The Hack-R Blog

Archive for the ‘Programming’ Category

The gutenbergr package is an excellent API wrapper for Project Gutenberg, which provides unlimited free access to public domain books and materials.

Inspired by an analytical philosopher working in DS, I decided to spin up a Shiny app to allow myself and the public to text-mine 1 or more works at a time.

I started a new blog for the project at:

https://hume.xyz/blog

The app itself can be found at https://hume.xyz, however the free ShinyApps.io is a bit slow, so it won’t load quickly until the project gets moved to AWS, Azure, a NameCheap VM or some other affordable production environment.

Please see the initial, and currently only, blog article at hume.xyz/blog for GitHub links to the full source code – and feel free to contribute!

library(data.table)
library(lightgbm)
data(agaricus.train, package = "lightgbm")
train <- agaricus.traindtrain <- lgb.Dataset(train$data, label = train$label, free_raw_data = FALSE)
data(agaricus.test, package = "lightgbm")
test <- agaricus.testdtest <- lgb.Dataset.create.valid(dtrain, test$data, label = test$label)
valids <- list(test = dtest)

grid_search <- expand.grid(Depth = 8,
                           L1 = 0:5,
                           L2 = 0:5)

model <- list()
perf <- numeric(nrow(grid_search))

for (i in 1:nrow(grid_search)) {
  model[[i]] <- lgb.train(list(objective = "regression",
                          metric = "l2",
                          lambda_l1 = grid_search[i, "L1"],
                          lambda_l2 = grid_search[i, "L2"],
                          max_depth = grid_search[i, "Depth"]),
                     dtrain,
                     2,
                     valids,
                     min_data = 1,
                     learning_rate = 1,
                     early_stopping_rounds = 5)
  perf[i] <- min(rbindlist(model[[i]]$record_evals$test$l2))
}

Result:
> cat("Model ", which.min(perf), " is lowest loss: ", min(perf), sep = "")
Model 1 is lowest loss: 1.972152e-31> print(grid_search[which.min(perf), ])
  Depth L1 L21     8  0  0

import sys
import math
 
import numpy as np
from sklearn.grid_search import GridSearchCV
 
sys.path.append('xgboost/wrapper/')import xgboost as xgb
 
 
class XGBoostClassifier():
    def __init__(self, num_boost_round=10, **params):
        self.clf = None
        self.num_boost_round = num_boost_round
        self.params = params
        self.params.update({'objective': 'multi:softprob'})
 
    def fit(self, X, y, num_boost_round=None):
        num_boost_round = num_boost_round or self.num_boost_round
        self.label2num = dict((label, i) for i, label in enumerate(sorted(set(y))))
        dtrain = xgb.DMatrix(X, label=[self.label2num[label] for label in y])
        self.clf = xgb.train(params=self.params, dtrain=dtrain, num_boost_round=num_boost_round)
 
    def predict(self, X):
        num2label = dict((i, label)for label, i in self.label2num.items())
        Y = self.predict_proba(X)
        y = np.argmax(Y, axis=1)
        return np.array([num2label[i] for i in y])
 
    def predict_proba(self, X):
        dtest = xgb.DMatrix(X)
        return self.clf.predict(dtest)
 
    def score(self, X, y):
        Y = self.predict_proba(X)
        return 1 / logloss(y, Y)
 
    def get_params(self, deep=True):
        return self.params
 
    def set_params(self, **params):
        if 'num_boost_round' in params:
            self.num_boost_round = params.pop('num_boost_round')
        if 'objective' in params:
            del params['objective']
        self.params.update(params)
        return self
   
   
def logloss(y_true, Y_pred):
    label2num = dict((name, i) for i, name in enumerate(sorted(set(y_true))))
    return -1 * sum(math.log(y[label2num[label]]) if y[label2num[label]] > 0 else -np.inf for y, label in zip(Y_pred, y_true)) / len(Y_pred)


def main():
    clf = XGBoostClassifier(
        eval_metric = 'auc',
        num_class = 2,
        nthread = 4,
        eta = 0.1,
        num_boost_round = 80,
        max_depth = 12,
        subsample = 0.5,
        colsample_bytree = 1.0,
        silent = 1,
        )
    parameters = {
        'num_boost_round': [100, 250, 500],
        'eta': [0.05, 0.1, 0.3],
        'max_depth': [6, 9, 12],
        'subsample': [0.9, 1.0],
        'colsample_bytree': [0.9, 1.0],
    }
    clf = GridSearchCV(clf, parameters, n_jobs=1, cv=2)
   
    clf.fit([[1,2], [3,4], [2,1], [4,3], [1,0], [4,5]], ['a', 'b', 'a', 'b', 'a', 'b'])
    best_parameters, score, _ = max(clf.grid_scores_, key=lambda x: x[1])
    print(score)
    for param_name in sorted(best_parameters.keys()):
        print("%s: %r" % (param_name, best_parameters[param_name]))
               
    print(clf.predict([[1,2]]))


if __name__ == '__main__':
    main()

Just a little Python script I wrote to make an LED blink on a Raspberry Pi and to print a message to the screen when there’s a #antisec tweet:

# Jason D. Miller
# github.com/hack-r

from twython import TwythonStreamer
import RPi.GPIO as GPIO
import time

C_KEY =  ""
C_SECRET = ""
A_TOKEN = "-"
A_SECRET = ""

GPIO.setmode(GPIO.BCM)
GPIO.setwarnings(False)
GPIO.setup(18,GPIO.OUT)

def blink():
    GPIO.output(18, GPIO.HIGH)
    time.sleep(1)
    GPIO.output(18, GPIO.LOW)

class MyStreamer(TwythonStreamer):
     def on_success(self, data):
         if 'text' in data:
              blink()
              print("Antisec Tweet detected. Call the FBI.")

stream = MyStreamer(C_KEY, C_SECRET, A_TOKEN, A_SECRET)

stream.statuses.filter(track="#antisec")

Has it really been a whole year?

On Pi Day 2015 hack-r.com posted a tribute to Pi (π) Day, published on GitHub, wherein we created fractals in R based on π, scraped and displayed information on Pi and other fun stuff.

This year, find out how Fibonacci numbers, which are sequences of integers, have a freaky relationship with π! View the entire script on GitHub.

# Pi Fibonacci Sequence ---------------------------------------------------
cat("This year, we'll look at the relationship between Pi and Fibonacci sequences. \n")
cat("Until very recently there were just two methods used to compute pi (π),
one invented by the Greek mathematician Archimedes,
and the other by the Scottish mathematician James Gregory. \n")

cat("If we use Sir Gregory's arc tangent method, you'll notice a pattern...")

pi/4
atan(1)

pi/4 == atan(1)

atan(1/3)
atan(1/5)  + atan(1/8)

atan(1/8)
atan(1/13) + atan(1/21)

cat("We can combine what we saw above")
pi/4
atan(1/2) + atan(1/3)
atan(1/2) + atan(1/5) + atan(1/8)

atan(1/21)
atan(1/34) + atan(1/55)

cat("You'll notice that the pattern is a Fibonacci sequence! \n")

cat(" We have just seen that there are infinitely many formulae for π using the Fibonacci numbers!")

cbind.fill <- function(…){
nm <- list(…)
nm <- lapply(nm, as.matrix)
n <- max(sapply(nm, nrow))
do.call(cbind, lapply(nm, function (x)
rbind(x, matrix(, n-nrow(x), ncol(x)))))
}

<div>detach_package <- function(pkg, character.only = FALSE)</div>
<div>{</div>
<div>    if(!character.only)</div>
<div>    {</div>
<div>        pkg <- deparse(substitute(pkg))</div>
<div>    }</div>
<div>    search_item <- paste("package", pkg, sep = ":")</div>
<div>    while(search_item %in% search())</div>
<div>    {</div>
<div>        detach(search_item, unload = TRUE, character.only = TRUE)</div>
<div>    }</div>
<div>}</div>
<div></div>
<div>