test your typing speed without leaving the terminal

This repository is home to a self-contained file typetest. As is it is a near clone of 10fastfingers with an added bonus of being able to see typing speed as you're typing.

Differences in the way typing speed is calculated and feedback across platforms got me interested in writing my own program for testing typing speed. I've come to love how simple and unrestrictive 10fastfingers and keybr feel compared to typingclub and typeracer. They all have great advantages for varying purposes but when it comes to warming up or just waiting for some program to compile (have you tried compiling chromium?) I am yet to find a rival to 10fastfingers. That is why I decided to clone its functionality and add some features I love from other sites.

typetest

typetest is a self-contained minimal typing test program written with blessed. It calculates typing speed as sum of spaces and characters from correctly written words divided by test duration. Adjustable settings are DURATION, SHUFFLE and NUMBER_OF_ROWS, which can be set using the command arguments. The input text for the typing test is read from the standard input or using the positional arguments. The results of typetest go into a file aptly named results positioned in the same directory as typetest. If you wish to redirect results to another file do so in a standard unix way, by adding >> my_results to the command arguments.

ideas for tests

Along with typetest this repository features sample tests. Try them like so: typetest -s -d 60 < common_200. Scrape something of the net, like a featured article on wikipedia.

#!/usr/bin/env python3
import re
import requests
from bs4 import BeautifulSoup

word_pattern = re.compile(r"['A-Za-z\d\-]+[,\.\?\!]?")  # symbols to keep
url = 'https://en.wikipedia.org/wiki/Special:RandomInCategory/Featured_articles'

r = requests.get(url)
soup = BeautifulSoup(r.text, 'html.parser')
for sup in soup.select('sup'):
    sup.extract()  # remove citations

text = ' '.join(p.text for p in soup.select('p'))
text = re.sub(r'\[.*?\]|\(.*?\)', '', text)  # remove parenthesis
print(' '.join(re.findall(word_pattern, text)))

If you create a file called wiki_random you can start the test with wiki_random | typetest. Write your own scraper, you may find some suggestions here.

usage

usage: typetest [-h] [-d DURATION] [-r ROWS] [-s] [words [words ...]]

example:
  typetest -d 3.5 The typing seems really strong today.
  echo 'I love typing' | typetest
  typetest < test.txt

positional arguments:
  words                 provide words via args in lieu of stdin

optional arguments:
  -h, --help            show this help message and exit
  -d DURATION, --duration DURATION
                        duration in seconds
  -r ROWS, --rows ROWS  number of test rows to show
  -s, --shuffle         shuffle words

shortcuts:
  ^c / ctrl+c           end the test and get results now
  ^h / ctrl+h           backspace
  ^r / ctrl+r           restart the same test
  ^w / ctrl+w           delete a word
  ^u / ctrl+u           delete a word

installation

*nix

1. install python 3
2. install blessed
3. clone this repository
4. run python typetest -s -d 60 < common_300
5. (optional) add typetest to path or make an alias like tt
6. (optional) store your results in some file and analyze

windows

Two caveats:

1. signal.SIGWINCH is not available on windows, you can get a similar behaviour by decreasing the inkey's timeout parameter.
2. a way for getting test words via stdin and then reading user input has to be found (file handles 0, 1, and 2 aren't the same file descriptor as in *nix), maybe a '/dev/tty' equivalent?

A way to completely avoid the aforementioned caveats is to use a linux subsystem (WSL), installation details can be found here. The rest of the installation steps are the same as for *nix.