phil8192

# desired matrix:
# 1 1
# 1 2
# 1 3
# 2 1
# 2 2
# 2 3
# 2 4
# 3 1
# 3 2

phil8192

def sample(x):
    """
          1
      2       3
    4   5   6   7
    =
    4 2 6 1 3 5 7
    >>> list(sample(['a', 'b', 'c', 'd', 'e', 'f', 'g']))
    [('d', 1), ('b', 2), ('f', 2), ('a', 3), ('c', 3), ('e', 3), ('g', 3)]
    """

phil8192

# 1 line
function(x) sum(head(tail(pmin(cummax(x),rev(cummax(rev(x))))-x,-1),-1))

# or with fancy ascii visualisation!
#
# > water(x)
# $water
# [1] 10
#
# $vis

phil8192

wapply <- function(x, window=1:10, fun=mean) {
    res <- rep(NA, length(x))
    names(res) <- names(x)
    
    i <- max(window) + 1
    
    while(i <= length(x)) { 
    
        res[i] <- fun(x[i - window])
        i <- i + 1

phil8192

# convert
# 7, 7, 6, 5, 5, 3, 3, 2, 0, 0 ->
# 1  1  2  3  3  4  4  5  NA NA
x <- c(7,7,6,5,5,3,3,2,0,0)

ifelse(x,cumsum(c(1,abs(sign(diff(x))))),NA)

# [1]  1  1  2  3  3  4  4  5 NA NA

phil8192

library(homomorpheR)

key.pair <- PaillierKeyPair$new(modulusBits=1024)
encrypt <- function(x) key.pair$pubkey$encrypt(x)
decrypt <- decrypt <- function(x) key.pair$getPrivateKey()$decrypt(x)
"%+%" <- function(a, b) key.pair$pubkey$add(a, b)
20 == decrypt(encrypt(1) %+% encrypt(19)) # TRUE

phil8192

def set_weights(intercept, coef, classes, model=linear_model.SGDClassifier()):
    model.intercept_ = intercept
    model.coef_ = coef
    model.classes_ = classes
    return model

phil8192

def batches(X, y, batch_size, randomise):
    rows = X.shape[0]
    i = np.arange(rows)
    if randomise:
        np.random.shuffle(i)
    splits = rows / batch_size
    for x_batch, y_batch in zip(np.array_split(X[i, ], splits),
                                np.array_split(y[i], splits)):
        yield x_batch, y_batch

phil8192

def batch_update(model, X, y, classes, batch_size, randomise):
    if batch_size is None:
        batch_size = X.shape[0]
    for x_batch, y_batch in batches(X, y, batch_size, randomise):
        model.partial_fit(x_batch, y_batch, classes)
    return model.intercept_, model.coef_

phil8192

def batch_train(model, X, y, classes, epochs, batch_size, randomise):
    for _ in range(0, epochs):
        batch_update(model, X, y, classes, batch_size, randomise)