FrankPortman

def closure(max_depth):
	def f(depth):
		if depth >= max_depth:
			return depth
		else:
			return f(depth + 1)
	return f

class CLASS(object):
	def __init__(self, max_depth):

FrankPortman

;; Project euler 1, sum of all multiples of 3 or 5 under 1000
(reduce + 
    (filter 
        (fn [x] 
            (or (zero? (mod x 3))
                (zero? (mod x 5))))
        (take 1000 (range))))

FrankPortman

compose_right_to_left <- function(funs) function(x) Reduce(function(y, f) f(y), funs, x)

some_composition <- compose_right_to_left(list(cumsum, sqrt, log))

# cumsum(sqrt(log(16:35))), these are all vectorized and return length n output 
some_composition(16:35)

# Creates a function that takes in a list of functions and returns a function
# that applies those functions to an input argument 'value'. These functions can `reduce` (sum)
# or apply a potentially vectorized operation (`log`, `sqrt`, etc) in whatever order you want.

FrankPortman

# trying to copy https://pymc-devs.github.io/pymc3/notebooks/getting_started.html

library(reticulate)

pymc <- import("pymc3", convert = FALSE)
py <- import_builtins()
np <- import("numpy")
scipy <- import("scipy")

alpha <- 1L

FrankPortman

library(microbenchmark)
library(Rcpp)

weighted_sample <- function(vec, weights, n) {
  
  cdf <- cumsum(weights / sum(weights))
  rand <- runif(n, 0, 1)
  
  sapply(rand, function(x) vec[min(which(cdf >= x))])
  

FrankPortman

## Benchmark
library(bayesAB)
library(microbenchmark)
library(doMC)

A_binom <- rbinom(100, 1, .5)
B_binom <- rbinom(100, 1, .6)

cores <- 4
registerDoMC(cores)

FrankPortman

1 '1 |n  feat_1:1.0 feat_11:1.0 feat_17:2.0 feat_22:1.0 feat_24:4.0 feat_25:1.0 feat_26:1.0 feat_29:2.0 feat_35:1.0 feat_40:1.0 feat_42:5.0 feat_48:2.0 feat_54:1.0 feat_57:2.0 feat_60:11.0 feat_62:1.0 feat_63:1.0 feat_65:1.0 feat_67:7.0 feat_71:1.0 feat_79:2.0 feat_80:1.0 feat_85:1.0
1 '2 |n  feat_8:1.0 feat_18:2.0 feat_37:1.0 feat_58:1.0 feat_64:1.0 feat_67:1.0 feat_73:2.0 feat_74:1.0 feat_76:1.0 feat_78:1.0
1 '3 |n  feat_8:1.0 feat_17:1.0 feat_33:1.0 feat_48:1.0 feat_55:1.0 feat_67:6.0 feat_70:2.0 feat_81:1.0
1 '4 |n  feat_1:1.0 feat_4:1.0 feat_5:6.0 feat_6:1.0 feat_7:5.0 feat_10:1.0 feat_11:1.0 feat_13:1.0 feat_16:1.0 feat_17:1.0 feat_24:7.0 feat_25:2.0 feat_26:2.0 feat_30:58.0 feat_32:10.0 feat_38:3.0 feat_44:2.0 feat_46:2.0 feat_48:1.0 feat_49:2.0 feat_50:1.0 feat_51:3.0 feat_54:3.0 feat_55:1.0 feat_65:2.0 feat_66:1.0 feat_67:5.0 feat_70:4.0 feat_73:2.0 feat_74:1.0 feat_76:1.0 feat_79:1.0 feat_80:1.0 feat_81:2.0 feat_82:2.0 feat_84:22.0 feat_86:1.0 feat_87:2.0
1 '5 |n  feat_17:4.0 feat_24:1.0 feat_41:1.0

FrankPortman

header read result: 16
DUMP [16]: 20 00 00 00 a4 00 00 00 00 00 00 00 00 00 00 00  | ...............
loading buffer (awaiting 164 bytes)
parsing parameters (buf=0x7fc037800000, len=164)
DUMP [164]: 04 08 00 00 2e 61 72 67 31 00 01 01 0a 94 00 00 21 90 00 00 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 00 08 40 00 00 00 00 00 ...  |.....arg1.......!..........@.......@.......@.......@.......@.......@.......@.......@.......@.......@.......@.......@.......@.....
PAR[0]: 000000a4 (PAR_LEN=8, PAR_TYPE=4, large=no, c=0x7fc037800000, ptr=0x7fc037800004)
PAR[1]: 000000a4 (PAR_LEN=148, PAR_TYPE=10, large=no, c=0x7fc03780000c, ptr=0x7fc037800010)
CMD=00000020, pars=2
>>CMD_set/assignREXP (.arg1, REXP)
decode: type=33, len=144

FrankPortman

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;

namespace sumApplication
{
    class Program
    {