Cleoparka

Mein neuer Beitrag Vimmi

2026-06-20T00:00:00+02:00

Die Sonne scheint. Doch das ist nur Schein.

Star Trek: Ryker treffen

2026-05-08T01:00:00+02:00

Wir freuen uns über das Leben in den Sternen.

Ich liebe Pferde wirklich

2025-01-25T03:29:20+01:00

Ich weiß gar nicht mehr, wie das kam, dass Timo mich mit zum Reiten nahm. Freundschaft war vermutlich der primäre Grund, der zweite: Neugier, der dritte: Die glorreichen Sieben. Wenn ich bislang mit Leuten übers Reiten sprach, dann waren es immer unsere zwei Ausflüge zu einem Reiterhof in Logabirum, die mir in den Sinn kamen. Und trotz des leichten Schreckens, den sie mit sich brachten, sitze ich jetzt in einem Bett, das zu einem Reiterhof in Irland gehört. Hier mache ich gerade Working Holidays, die mit Holidays gehörig wenig zu tun haben. Und – ich greife vor – ich sitze hier auf Pferden, die meistens nicht nur in der Landschaft herumstehen, sondern auch darin laufen, schreiten, traben, galoppieren, meist mit mir obendrauf. Davon hätte ich höchstens zu träumen gewagt nach meinen Erlebnissen mit Timo und meinem ersten Leihpferd.

Pferdchen.

Ich war ungefähr neun. Mit dem Fahrrad zum Reiterhof zu fahren, war allein schon aufregend. So weite Strecken selbstständig hinter mich zu bringen, war ich nicht gewohnt. Das waren bestimmt zwölf Kilometer oder so. Oft mit Gegenwind – auf dem Hin- und dem Rückweg.

Ich weiß noch, dass Timo ein super Voltigierer war. Zumindest hat er an Wettbewerben teilgenommen. Und wohl deswegen durfte er sich einfach Pferde für einen Ausritt ins Grüne ausleihen und musste nicht im Kreis durch die Arena marschieren. Es war auch kein Thema, dass ich als Reitunkundiger auf so einem Ausflug mitkam. Bronko hieß mein Pferd, ein Name, der für meinen Geschmack zu sehr nach Gemüse klingt, um eines edlen Rosses würdig zu sein.

Fig 2. bagan kernel.

Zwei Gründe, die dafür sprechen , nach meiner Reiterfahrung mit Bronko nie wieder auf ein Pferd zu steigen:

Kurz nach dem Losreiten, vermutlich beim zweiten Besuch in Logabirum, fiel ich von Bronko herunter. Ich geriet so ins Gras, in knapper Entfernung einer Hausmauer, gegen die ich auch problemlos hätte knallen können. Ich tat das nicht, weiß aber spätestens seitdem, dass Pferderücken nicht nur optisch einen wesentlich höheren Abstand zum Boden aufweisen als mein Fahrradsattel. Und mein Fahrradunfall, der mir mein Untergebiss durcheinander haute, war schon schlimm.
Als Timo mit seinem Pferd in den Trab ging, rannte Bronko hinterher. Mit der Folge, dass ich wie blöde im Sattel hoch- und runterhüpfte und feststellen musste, dass ich keine Kontrolle über das Pferd hatte. Timo stoppte mein Pferd und rettete mich (danke nochmal dafür!) vor dem erneuten Runterplumsen und einem vorzeitigen Verlust der Gradheit meines Untergebisses, das ja erst ein paar Jahre später beim Fahrradunfall seine endgültige Form gewinnen sollte. Seit diesem Vorfall vertrat ich mit festem Glauben die These vom Alphapferd, das die Herde führt, so wie Timos Pferd das meine in den Trab gebracht hat.

Und trotzdem: Ich bin jetzt in Irland, bewohne ein Zimmer mit quietschendem Bett und meiner Freundin Johanna, die darin neben mir liegt und die Ursache für meinen Weg hierhin ist. Vier Wochen war sie schon hier, bevor ich nachkam. Ich nahm den Ryanair-Flieger (schrecklich: diese gelben Sitze) von Weeze nach Shannon, den Bus vom Flughafen nach Limerick Bus Station, von dort den Bus nach Abbeyfeale. Gegenüber dem Spar-Markt (gibt’s die eigentlich nur noch außerhalb von Deutschland? Nee, oder? Ich hab im Rheiderland auch mal wieder einen gesehen, ist aber schon länger her), also: gegenüber vom Spar-Markt bin ich ausgestiegen, dann zum Tesco-Markt gelaufen, dann da Süßigkeiten gekauft, dann bei der Familie Fitzgerald angerufen. Leona ist eine Tochter des Hauses, die mich abholte und erst mal Futtersäcke für die Hühner kaufte und zusammen mit dem Verkäufer in ihren Peugout donnerte, in dem ich saß und darüber staunte, wie selbstverständlich das Volk der Iren auf der falschen Seite fährt und die Gangschaltung mit links bedient.

Und das war erst der Anfang. Wie nur sollten in einem Land, in dem Autofahrer links in Kreisel hineinfahren, die mit „Yield“ beschriftet sind, die Pferde ordnungsgemäß funktionieren? Ich sollte es bald erfahren …

Ich liebe Phasen wirklich

2025-01-22T03:29:20+01:00

Pferdchen.

Fig 2. bagan kernel.

Zwei Gründe, die dafür sprechen , nach meiner Reiterfahrung mit Bronko nie wieder auf ein Pferd zu steigen:

Kurz nach dem Losreiten, vermutlich beim zweiten Besuch in Logabirum, fiel ich von Bronko herunter. Ich geriet so ins Gras, in knapper Entfernung einer Hausmauer, gegen die ich auch problemlos hätte knallen können. Ich tat das nicht, weiß aber spätestens seitdem, dass Pferderücken nicht nur optisch einen wesentlich höheren Abstand zum Boden aufweisen als mein Fahrradsattel. Und mein Fahrradunfall, der mir mein Untergebiss durcheinander haute, war schon schlimm.
Als Timo mit seinem Pferd in den Trab ging, rannte Bronko hinterher. Mit der Folge, dass ich wie blöde im Sattel hoch- und runterhüpfte und feststellen musste, dass ich keine Kontrolle über das Pferd hatte. Timo stoppte mein Pferd und rettete mich (danke nochmal dafür!) vor dem erneuten Runterplumsen und einem vorzeitigen Verlust der Gradheit meines Untergebisses, das ja erst ein paar Jahre später beim Fahrradunfall seine endgültige Form gewinnen sollte. Seit diesem Vorfall vertrat ich mit festem Glauben die These vom Alphapferd, das die Herde führt, so wie Timos Pferd das meine in den Trab gebracht hat.

Was macht die Kunst!

2024-08-29T04:29:20+02:00

Pizza ipsum dolor amet buffalo sauce anchovies mushrooms, NY style peppers melted cheese platter burnt mouth. Steak white garlic green bell peppers bbq sauce thin crust pepperoni string cheese deep crust party bbq. Pan pizza deep crust extra sauce platter burnt mouth peppers. Philly chicken burnt mouth beef chicken and bacon, bbq rib pie mozzarella hand tossed sauteed onions stuffed.

Peppers black olives lasagna, extra sauce white garlic steak pie stuffed beef sausage stuffed crust meat lovers bbq marinara deep crust. Buffalo sauce pineapple bacon chicken and bacon NY style anchovies pork spinach stuffed thin crust philly chicken party. Extra sauce lasagna hawaiian, Chicago style pan white pizza pork pineapple bacon thin crust. Ricotta hand tossed lasagna, stuffed bbq sauce pineapple thin crust fresh tomatoes melted cheese. Pizza ipsum dolor amet buffalo sauce anchovies mushrooms, NY style peppers melted cheese platter burnt mouth. Steak white garlic green bell peppers bbq sauce thin crust pepperoni string cheese deep crust party bbq. Pan pizza deep crust extra sauce platter burnt mouth peppers. Philly chicken burnt mouth beef chicken and bacon, bbq rib pie mozzarella hand tossed sauteed onions stuffed.

Peppers black olives lasagna, extra sauce white garlic steak pie stuffed beef sausage stuffed crust meat lovers bbq marinara deep crust. Buffalo sauce pineapple bacon chicken and bacon NY style anchovies pork spinach stuffed thin crust philly chicken party. Extra sauce lasagna hawaiian, Chicago style pan white pizza pork pineapple bacon thin crust. Ricotta hand tossed lasagna, stuffed bbq sauce pineapple thin crust fresh tomatoes melted cheese.Pizza ipsum dolor amet buffalo sauce anchovies mushrooms, NY style peppers melted cheese platter burnt mouth. Steak white garlic green bell peppers bbq sauce thin crust pepperoni string cheese deep crust party bbq. Pan pizza deep crust extra sauce platter burnt mouth peppers. Philly chicken burnt mouth beef chicken and bacon, bbq rib pie mozzarella hand tossed sauteed onions stuffed.

Maths Example with Mathjax

2021-06-17T00:57:49+02:00

Below is an example of maths using mathjax.

Any page needing maths should start with the frontmatter:

usemathjax: true

\[\begin{align*} y = y(x,t) &= A e^{i\theta} \\ &= A (\cos \theta + i \sin \theta) \\ &= A (\cos(kx - \omega t) + i \sin(kx - \omega t)) \\ &= A\cos(kx - \omega t) + i A\sin(kx - \omega t) \\ &= A\cos \Big(\frac{2\pi}{\lambda}x - \frac{2\pi v}{\lambda} t \Big) + i A\sin \Big(\frac{2\pi}{\lambda}x - \frac{2\pi v}{\lambda} t \Big) \\ &= A\cos \frac{2\pi}{\lambda} (x - v t) + i A\sin \frac{2\pi}{\lambda} (x - v t) \end{align*}\]

Inline maths can be written with the \\( and \\) characters, producing inline maths such as \(\delta(t) \xrightarrow{\mathscr{F}} 1\).

The above is accomplished with thanks to Alan Duan and Zichen Vincent Zhang.

Welcome to Jekyll!

2020-08-29T04:29:20+02:00

You’ll find this post in your _posts directory. Go ahead and edit it and re-build the site to see your changes. You can rebuild the site in many different ways, but the most common way is to run jekyll serve, which launches a web server and auto-regenerates your site when a file is updated.

Jekyll requires blog post files to be named according to the following format:

YEAR-MONTH-DAY-title.MARKUP

Where YEAR is a four-digit number, MONTH and DAY are both two-digit numbers, and MARKUP is the file extension representing the format used in the file. After that, include the necessary front matter. Take a look at the source for this post to get an idea about how it works.

Jekyll also offers powerful support for code snippets:

def print_hi(name)
  puts "Hi, #{name}"
end
print_hi('Tom')
#=> prints 'Hi, Tom' to STDOUT.

Check out the Jekyll docs for more info on how to get the most out of Jekyll. File all bugs/feature requests at Jekyll’s GitHub repo. If you have questions, you can ask them on Jekyll Talk.

Cara ribet memperbarui forked repo

2020-01-28T19:00:00+01:00

Berawal dari saya pengen memperbarui repo yang tua dari suatu organisasi, niatnya pengen rumat ulang nih, ternyata dari orginal reponya ada update, sekalian buat artikel deh, lebih kurang gambaranya seperti ini.

Fig 1. Gambaran ribetnya.

Ada dua cara untuk memperbarui forked repository menggunakan web interface yang disediakan oleh github tapi ribet, atau melalui terminal yang lebih ribet lagi.

Melalui Github (boring way) 💻

Buka repo yang hasil fork di Github.
Klik Pull Requests di sebelah kanan, lalu New Pull Request.
Akan memunculkan hasil compare antara repo upstream dengan repo kamu(forked repo), dan jika menyatakan “There isn’t anything to compare.”, tekan link switching the base, yang mana sekarang repo kamu(forked repo) akan dibalik menjadi base repo dan repo upstream menjadi head repo.
Tekan Create Pull Request, beri judul pull request, Tekan Send Pull Request.
Tekan Merge Pull Request dan Confirm Merge.

* pastikan kamu tidak merubah apapun pada forked repo, supaya melakukan merge secara otomatis, kalo tidak ya paling2 konflik.

Melalui terminal ⌨️

Tambahkan remote alamat repository yang aslinya disini tak beri nama upstream., ganti ORIGINAL_OWNER dan ORIGINAL_REPO dengan alamat repo aslimu.

$ git add remote upstream git@github.com:ORIGINAL_OWNER/ORIGINAL_REPO.git
$ git remote -v
> origin    git@github.com:piharpi/www.git (fetch) # forked repo
> origin    git@github.com:piharpi/www.git (push) # forked repo
> upstream    git@github.com:ORIGINAL_OWNER/ORIGINAL_REPO.git (fetch) # upstream repo / original repo
> upstream    git@github.com:ORIGINAL_OWNER/ORIGINAL_REPO.git (push) # upstream repo / original repo

Checkout ke local branch master.

$ git checkout master
> Switched to branch 'master'

Jika sudah, Merge local repo dengan remote upstream/master.

$ git merge upstream/master

Terakhir push local repo ke remote origin.

$ git add -A
$ git commit -m "updating origin repo" && git push -u origin master

Selamat mencoba cara ribet ini, semoga bisa dipahami, saya sendiri lebih senang melalui terminal, klo ada yang ribet kenapa cari yang mudah.

Resources

How this site was born 👶

2018-09-28T04:45:47+02:00

The website was made using Jekyll the one of open source static sites generator, and using my own simple theme, I called klisé.

All the services are free, source code the site was placed on my github repository and intergration with ~~netlify~~ service, another service that you can use is github page for hosting your own static site.

Let’s do this

So, before we start create a site, you need some tools, you can self paced for how to installing each tools, on this guide i’m just want to show you how to install jekyll and deploying in netlify, but make sure you have each tools below.

Prerequisites

Requirements before we doing magic show.

Ruby programming language
Git (version control)
Netlify and Github account
Bundler

Installation

First, you need some SSG, there are many kind ssg, but in case i’m using Jekyll cause i’m already familiar with it, open your terminal and type command on below

$ bundle install jekyll # installing jekyll in your machine
$ jekyll new my-site && cd my-site # create new jekyll project
$ jekyll s # run jekyll server

Now, jekyll is running on your local machine, open your browser and go to localhost:4000 is default address from jekyll, press CTRL + C to stop the jekyll server.

Adding remote repository

Before we adding remote repository, you must have github repository, if already have repository, just add github remote address to your local folder, with the following commands

$ git init # initializing project folder
$ git remote add origin https://github.com/YOUR-USERNAME/YOUR-REPO.git # change UPPERCASE with your own!
$ git add -A && git commit -m "Initialize" && git push -u origin master # push code to github

Now check your github repository, make sure the files is uploaded correctly.

Deploying to netlify

Go netlify dashboard, and following this step.

click new site from git, then choose Github.
then choose your repository where is the jekyll sources uploaded.
netlify smart enough to configuring, we just need’s are hosting’s are hosting’s are hosting’s are hosting to click Deploy site button.

Wait for moment, and voila..! your site’s are hosting and using .netlify.com tld, if your website wants to look professional, just buy a domain from your favorite domain store. or if you the first time, I advice using namecheap.com(isn’t sponsor) *based on my experienced it provides good service and have various TLDs.

So, what you waiting for, just create your own website for free.

Python Notes from Intro to Machine Learning

2018-04-14T05:00:00+02:00

This article is for Demo purpose

The article was originally on this repo

I rarely use Python. I only have one repository at Github that is written in Python: iris-flower-classifier. And it was written two years ago!

A few days ago I took this free course from Udacity: Intro to Machine Learning. The machine learning related codes are quite easy to grasp since it simply uses the scikit-learn modules. But most of the supporting Python modules that are provided by this course were like a black-box to me. I had no idea how to download a file in Python or what’s the difference between a list, a tuple and a dictionary.

That’s why I decided to read all of the provided Python modules and implement it myself. I ended up refactor most of the code so it’s easier to understand: github.com/risan/intro-to-machine-learning.

So here are some notes and snippets of Python that I’ve been collecting so far (I’m not even halfway through the course 😝). Also, note that the codes here are still using Python version 2.7.

Example Table overflow-table

Markdown	HTML	Rendered Output
`[Example Link](#link)`	`Example Link`	Example Link
`_Be Italic_`	`Be Italic`	Be Italic
`Be Bold`	`Be Italic`	Be Bold

*) resize to see difference

Example Table without-overflow

Markdown	HTML	Rendered Output
`[Example Link](#link)`	`Example Link`	Example Link
`_Be Italic_`	`Be Italic`	Be Italic
`Be Bold`	`Be Italic`	Be Bold

*) resize to see difference

Main Entry File

Suppose our Python project is stored in /foo/bar directory. And this application has one file that serves as the single entry point. We can name this file __main__.py so we can run this project simply be referencing its directory path:

# Referencing its directory.
$ python /foo/bar

# It's equivalent to this.
$ python /foo/bar/__main__.py

Import Python Module Dynamically

Suppose we would like to import a Python module dynamically based on a variable value. We can achieve this through the __import__ function:

module_name = "numpy"

__import__(module_name)

Multiple Returns in Python

In Python, it’s possible for a function or a method to return multiple values. We can do this simply by separating each return value by a comma:

def test():
    return 100, "foo"

someNumber, someString = test()

Importing Modules Outside of the Directory

In order to import a module from outside of the directory, we need to add that module’s directory path into the current file with sys.path.append. Suppose we have the following directory structure:

|--foo
| |-- bar.py
|
|-- tools
| |-- speak_yoda.py

If we want to use the speak_yoda.py module within the bar.py, we can do the following:

# /foo/bar.py
import os

# Use relative path to tools directory.
sys.path.append("../tools")

import speak_yoda

However, this won’t work if we run the baz.py file from outside of its foo directory:

# It works inside of the /foo directory.
$ cd /foo
$ python bar.py

# But it won't work if the code runs from outside of /foo directory.
$ python foo/bar.py

To solve this problem we can refer to the tools directory using its absolute path.

# /foo/bar.py
import os
import sys

# Get the directory name for this file.
current_dirname = os.path.dirname(os.path.realpath(__file__))

# Use the absolute path to the tools directory
tools_path = os.path.abspath(os.path.join(dirname, "../tools"))
sys.path.append(tools_path)

import speak_yoda

Output

Print The Emojis

It turns out you can’t just print an emoji or any other Unicode characters to the console. You need to specify the encoding type beforehand:

# coding: utf8

print("😅")

Pretty Print

We can use the pprint module to pretty-print Python data structure with a configurable indentation:

import pprint
pp = pprint.PrettyPrinter(indent=2)

pp.pprint(people)

Working with Pathname

Read more about pathname manipulations in the os.path documentation.

Get Filename From URL

Suppose the last segment of the URL contains a filename that we would like to download. We can extract this filename with the following code:

import os
from urlparse import urlparse

url = "https://example.com/foo.txt"

url_components = urlparse(url)

filename = os.path.basename(url_components.path) # foo.txt

Check if File Exists

To check whether the given file path exists or not:

import os

is_exists = os.path.isfile("foo.txt")

Create a Directory if It Does Not Exists

To create a directory only if it does not exist:

import os
import errno

try:
    os.makedirs(directory_path)
except OSError, e:
    if e.errno != errno.EEXIST:
        raise

Working with Files

Downloading a File

We can use the urllib module to download a file in Python. The first argument is the file URL that we would like to download. The second argument is the optional filename that will be used to store the file.

import urllib

urllib.urlretrieve("https://example.com/foo.txt", "foo.txt")

Extracting Tar File

There’s a built-in tarfile module that we can use to work with Tar file in Python. To extract the tar.gz file we can use the following code:

import tarfile

# Open the file.
tfile = tarfile.open("foo.tar.gz")

# Extract the file to the given path.
tfile.extractall(path)

We can pass the mode argument to the open method. By default, the mode would be r—reading mode with transparent compression. There are also other mode options that we can use:

r:gz: Reading mode with gzip compression.
r:: Reading mode without compression.
a: Appending mode without compression.
w: Writting mode without compression.
Checkout other available options in tarfile documentation.

Working with List

Generate a List of Random Numbers

Use the for..in syntax to generate a list of random numbers in a one-liner style.

import random

# Initialize internal state of random generator.
random.seed(42)

# Generate random points.
randomNumbers = [random.random() for i in range(0, 10)]
# [0.6394267984578837, 0.025010755222666936, 0.27502931836911926, ...]

Pair Values from Two Lists

The built-in zip function can pair values from two lists. However, this zip function will return a list of tuples instead. To get a list of value pairs, we can combine it with for..in syntax:

coordinates = [[x, y] for x,y in zip([5,10,15], [0,1,0])]
# [[5, 0], [10, 1], [15, 0]]

Splitting a List

We can easily split a list in Python by specifying the starting index and it’s ending index. Note that the ending index is excluded from the result.

We can also specify a negative index. And also note that both of these indices are optional!

a = [0,1,2,3,4,5]

a[0:3]  # 0,1,2
a[1:3]  # 1,2
a[2:]   # 2,3,4,5
a[:3]   # 0,1,2
a[0:-2] # 0,1,2,3
a[-2:]  # 4,5
a[:]    # 0,1,2,3,4,5

Filtering a List In One Line

We can easily filter a list in Python by combining the for..in and the if syntax together:

numbers = range(1,11)

# Filter even numbers only.
[numbers[i] for i in range(0, len(numbers)) if numbers[i] % 2 == 0]
# [2, 4, 6, 8, 10]

Sorting a List in Ascending Order

In Python, we can sort a list in ascending order simply by calling the sort method like so:

people = ["John", "Alice", "Poe"]
people.sort()
print(people) # ["Alice", "John", "Poe"]

Using Filter Function with a List

Just like its name, we can use the filter function to filter out our list:

numbers = range(1, 11)

even_numbers = filter(lambda number: number % 2 == 0, numbers)
# [2, 4, 6, 8, 10]

We can break the above statement into two parts:

lambda number: statement: The first part is the function that we would like to run to every item on the list. number is the variable name we’d like to use in this function to refer to a single item from the numbers list. The following function body must evaluate to truthy/falsy value—falsy means the current item will be removed from the final result.
numbers: The second parameter is the list that we’d like to filter.

Using Reduce with a List of Dictionary

We can use the reduce function to calculate the total of a particular key in a list of a dictionary:

items = [{value:10}, {value:20}, {value:50}]

# Calculate the total of value key.
totalValues = reduce(lambda total, item: total + item["value"], items, 0) # 80

It can be broken down into 4 parts:

lambda total: It’s the variable name that we’d like to use in the function body to refer to the carried or the accumulative value that will finally be returned.
item: statement: item is the name of the variable we’d like to use within the function body to refer to the single item in the items list. The following function body will be executed in order to define the accumulative value of total for the next iteration.
items: It’s the list of item that we would like to “reduce”.
0: The last parameter is optional and it’s the initial accumulative value for the first iteration.

We can also use this reduce function to find a single item from the list. Here’s an example of code to find the person with the biggest total_payments within the given list of people dictionary.

people = [
    {"name": "John", "total_payments": 100},
    {"name": "Alice", "total_payments": 1000},
    {"name": "Poe", "total_payments": 800}
]

person_biggest_total_payments = reduce(lambda paid_most, person: person if person["total_payments"] > paid_most["total_payments"] else paid_most, people, { "total_payments": 0 })
# {'name': 'Alice', 'total_payments': 1000}

Working with Dictionary

Loop Through Dictionary

We can use the itervalues method to loop through a dictionary:

for person in people.itervalues():
    print(person["email_address"])

We can also use the iteritems method if we want to access the key too:

for person in people.iteritems():
    print(person[0] + ": " + person[1]["email_address"])

Calculate Total of Particular Dictionary Key

Suppose we would like to calculate the total amount of salary key on a people dictionary. We can extract the salary key and use the sum function to get the total:

total_salary = sum([person["salary"] for person in people.itervalues()])

Working with Numpy

Numpy Create Range of Values with The Given Interval

Use the arange method to create an array with an evenly spaced interval.

import numpy as np

np.arange(0, 5, 1)
# array([0,1,2,3,4])

np.arange(1, 4, 0.5)
# array([1. , 1.5, 2. , 2.5, 3. , 3.5])

Numpy Create Coordinate Matrices from Coordinate Vectors

We can use the Numpy meshgrid method to make coordinate matrices from one-dimentional coordinate arrays.

import numpy as np

np.meshgrid([1, 2, 3], [0, 7])
# [
#   array([[1,2,3], [1,2,3]]),
#   array([[0,0,0], [7,7,7]])
# ]

Flatten Numpy Array

When we have a multi-dimensional Numpy array, we can easily flatten it with the ravel method:

import numpy as np

arr = np.array([[1,2], [3,4]])
arr.ravel()
# array([1, 2, 3, 4])

Pairing Array Values with Second Axis

We can use Numpy c_ function to pair array values with another array that will be it’s second axis. Read the numpy.c_ documentation.

import numpy as np

x = [1,2]
y = [10,20]

np.c_[x, y]
# array([1,10], [2,20])

Generate Coordinates Across The Grid

With the knowledge of Numpy arange, meshgrid, ravel and c_ methods, we can easily generate an evenly spaced coordinates across the grid so we can pass it to the classifier and plot the decision surface.

import numpy as np

# Generate an evenly spaced coordinates.
x_points, y_points = np.meshgrid(np.arange(x_min, x_max, step), np.arange(y_min, y_max, step))

# Pair the x and y points.
test_coordinates = np.c_[x_points.ravel(), y_points.ravel()]

Plotting the Data

Plot The Surface Decision

We can pass an evenly spaced coordinates across the grid to the classifier to predict the output on each of that coordinate. We can then use matplotlib.pyplot to plot the surface decision.

import matplotlib.pyplot as plt
import pylab as pl

# Pass coordinates across the grid.
predicted_labels = classifier.predict(test_coordinates)

# Don't forget to reshape the output array dimension.
predicted_labels = predicted_labels.reshape(x_points.shape)

# Set the axes limit.
plt.xlim(x_points.min(), x_points.max())
plt.ylim(y_points.min(), y_points.max())

# Plot the decision boundary with seismic color map.
plt.pcolormesh(x_points, y_points, predicted_labels, cmap = pl.cm.seismic)

The classifier output would be a one-dimensional array, so don’t forget to reshape it back into a two-dimensional array before plotting. The cmap is an optional parameter for the color map. Here we use the seismic color map from pylab module. It has the red-blue colors.

Scatter Plot

We need to separate the test points based on its predicted label (the speed). So we can plot the test points with two different colors.

# Separate fast (label = 0) & slow (label = 1) test points.
grade_fast = [features_test[i][0] for i in range(0, len(features_test)) if labels_test[i] == 0]
bumpy_fast = [features_test[i][1] for i in range(0, len(features_test)) if labels_test[i] == 0]
grade_slow = [features_test[i][0] for i in range(0, len(features_test)) if labels_test[i] == 1]
bumpy_slow = [features_test[i][1] for i in range(0, len(features_test)) if labels_test[i] == 1]

# Plot the test points based on its speed.
plt.scatter(grade_fast, bumpy_fast, color = "b", label = "fast")
plt.scatter(grade_slow, bumpy_slow, color = "r", label = "slow")

# Show the plot legend.
plt.legend()

# Add the axis labels.
plt.xlabel("grade")
plt.ylabel("bumpiness")

# Show the plot.
plt.show()

If we want to save the plot into an image, we can use the savefig method instead:

plt.savefig('scatter_plot.png')

Dealing with Data

Deserializing Python Object

We can use pickle module for serializing and deserializing Python object. There’s also the cPickle—the faster C implementation. We use both of these modules to deserialize the email text and author list.

import pickle
import cPickle

# Unpickling or deserializing the texts.
texts_file_handler = open(texts_file, "r")
texts = cPickle.load(texts_file_handler)
texts_file_handler.close()

# Unpickling or deserializing the authors.
authors_file_handler = open(authors_file, "r")
authors = pickle.load(authors_file_handler)
authors_file_handler.close()

Split Data for Training and Testing

We can use the built-in train_test_split function from scikit-learn to split the data both for training and testing.

from sklearn.model_selection import train_test_split

features_train, features_test, labels_train, labels_test = train_test_split(texts, authors, test_size = 0.1, random_state = 42)

The test_size argument is the proportion of data to split into the test, in our case we split 10% for testing.

Vectorized the Strings

When working with a text document, we need to vectorize the strings into a list of numbers so it’s easier and more efficient to process. We can use the TfidfVectorizer class to vectorize the strings into a matrix of TF-IDF features.

from sklearn.feature_extraction.text import TfidfVectorizer

vectorizer = TfidfVectorizer(sublinear_tf = True, max_df = 0.5, stop_words = "english")
features_train_transformed = vectorizer.fit_transform(features_train)
features_test_transformed = vectorizer.transform(features_test)

Word with a frequency higher than the max_df will be ignored. Stop words are also ignored—stop words are the most common words in a language (e.g. a, the, has).

Feature Selection

Text can have a lot of features thus it may slow to compute. We can use scikit SelectPercentile class to select only the important features.

selector = SelectPercentile(f_classif, percentile = 10)
selector.fit(features_train_transformed, labels_train)
selected_features_train_transformed = selector.transform(features_train_transformed).toarray()
selected_features_test_transformed = selector.transform(features_test_transformed).toarray()

The percentile is the percentage of features that we’d like to select based on its highest score.

This article is for Demo purpose

The article was originally on this repo