本文目录
- python 井字棋 ALPHA-BETA剪枝算法和暴力算法 具体代码
- Python-openpyxl教程11 - 注释和样式
- Python中如何从键盘中输入字符串，统计字母，数字，符号和空格的个数
- python3的画直方图的程序hist中的alpha参数是什么意思啊
- python图像灰度化怎么处理alpha值_python将图像转化为灰度图像
- 关于python pygame convert()以及convert_alpha()函数怎么理解..
- python 有Nemenyi 函数吗
- python 岭回归
python 井字棋 ALPHA-BETA剪枝算法和暴力算法 具体代码
#!/usr/bin/env python’’’Tic tac toe in python, Minimax with alpha-beta pruning.’’’import sysimport randomimport getopt# Board: array of 9 int, positionally numbered like this:# 0 1 2# 3 4 5# 6 7 8# Well-known board positionsWINNING_TRIADS = ((0, 1, 2), (3, 4, 5), (6, 7, 8), (0, 3, 6), (1, 4, 7), (2, 5, 8), (0, 4, 8), (2, 4, 6))PRINTING_TRIADS = ((0, 1, 2), (3, 4, 5), (6, 7, 8))# The order in which slots get checked for absence of a player’s token:SLOTS = (0, 1, 2, 3, 4, 5, 6, 7, 8)# Internal-use values. Chosen so that the "winner" of a finished# game has an appropriate value, as X minimizes and O maximizes# the board’s value (function board_valuation() defines "value")# Internally, the computer always plays Os, even though the markers# array can change based on -r command line flag.X_token = -1Open_token = 0O_token = 1# Strings for output: player’s markers, phrase for end-of-gameMARKERS = END_PHRASE = (’draw’, ’win’, ’loss’)HUMAN = 1COMPUTER = 0def board_valuation(board, player, next_player, alpha, beta): ’’’Dynamic and static evaluation of board position.’’’ # Static evaluation - value for next_player wnnr = winner(board) if wnnr != Open_token: # Not a draw or a move left: someone won return wnnr elif not legal_move_left(board): # Draw - no moves left return 0 # Cat # If flow-of-control gets here, no winner yet, not a draw. # Check all legal moves for "player" for move in SLOTS: if board == Open_token: board = player val = board_valuation(board, next_player, player, alpha, beta) board = Open_token if player == O_token: # Maximizing player if val 》 alpha: alpha = val if alpha 》= beta: return beta else: # X_token player, minimizing if val 《 beta: beta = val if beta 《= alpha: return alpha if player == O_token: retval = alpha else: retval = beta return retvaldef print_board(board): ’’’Print the board in human-readable format. Called with current board (array of 9 ints). ’’’ for row in PRINTING_TRIADS: for hole in row: print MARKERS, printdef legal_move_left(board): ’’’ Returns True if a legal move remains, False if not. ’’’ for slot in SLOTS: if board == Open_token: return True return Falsedef winner(board): ’’’ Returns -1 if X wins, 1 if O wins, 0 for a cat game, 0 for an unfinished game. Returns the first "win" it finds, so check after each move. Note that clever choices of X_token, O_token, Open_token make this work better. ’’’ for triad in WINNING_TRIADS: triad_sum = board if triad_sum == 3 or triad_sum == -3: return board # Take advantage of "_token" values return 0def determine_move(board): ’’’ Determine Os next move. Check that a legal move remains before calling. Randomly picks a single move from any group of moves with the same value. ’’’ best_val = -2 # 1 less than min of O_token, X_token my_moves = for move in SLOTS: if board == Open_token: board = O_token val = board_valuation(board, X_token, O_token, -2, 2) board = Open_token print "My move", move, "causes a", END_PHRASE if val 》 best_val: best_val = val my_moves = if val == best_val: my_moves.append(move) return random.choice(my_moves)def recv_human_move(board): ’’’ Encapsulate human’s input reception and validation. Call with current board configuration. Returns an int of value 0..8, the Human’s move. ’’’ looping = True while looping: try: inp = input("Your move: ") yrmv = int(inp) if 0 《= yrmv 《= 8: if board == Open_token: looping = False else: print "Spot already filled." else: print "Bad move, no donut." except EOFError: print sys.exit(0) except NameError: print "Not 0-9, try again." except SyntaxError: print "Not 0-9, try again." if looping: print_board(board) return yrmvdef usage(progname): ’’’Call with name of program, to explain its usage.’’’ print progname + ": Tic Tac Toe in python" print "Usage:", progname, "" print "Flags:" print "-x, -X: print this usage message, then exit." print "-h: human goes first (default)" print "-c: computer goes first" print "-r: computer is X, human is O" print "The computer O and the human plays X by default."def main(): ’’’Call without arguments from __main__ context.’’’ try: opts, args = getopt.getopt(sys.argv, "chrxX", ) except getopt.GetoptError: # print help information and exit: usage(sys.argv) sys.exit(2) next_move = HUMAN # Human goes first by default for opt, arg in opts: if opt == "-h": next_move = HUMAN if opt == "-c": next_move = COMPUTER if opt == "-r": MARKERS = ’O’ MARKERS = ’X’ if opt in ("-x", "-X", "--help"): usage(sys.argv) sys.exit(1) # Initial state of board: all open spots. board = # State machine to decide who goes next, and when the game ends. # This allows letting computer or human go first. while legal_move_left(board) and winner(board) == Open_token: print print_board(board) if next_move == HUMAN and legal_move_left(board): humanmv = recv_human_move(board) board = X_token next_move = COMPUTER if next_move == COMPUTER and legal_move_left(board): mymv = determine_move(board) print "I choose", mymv board = O_token next_move = HUMAN print_board(board) # Final board state/winner and congratulatory output. try: # "You won" should never appear on output: the program # should always at least draw. print except IndexError: print "Really bad error, winner is", winner(board) sys.exit(0)#-------if __name__ == ’__main__’: try: main() except KeyboardInterrupt: print sys.exit(1)
Python-openpyxl教程11 - 注释和样式
注释具有text属性和author属性，必须同时设置它们。
加载时工作薄中存在的注释会自动存储在其相应单元格的注释属性中。格式信息(如字体大小，粗体和斜体)以及注释的容器框的原始尺寸和位置都将丢失。
保存工作薄时保留在工作薄中的注释会自动保存到工作薄文件中
注释尺寸可以指定为只写。评论尺寸以像素为单位。
如果需要, openpyxl.utils.units 包含用于从其他度量单位(例如mm或点)转换为像素的辅助函数:
样式用于屏幕上显示时更改数据的外观。他们还用于确定数字的格式。
样式可以应用于以下方面： - 用于设置字体大小，颜色，下划线等的字体 - 填充以设置图案或颜色渐变 - border可以设置单元格的边框 - 单元格对齐 - 保护
以下是默认值：
有两种类型的样式：单元样式和命名样式，也成为样式模板
单元格样式在对象之间共享，并且一旦分配了它们就无法更改。这样可以避免不必要的副作用，例如，仅更改一个单元格时就可以更改许多单元格的样式。
样式也可以复制
字体，背景，边框等的颜色都可以通过三种方式设置：索引，aRGB或主题。 索引颜色是旧版实现，颜色本身取决于工作薄或应用程序默认提供的索引。主题颜色可用于互补色，但也取决于工作薄中存在的主题，因此，建议使用RGB颜色。
RGB颜色使用红色，绿色和蓝色的十六进制值设置
理论上，alpha值是指颜色的透明度，但这与单元格样式无关。默认值00将加在任何简单的RGB值之前:
还支持传统索引颜色以及主题和色彩。
索引64和65不能设置，并且分别留给系统前景色和背景色
样式直接应用于单元格
样式也可以应用于行和列，但是请注意，这仅适用于关闭文件后再Excel中创建的单元格。如果要将样式应用于整个行和列，则必须自己将样式应用于每个单元格。这是文件格式的限制：
合并的单元格的行为与其他单元格对象相似。其值和格式在其左上角的单元格中定义。为了更改整个合并单元格的边框。请更改其左上角单元格的边框。格式化是出于编写目的而生成的。
与单元格样式相反，命名样式是可变的。当您想一次将格式应用于许多不同的单元格时，它们很有意义。注意: 将命名样式分配给单元格之后，对样式的其他更改将不会影响该单元格。
一旦将命名样式注册到工作薄中，就可以简单的通过名称来引用它。
创建命名样式后，可以将其注册到工作薄中： wb.add_name_style(highlight)
但是，命名样式在首次分配给单元时也将自动注册： ws.style = highlight
注册后，仅使用名称分配样式： ws.style = ’highlight’
该规范包括一些内置样式，也可以使用，不幸的是，这些样式的名称以本地化形式存储。 openpyxl仅会识别英文名称，并且只能与此处的文字完全一样。如下：
Number formats
Informative
Text Styles
Comparisons
Highlights
Python中如何从键盘中输入字符串，统计字母，数字，符号和空格的个数
初学者的话确实可以通过 ascii table 来判断字母和数字的区别。Python 里面有两个内置函数 ord 和 chr 可用。
ascii table
通过判断其字的范围来确定是字母还是别的。确实可以达到你现在想要的目的。
不过我个人建议是，暂时不用太在意这种问题。这个也不是解决这些问题的理想方法。
当你以后学会使用正则表达式之后，回头看这个问题就十分简单了！
正则表达式才是解决字符串这类问题更理想的方法。这种方式的优势和能力之强等你学习到后就知道了。
python3的画直方图的程序hist中的alpha参数是什么意思啊
理解为填充颜色的深度，你把alpha设置成0.99和0.01，看看画出来的结果就知道了
python图像灰度化怎么处理alpha值_python将图像转化为灰度图像
能不能把PIL（PythonImageLibrary)也弄进来，那就简单了。
PIL中可以把图像模式convert，例如转成“L”模式，8位，就变成灰色了。
关于python pygame convert()以及convert_alpha()函数怎么理解..
convert函数作用是将图片转化为Surface对象，pygame现在会自动这么做，不用你去写；convert_alpha相对于convert，保留了图像的Alpha 通道信息，这个你可以去百度一下，可以认为是保留了透明的部分，这样图片才可以是不规则的。
python 有Nemenyi 函数吗
没有，网上有人写好的，可以当成一个库来使用：import numpy as np from scipy import stats from itertools import combinations from statsmodels.stats.multitest import multipletests from statsmodels.stats.libqsturng import psturng import warnings def kw_nemenyi(groups, to_compare=None, alpha=0.05): """ Kruskal-Wallis 1-way ANOVA with Nemenyi’s multiple comparison test Arguments: --------------- groups: sequence arrays corresponding to k mutually independent samples from continuous populations to_compare: sequence tuples specifying the indices of pairs of groups to compare, e.g. would compare group 0 with 1 & 2. by default, all possible pairwise comparisons between groups are performed. alpha: float family-wise error rate used for correcting for multiple comparisons (see statsmodels.stats.multitest.multipletests for details) Returns: --------------- H: float Kruskal-Wallis H-statistic p_omnibus: float p-value corresponding to the global null hypothesis that the medians of the groups are all equal Z_pairs: float array Z-scores computed for the absolute difference in mean ranks for each pairwise comparison p_corrected: float array corrected p-values for each pairwise comparison, corresponding to the null hypothesis that the pair of groups has equal medians. note that these are only meaningful if the global null hypothesis is rejected. reject: bool array True for pairs where the null hypothesis can be rejected for the given alpha Reference: --------------- """ # omnibus test (K-W ANOVA) # ------------------------------------------------------------------------- groups = k = len(groups) n = np.array() if np.any(n 《 5): warnings.warn("Sample sizes 《 5 are not recommended (K-W test assumes " "a chi square distribution)") allgroups = np.concatenate(groups) N = len(allgroups) ranked = stats.rankdata(allgroups) # correction factor for ties T = stats.tiecorrect(ranked) if T == 0: raise ValueError(’All numbers are identical in kruskal’) # sum of ranks for each group j = np.insert(np.cumsum(n), 0, 0) R = np.empty(k, dtype=np.float) for ii in range(k): R.sum() # the Kruskal-Wallis H-statistic H = (12. / (N * (N + 1.))) * ((R ** 2.) / n).sum() - 3 * (N + 1) # apply correction factor for ties H /= T df_omnibus = k - 1 p_omnibus = stats.chisqprob(H, df_omnibus) # multiple comparisons # ------------------------------------------------------------------------- # by default we compare every possible pair of groups if to_compare is None: to_compare = tuple(combinations(range(k), 2)) ncomp = len(to_compare) Z_pairs = np.empty(ncomp, dtype=np.float) p_uncorrected = np.empty(ncomp, dtype=np.float) Rmean = R / n for pp, (ii, jj) in enumerate(to_compare): # standardized score Zij = (np.abs(Rmean) / np.sqrt((1. / 12.) * N * (N + 1) * (1. / n))) Z_pairs = Zij # corresponding p-values obtained from the upper quantiles of the # studentized range distribution p_corrected = psturng(Z_pairs * np.sqrt(2), ncomp, np.inf) reject = p_corrected 《= alpha return H, p_omnibus, Z_pairs, p_corrected, reject
python 岭回归
所求参数是alpha的函数，比如记为f(alpha), f(alpha)随alpha的改变的轨迹就是岭迹。实际计算中可选非常多的alpha值，做出一个岭迹图，看看这个图在取哪个值的时候变稳定了，那就确定alpha值了，从而确定参数。Ridge(alpha=1.0， fit_intercept=False)model.fit(x,y)这样就等于你算的，因为你numpy是用增广矩阵算的，所以应该将set fit_intercept=Falsemodel.coef_array()
