confidence["low"]
& negative_confidence["low"],
final_decision["neutral"],
),
ctrl.Rule(
neutral_confidence["high"]
& positive_confidence["medium"]
& negative_confidence["low"],
final_decision["neutral"],
),
ctrl.Rule(
neutral_confidence["high"]
& negative_confidence["medium"]
& positive_confidence["low"],
final_decision["neutral"],
),
ctrl.Rule(
positive_confidence["medium"] & negative_confidence["medium"],
final_decision["neutral"],
),
ctrl.Rule(
positive_confidence["medium"] & neutral_confidence["medium"],
final_decision["neutral"],
),
ctrl.Rule(
negative_confidence["medium"] & neutral_confidence["medium"],
final_decision["neutral"],
),
ctrl.Rule(
positive_confidence["high"] & negative_confidence["high"],
final_decision["neutral"],
),
ctrl.Rule(
positive_confidence["low"]
& negative_confidence["low"]
& neutral_confidence["low"],
final_decision["neutral"],
),
]
control_system = ctrl.ControlSystem(rules)
return ctrl.ControlSystemSimulation(control_system)
def fuzzy_label_from_probabilities(probability_vector: np.ndarray) -> tuple[str, float]:
probabilities = probability_dict_from_vector(probability_vector)
fuzzy_system = build_fuzzy_system()
fuzzy_system.input["positive_confidence"] = probabilities["positive"]
fuzzy_system.input["neutral_confidence"] = probabilities["neutral"]
fuzzy_system.input["negative_confidence"] = probabilities["negative"]
fuzzy_system.compute()
decision_score = float(fuzzy_system.output["final_decision"])
if decision_score >= 65:
label = "Positive"
elif decision_score <= 35:
label = "Negative"
else:
label = "Neutral"
return label, decision_score
def predict_image_bytes(image_bytes: bytes) -> dict[str, Any]:
image = preprocess_image_bytes(image_bytes)
model = load_model()
batch = np.expand_dims(image, axis=0).astype(np.float32)
model_output = np.asarray(model.predict(batch, verbose=0)[0], dtype=np.float32)
probability_vector = normalize_probability_vector(model_output)
probabilities = probability_dict_from_vector(probability_vector)
fuzzy_label, fuzzy_score = fuzzy_label_from_probabilities(probability_vector)
cnn_prediction_name = LABEL_NAMES[int(np.argmax(probability_vector))]
return {
"label": fuzzy_label,
"fis_score": fuzzy_score,
"cnn_prediction_name": cnn_prediction_name,
"probabilities": probabilities,
}
|