brezniczky

""" Partial's frozen arguments are not 'as frozen' as I would have thought
    at first.
"""
from functools import partial


f = partial(int, base=2)


# results in 9 - great!

brezniczky

"""
Depending on the scenario, lru_cache() in functools may lead to inadvertent or
at least unexpected cache misses.
"""
from functools import lru_cache


eval_count = 0

brezniczky

def mapper(input):
  for each edge node1 -> node2 in input:
    emit (key: node1, node1 -> node2)
    emit (key: node2, node2 <- node1)

def reducer(key: node, edges):
  for each distinct edge node1 -> node2 in edges:
    if node1 <= node2:
      if node2 <- node1 in edges:
        emit(node1, node2)

brezniczky

get.perm = function(n, FUN = print) {
  perm = n:1

  switch_loop = function(L) {
    # evaluate all pair exchanges on the given level (L)
    if (L == 1)
      FUN(perm)
    else {
      switch_pair = function(i1, i2) {
        temp = perm[i1]

brezniczky

# Verification of the is_ordered.R solution.

library(testthat)

test.generate.random.sequences = function() {
  # generate some random sequence
  #
  # first: a random walk with a drift of unit increases or stagnation
  n = 1000
  root.sequence = c(1, 1 + cumsum(sample(c(0, 1), n, replace = TRUE)))

brezniczky

# This code exhibits a vectorized function telling whether a series of numbers
# is possibly sorted according to some (not necessarily increasing or
# decreasing) order.
#
# "Possibly" as in practice a random unsorted series may not contradict being
# sorted - e.g. 1, 3, 3, 2, 2 could be just a random discrete uniform sample.
#
# E.g. (1, 2, 5, 7, 7, 3, 3, 3, 8, 10, 10) is possibly sorted, but
#      (1, 3, 2, 5, 2) cannot be, as 2 occurs in a disjoint way, separated by 5.

brezniczky

# install.packages("animation")

# A sinus plotter in R just to demo animations.

library(animation)

n.frames = 200
max.t = 60
max.i = 60
all.i = 1:max.i

brezniczky

# Here I will try to visualize how an xgboost random forest approximates the
# training set in case of a binary classification as the number of weak
# classifiers in the ensemble increase, and/or as the maximum depth of trees
# change.

library(animation)
library(xgboost)

n.iters = 400
video.speed = 2 # factor, should be an integer

brezniczky

file.size.bytes = 1024 * 256
default.window.width = 256

# symbols in the range [0:(symbol.count - 1)] are expected
symbol.count = 256

set.seed(0)

test.data.bytes =
  sample(x = 0:(symbol.count - 1), replace = TRUE, size = file.size.bytes)

brezniczky

quicker.count = function(data, window.width = default.window.width) {
# ...
  # start and do the counting
  for(i in 2:n.positions) {
    # the value leaving the window is the first item in the
    # previous window, i.e. data[i - 1]
    old.data.idx = x[i - 1] + 1
    frequency[old.data.idx] = frequency[old.data.idx] - 1
    # the value entering the window is the last item in the
    # current window