from sklearn.datasets import fetch_olivetti_faces

import matplotlib.pyplot as plt

import numpy as np

from sklearn.decomposition import PCA

faces_all = fetch_olivetti_faces()

K = 7  # 7번 인물의 사진만 선택

faces = faces_all.images[faces_all.target == K]

# faces.shape = (10, 64, 64) 인 3차원 데이터

# 이 3차원 데이터를 2차원으로 pca 해서 주성분 분석하기

# 7번 인물의 olivetti_faces를 plot하는 코드

################################################################

# N = 2

# M = 5

# fig = plt.figure(figsize=(10, 5))

# plt.subplots_adjust(top=1, bottom=0, hspace=0, wspace=0.05)

# for i in range(N):

#     for j in range(M):

#         k = i * M + j

#         ax = fig.add_subplot(N, M, k+1)

#         ax.imshow(faces[k], cmap=plt.cm.bone)

#         ax.grid(False)

#         ax.xaxis.set_ticks([])

#         ax.yaxis.set_ticks([])

# plt.suptitle("olivetti_faces of 7th person")

# plt.tight_layout()

# plt.show()

################################################################

pca3 = PCA(n_components=2)

# .images는 64 by 64, .data는 4096로 vectorized form으로 불러온다.

X3 = faces_all.data[faces_all.target == K]

W3 = pca3.fit_transform(X3)

X32 = pca3.inverse_transform(W3)

# pca version - olivetti_faces of 7th person

################################################################

# N = 2

# M = 5

# fig = plt.figure(figsize=(10, 5))

# plt.subplots_adjust(top=1, bottom=0, hspace=0, wspace=0.05)

# for i in range(N):

#     for j in range(M):

#         k = i * M + j

#         ax = fig.add_subplot(N, M, k+1)

#         ax.imshow(X32[k].reshape(64, 64), cmap=plt.cm.bone)

#         ax.grid(False)

#         ax.xaxis.set_ticks([])

#         ax.yaxis.set_ticks([])

# plt.suptitle("pca version - olivetti_faces of 7th person")

# plt.tight_layout()

# plt.show()

################################################################

# 평균과 주성분1, 2 계산

face_mean = pca3.mean_.reshape(64, 64)

face_p1 = pca3.components_[0].reshape(64, 64)

face_p2 = pca3.components_[1].reshape(64, 64)

# 평균 이미지에 주성분1 이미지의 비율을 달리하며 더해서 플롯하는 코드

################################################################

# N = 2

# M = 5

# fig = plt.figure(figsize=(10, 6))

# plt.subplots_adjust(top=1, bottom=0, hspace=0, wspace=0.05)

# for i in range(N):

#     for j in range(M):

#         k = i * M + j

#         ax = fig.add_subplot(N, M, k+1)

#         w = 1.5 * (k - 5) if k < 5 else 1.5 * (k - 4)

#         ax.imshow(face_mean + w * face_p1, cmap=plt.cm.bone)

#         ax.grid(False)

#         ax.xaxis.set_ticks([])

#         ax.yaxis.set_ticks([])

#         plt.title("ratio of pc1={}".format(w))

# plt.suptitle("mean face + principal comp1")

# plt.tight_layout()

# plt.show()

################################################################

# 평균 이미지에 주성분2 이미지의 비율을 달리하며 더해서 플롯하는 코드

################################################################

# N = 2

# M = 5

# fig = plt.figure(figsize=(10, 6))

# plt.subplots_adjust(top=1, bottom=0, hspace=0, wspace=0.05)

# for i in range(N):

#     for j in range(M):

#         k = i * M + j

#         ax = fig.add_subplot(N, M, k+1)

#         w = 1.5 * (k - 5) if k < 5 else 1.5 * (k - 4)

#         ax.imshow(face_mean + w * face_p2, cmap=plt.cm.bone)

#         ax.grid(False)

#         ax.xaxis.set_ticks([])

#         ax.yaxis.set_ticks([])

#         plt.title("ratio of pc2={:.1f}".format(w))

# plt.suptitle("mean face + principal comp2")

# plt.tight_layout()

# plt.show()

################################################################