from sklearn.datasets import load_digits
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.neural_network import MLPClassifier
from sklearn.metrics import ConfusionMatrixDisplay, confusion_matrix, davies_bouldin_score
from sklearn.cluster import KMeans
import matplotlib.pyplot as plt

# 1. feladat

df = load_digits(as_frame=True)

print('rekordok:', df.data.shape[0])
print('attributumok:', df.data.shape[1])
print('osztalyok:', len(df.target_names))

# 2. feladat

"""
sns.set(style='ticks')
sns.pairplot(df.frame)
plt.show()
"""

# 3. feladat

X_train, X_test, y_train, y_test = train_test_split(
    df.data, df.target, test_size=0.2, shuffle=True, random_state=2025)

# 4. feladat

dec_tree = DecisionTreeClassifier(max_depth=4, criterion='entropy')
dec_tree.fit(X_train, y_train)
dec_tree_score = dec_tree.score(X_test, y_test)
print('dec_tree:', dec_tree_score)

logreg = LogisticRegression(solver='liblinear')
logreg.fit(X_train, y_train)
logreg_score = logreg.score(X_test, y_test)
print('logreg:', logreg_score)

neural = MLPClassifier(hidden_layer_sizes=(4), activation='logistic')
neural.fit(X_train, y_train)
neural_score = neural.score(X_test, y_test)
print('neural:', neural_score)

# 5. feladat

prediction = logreg.predict(X_test)
cm = confusion_matrix(y_test, prediction)

# 6. feladat

disp = ConfusionMatrixDisplay(cm)
disp.plot(cmap=plt.cm.Blues)
plt.show()

# 7. feladat

kmeans = KMeans(n_clusters=2)
kmeans.fit(X_train, y_train)

DB = [davies_bouldin_score(X_train, kmeans.labels_)]

K = 3
while K != 31:
    kmeans = KMeans(n_clusters=K)
    kmeans.fit(X_train, y_train)
    DB.append(davies_bouldin_score(X_train, kmeans.labels_))
    K+=1
optimalis_K = DB.index(max(DB))+2
print('klaszterszam:', optimalis_K)

cluster = KMeans(n_clusters=optimalis_K)
cluster.fit(X_train, y_train)

centers = cluster.cluster_centers_;  # centroid of clusters
distX = cluster.transform(df.data);
dist_center = cluster.transform(centers);

# Visualizing of clustering in the distance space where coordinates are the distances from the cluster centers
fig = plt.figure(2);
plt.title('Iris data in the distance space');
plt.xlabel('Cluster 1');
plt.ylabel('Cluster 2');
plt.scatter(distX[:,0],distX[:,1],s=50,c=cluster.labels_);  # data
plt.scatter(dist_center[:,0],dist_center[:,1],s=200,c='red',marker='X');  # centroids
plt.show();