• Stars
    star
    318
  • Rank 131,872 (Top 3 %)
  • Language
    Python
  • License
    GNU Affero Genera...
  • Created over 7 years ago
  • Updated over 3 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Typeracer-like console app for measuring your WPM

wpm β€” measure and improve your typing speed

Supported Python versions Project License pypi

wpm is a curses-based UNIX terminal program for measuring and improving your typing speed (measured in words per minute, or WPM).

It depends only on standard Python libraries and therefore works with Python 2, 3 and PyPy.

Features

  • Over 4900 quotes in the database, shamelessly stolen from typeracerdata.com
  • Extremely low typing latency!
  • Timer starts when you strike the first key
  • Completed text is darkened, helping you to focus ahead
  • Keep separate scores for, e.g. type of keyboard, layout etc.
  • Saves race scores in a CSV file that is a superset of TypeRacer's export format. Loads fine in Excel as well.
  • Launches quickly in your terminal window for "in-between moments"

How to install

The recommended way is to install via PyPi

$ pip install wpm

The above usually requires sudo. If you don't want to install it system-wide, you can use pip install wpm --user.

Remember to check for upgrades with pip install --upgrade wpm. You can also install it from the source repository with

$ pip install . [--user]

To just test the app without installing, type make run.

How to run

Just type wpm to start the program. The timer will start when you press the first key. At any time, you can hit ESCAPE to quit.

You can backspace for the current word you're editing, if you make a mistake. Mistakes will lower the accuracy score.

If you have problems finding the wpm file, you can also start it by typing python -m wpm. You can also see options with python -m wpm --help.

Calculating WPM

The WPM is calculated by dividing characters per second by five and then multiplying that with 60. This is a well-known formula, but gives slightly higher scores than on sites like typeracer.com. It is, however, good enough to gauge your typing speed. And it works offline, and with your own texts.

Regarding TypeRacer, I really suggest everyone check it out. I use this program merely to warm up before heading over to typeracer.com, where you can race against others.

How to get the lowest typing latency

Run outside of tmux, and use a really speedy terminal window. On my macOS system, I found the best latency using the built-in Terminal.app, which easily beats iTerm. I also found the Kitty terminal to provide very low latency.

On Linux, the ultimate typing latency is achieved if you open up one of the virtual consoles. For example, hit CTRL+ALT+F2 and log in, set your TERM=xterm-color and run wpm. Many terms also have quite a high latency. Try using uxterm if you need to run it inside X.

How to improve your typing speed

I believe that everyone can type at 100 WPM with enough practice. If you are currently typing slower than that, my suggestions are:

  • Learn to type without looking at the keyboard
  • Learn to use all your fingers
  • Sit up straight and type in a comfortable situation

If you are consistently above 100 WPM:

  • Focus on the next word
  • Type words instead of characters
  • Train muscle memory
  • Type hard parts slower
  • Raise your wrists

Practice a little bit every few days, but don't overdo it. Stop when you're tired or feeling unmotivated.

Loading custom texts

If you want to type a custom text, run

$ wpm --load yourfile.txt

If you use --load, the author will currently be empty, the title will be the basename of the file. The text ID will be its inode, just to make them somewhat unique, so your stats will work.

You can also bundle up several texts into a single JSON file, using wpm --load-json yourfile.json. It must have the following format:

[
  {
    "author": "Author Name",
    "title": "Title of Work",
    "text": "The text to type here ..."
    "id": 123,
  },
  ...
]

The id is an optional integer. If you leave it out, an increasing, zero-based integer will be used.

Format of race history

wpm will save scores in a CSV file in ~/.wpm.csv. This file can be loaded directly into Excel. It uses the same format as TypeRacer, with the addition of a few extra columns at the end. That means is should be possible to use existing TypeRacer score history tools with this file with minor modifications.

The column order is:

Column Datatype Explanation
race int Race number, always increasing and tied to timestamp
wpm float The average WPM for that quote that single time
accuracy float From 0 to 1, where 1 means no mistakes
rank int Always 1
racers int Always 1
text_id int Item number of text in given database
timestamp str UTC timestamp in strptime format %Y-%m-%d %H:%M:%S.%f
database str Either "default" or the basename of the file used
tag str A user supplied tag for that score (e.g., keyboard)

Should there be any problem saving or loading the score history, it will copy the existing file into ~/.wpm.csv.backup and create a new one.

Tagging races

If you use --tag=... to tag your scores, this will be used until you change it. It is just a free text field that is saved along with each race result. It is useful to compare how well you are typing in various situations.

For example, perhaps you want to check if you are typing faster (but perhaps less accurate?) on different keyboards, or you are learning a new keyboard layout like Dvorak or Colemak and then use the tags --tag=qwerty and --tag=dvorak. If you are learning to touch type, or type with more fingers, you often start out slower than your normal speed. Tagging is a great way to keep track of your progress.

By running wpm --stats (or just -s), you will see a table of statistics, grouped by each tag. It shows things like the average over time, along with confidence and prediction intervals. An item like n-10 means "the last 10 games".

The ~/.wpmrc file

The first time you start wpm, it writes a .wpmrc file to your home directory. It contains user settings that you can change. They are given in the table below.

Section Name Default Description
curses escdelay 15 Time in ms to wait for follow-up key after ESC
curses window_timeout 20 Time in ms until giving up waiting for a keypress. If negative, wait forever.
wpm confidence_level 0.95 The confidence level for WPM statistics
wpm cpm 0 If positive, report CPM in stats instead of WPM
wpm tab_spaces 1 Number of spaces to expand tabs to
wpm wrap_width -1 If positive, wrap text at this width
xterm256colors Β  Β  Color codes for 256-color terminals (foreground, background)
xtermcolors Β  Β  Color codes for ordinary terminals (foreground, background)

Development features

You can enable certain unofficial features by seting the environment variable WPM_DEVFEATURES=feature1:feature2:etc..

Look in the file wpm/devfeatures.py for a list.

License

Copyright 2017, 2018 Christian Stigen Larsen

Distributed under the GNU Affero General Public License (AGPL) v3 or later. See the file LICENSE.txt for the full license text. This software makes use of open source software.

The quotes database is not covered by the AGPL!

More Repositories

1

jp2a

Converts jpg images to ASCII
HTML
608
star
2

mandelbrot-js

Fast rendering of the Mandelbrot set in HTML5 canvas using JavaScript
JavaScript
344
star
3

minijit

A basic x86-64 JIT compiler written from scratch in stock Python
Python
210
star
4

stack-machine

A simple stack-based virtual machine in C++ with a Forth like programming language
C++
162
star
5

python-simple-vm

A simple virtual machine w/constant folding implemented in Python
Python
117
star
6

arv

A fast 23andMe DNA parser and inferrer for Python
C++
114
star
7

mersenne-twister

This Mersenne Twister is a fast pseudo-random number generator (PRNG) in C++
C++
83
star
8

dna-traits

A fast 23andMe genome text file parser, now superseded by arv
Python
65
star
9

mickey-scheme

Mickey Scheme is an interpreter for R7RS Scheme written in pure C++
C++
62
star
10

crianza

A stack machine VM, interpreter and genetic programming library
Python
47
star
11

miller-rabin

The Miller-Rabin probabilistic primality test in C++ w/GMP
C++
30
star
12

libunwind-examples

A few libunwind examples
C++
27
star
13

lua-cpp

Tutorial code for Lua and C++ integration
C++
20
star
14

luajit-cpp

Example C++ shared library loaded in LuaJIT through FFI
C++
19
star
15

brainfuck-jit

Brainfuck JIT VMs
Python
17
star
16

c64-examples

Simple C64 programs compiled from the command line
Assembly
15
star
17

lyn

Python bindings for GNU Lightning
Python
9
star
18

vimp

Command line plugin manager for vim
Python
8
star
19

busy-beaver

Calculates the uncomputable Busy Beaver Ξ£-function
Python
7
star
20

gameboy

A Gameboy emulator in Python
Python
7
star
21

eulers-totient-function

A fast implementation of Euler's totient function phi(n) in C++
C++
7
star
22

skall

SKALL is a minimal, experimental UNIX shell
C
6
star
23

cellular-automaton

Cellular automaton using the HTML5 canvas.
6
star
24

nash-cipher

John Nash's encryption scheme from his 1955 letter to the NSA in C++.
C++
5
star
25

chicken-play

Chicken Scheme HTML5 rendering library
Scheme
3
star
26

poseur

Simple presentation tool
Python
3
star
27

vev

Simple HTTP server request routing in Python
Python
3
star
28

mickey-scheme-historic

Mickey Scheme is an interpreter for R7RS Scheme written in C++
C++
3
star
29

rosalind

Solutions to the Rosalind.info puzzles
C++
2
star
30

impute-me

This is the code behind the www.impute.me site. It contains algorithms for imputing personal genomes, as well as a range of custom-made analysis for genetics-based disease and health.
R
2
star
31

elv

Parses bank CSV files
Python
1
star
32

q

Prints C/C++ definitions from current directory
Python
1
star
33

latex-template

A small LaTeX template to get you started
TeX
1
star
34

presentation-vm

The presentation "How to make a simple virtual machine"
Brainfuck
1
star
35

project-euler

My Project Euler solutions
C++
1
star
36

dotfiles

Collection of my personal dot-files
Shell
1
star
37

armstrong-notes

Notes for M.A. Armstrong's Groups and Symmetry
TeX
1
star
38

py-html-generator

A quick-and-dirty HTML document generator and renderer
Python
1
star
39

callcc-c

Experimental undelimited continuations in C via x86-64 assembly
C
1
star