# ==========================================
# 练习1：图像矩阵操作 (3×3灰度图像)
# ==========================================
import numpy as np

# 定义3×3灰度图像
print("===== 练习1开始 =====")
image = np.array([
    [100, 150, 200],
    [80, 120, 180],
    [60, 90, 140]
], dtype=np.uint8)

print("原图: ")
print(image)

# 1. 变暗20：每个像素值减20
image_dark = image - 20
print("\n变暗20后: ")
print(image_dark)

# 2. 裁剪左上角：保留2×2区域
image_crop = image[0:2, 0:2]
print("\n裁剪左上角后: ")
print(image_crop)

# 3. 水平翻转：使用np.fliplr()
image_flip = np.fliplr(image)
print("\n水平翻转后: ")
print(image_flip)
print("===== 练习1结束 =====\n")


# ==========================================
# 练习2：看图写代码 (4×4图像矩阵)
# ==========================================
print("===== 练习2开始 =====")
# 定义4×4图像矩阵
img = np.array([
    [255, 255, 0, 0],
    [255, 255, 0, 0],
    [0, 0, 255, 255],
    [0, 0, 255, 255]
], dtype=np.uint8)

print("原始4x4图像:")
print(img)

# 1. 统计白色(255)和黑色(0)像素数量
white_count = np.sum(img == 255)
black_count = np.sum(img == 0)
print(f"\n白色像素(255)数量: {white_count}")
print(f"黑色像素(0)数量: {black_count}")

# 2. 水平翻转并打印
img_flipped = np.fliplr(img)
print("\n水平翻转后的图像:")
print(img_flipped)

# 3. 逆时针旋转90度 (转置 + 上下翻转)
# 方法：先转置(.T)，再上下翻转(np.flipud)
img_rotated = np.flipud(img.T)
print("\n逆时针旋转90度后的图像:")
print(img_rotated)
print("===== 练习2结束 =====\n")


# ==========================================
# 练习3：特征向量计算
# ==========================================
print("===== 练习3开始 =====")
# 定义两个特征图
feature_map1 = np.array([[1, 0, 1], [0, 1, 0], [1, 0, 1]])
feature_map2 = np.array([[1, 1, 1], [1, 0, 0], [1, 0, 0]])

# 展平为向量
vector1 = feature_map1.flatten()
vector2 = feature_map2.flatten()

print("feature_map1 展平向量 vector1: ", vector1)
print("feature_map2 展平向量 vector2: ", vector2)

# 1. 计算欧几里得距离
# 公式: sqrt(sum((a-b)^2))
euclidean_dist = np.linalg.norm(vector1 - vector2)
print(f"\nvector1 和 vector2 的欧几里得距离: {euclidean_dist:.4f}")

# 2. 计算余弦相似度
# 公式: dot(a,b) / (||a|| * ||b||)
cos_sim = np.dot(vector1, vector2) / (np.linalg.norm(vector1) * np.linalg.norm(vector2))
print(f"vector1 和 vector2 的余弦相似度: {cos_sim:.4f}")
print("===== 练习3结束 =====\n")


# ==========================================
# 练习4：文本向量化 (预习挑战)
# ==========================================
print("===== 练习4开始 =====")
# 原始词汇表
vocab = ["Python", "学习", "数据", "人工智能", "编程"]

# 两句话
doc1 = "Python学习编程"
doc2 = "Python人工智能数据"

# 定义向量化函数
def text_to_vector(text, vocab):
    words = text.split()
    vector = np.zeros(len(vocab))
    for i, word in enumerate(vocab):
        vector[i] = words.count(word)
    return vector

# 生成向量
v1 = text_to_vector(doc1, vocab)
v2 = text_to_vector(doc2, vocab)

print("doc1向量: ", v1)
print("doc2向量: ", v2)

# 2. 计算余弦相似度
cos_sim_text = np.dot(v1, v2) / (np.linalg.norm(v1) * np.linalg.norm(v2))
print(f"\ndoc1 和 doc2 的余弦相似度: {cos_sim_text:.4f}")

# 3. 修改vocab，加入"机器"，测试doc3 = "机器学习"
print("\n--- 修改词汇表并测试 doc3 ---")
new_vocab = ["Python", "学习", "数据", "人工智能", "编程", "机器"]
doc3 = "机器学习"
v3 = text_to_vector(doc3, new_vocab)
print(f"新词汇表: {new_vocab}")
print(f"doc3('{doc3}')的向量: ", v3)
print("===== 练习4结束 =====")