Installing Humdrum Tools from GitHub¶
This chapter describes the primary method of installing command-line Humdrum processing tools on your computer. Two base packages can be downloaded from GitHub: The Humdrum Toolkit and Humdrum Extras. If you only want to install one or the other, go the respective GitHub repository; however, the two packages usually should be installed together from a meta repository called humdrum-tools.
The humdrum-tools GitHub repository contains submodules for both The Humdrum Toolkit and Humdrum Extras. Documentation (such as this chapter) as well as sample data can optionally be downloaded through the humdrum-tools metapackage.
cd # Go to installation directory (home directory in this example). git clone --recursive https://github.com/humdrum-tools/humdrum-tools cd humdrum-tools # Go into repository to run make commands. make update # Make sure you have the most recent humdrum/humextra code. make # Compile C/C++ programs and create bin directories. make install # Add bin directories to $PATH environment variable. make regression # Test tools to make sure they are working on your computer. make data # Download sample Humdrum data files to work with. make doc # Download documentation for offline reference.
Otherwise, follow the more detailed instructions below. Super-users can instead configure humdrum-tools for all users on a computer by typically installing in /usr/local/humdrum-tools and adding the humdrum/bin and humextra/bin directories to the PATH environment variable in the login script for all users (which depends on the shell: /etc/profile for most shells, /etc/zshenv for zsh and /etc/csh.cshrc for csh/tcsh).
The Humdrum Toolkit is a set of unix command-line programs which process data files, including musical scores, in the Humdrum file format. Most Humdrum Toolkit programs are written in AWK that are called from shell scripts which handle command-line options, although the toolkit also includes a few C-language programs.
The Humdrum Toolkit can be used on unix-based computer systems such as linux and Apple OS X. To use in MS Windows, install a unix terminal emulator. A comprehensive package for linux tools in Windows can be downloaded from http://www.cygwin.com. The simplest method is to download all cygwin packages when installing so that components do not have to be added later when you notice that they are missing.
To download this repository, first make sure that git is installed on your computer. Git is a version control program which is the main interface to online repositories on GitHub. To check if the git program is available on your computer, type the following in a terminal:
If the terminal replies with a line such as /usr/local/bin/git, then git is present and you can run the installation commands given further above. If the which command replies with an error that git cannot be found, you need to install git. How to do this will depend on your operating system. Here are installation hints for various computer systems:
In linux, the installation command for git is usually one of these two possibilities:
sudo yum install git sudo apt-get install git
For cygwin on MS Windows computers, you should have included git when installing packages when you first installed cygwin. Re-run the installation program and include git in the package installation list if the which command does not find it.
For Apple OS X, the easiest method is to download git from this link. More advanced Mac users can use Homebrew to install git with this command (but installing Homebrew probably also includes automatic installation of git):
brew install git
For individual user installations, the humdrum-tools repository can reside anywhere within a user’s home directory. For system-wide installation, the recommended location is /usr/local/humdrum-tools. The following instructions are for individual account installations, but system-wide installation will be similar. The main difference will be the location of the shell startup script where the PATH needs to be set (see the Installing section below for setting up the PATH shell variable).
To download humdrum-tools, type these commands:
cd # Go to home directory or wherever you want to install, # such as "cd /usr/local" for system-wide installations. git clone --recursive https://github.com/humdrum-tools/humdrum-tools
The --recursive option is needed to download each of the individual repositories inside the humdrum-tools meta-repository. Sample Humdrum file data can also be downloaded by typing the following command within the humdrum-tools directory:
cd ~/humdrum-tools # or wherever humdrum-tools was downloaded. make data
If you want to download the data files outside of the humdrum-tools directory (for example into a user directory when the tools are installed in /usr/local/humdrum-tools) then download with this command instead:
cd # or wherever you want to install the data. git clone --recursive https://github.com/humdrum-tools/humdrum-data
A local copy of the http://www.humdrum.org website can also be downloaded for off-line use with these commands:
cd ~/humdrum-tools # or wherever humdrum-tools was downloaded. make doc # or to install outside of the humdrum-tools directory: git clone https://github.com/humdrum-tools/humdrum-tools.github.io humdrum-documentation
Note that the humdrum-tools and humdrum-data repositories cannot be downloaded in a very convenient format from the ZIP link on the Github website since repository submodules will not be included in the ZIP file. GitHub may allow submodule inclusion in their ZIP downloads in the future.
To compile programs in the two humdrum-tools submodules, type make inside of the humdrum directory:
cd ~/humdrum-tools # or wherever humdrum-tools was downloaded. make
Note that to use the make command or gcc for compiling the C/C++ programs, these must already be installed. Check to see if gcc is installed by typing these commands:
which gcc which make
The terminal should reply with something like /usr/bin/gcc and /usr/bin/make.
If gcc is not installed, then you will have to figure out how to install it on your computer first. Linux/Unix computers usually have it pre-installed along with the operating system; if not, then typing sudo yum install gcc or sudo apt-get install gcc will typically install it.
Apple OS X does not include it by default, so you will have to install it. If you are using OS X Mavericks or later, then type xcode-select --install to install the Xcode command line tools, which includes gcc (or actually a similar compiler). The make command will be installed at the same time that gcc is installed.
Cygwin users would have to re-run the installation program and include the compile tools if gcc was not initially installed with cygwin (a minimal installation will not include gcc).
To use humdrum-tools commands within any directory, you must add the humdrum/bin and humextra/bin directories to the PATH environment variable. First, determine the shell (unix command-line interpreter) which you are using in the terminal by running this command: echo $SHELL For bash shells (most common shell), the above command should reply with the text /bin/bash. If you are using another shell, bash will be replaced with the name of the shell you are using.
To temporarily adjust the PATH variable so that you can immediately start using the tools in the current terminal session, here are the two main methods of setting the PATH variable in various shells:
In bash, sh, ksh, zsh, type these commands:
cd ~/humdrum-tools # or wherever humdrum-tools was downloaded. PATH=`pwd`/humdrum/bin:$PATH PATH=`pwd`/humextra/bin:$PATH
Or in csh or tcsh, do these commands:
cd ~/humdrum-tools # or wherever humdrum-tools was downloaded. set PATH=`pwd`/humdrum/bin:$PATH set PATH=`pwd`/humextra/bin:$PATH
For a persistent installation of humdrum tools whenever you open a new terminal, the PATH environment variable needs to be amended during login with the paths of the humdrum-tools executables. The above PATH= lines must be added to the shell login script. The name of this shell login file is different for different shells. If you type:
cd ~/humdrum-tools make install
then the humdrum-tools makefile will attempt to place those lines in the correct file based on your login shell; otherwise, you can add the lines manually to the shell startup scripts as outlined below. The command make install-hint will suggest the commands needed to add the humdrum-tools bin directories permanently to the PATH environment variable if you want to manually configure it. Don’t run the temporary installation given further above before running make install, since this make target looks at the PATH environment variable to decide if the PATH needs to be updated. Login again if you already ran the temporary PATH update. Initialization files for various shells are given in Table 1.1. Choose the shell and installation type to select the correct setup file to edit.
Table 1.1 Locations of startup scripts for various unix shells. The PATH environment variable should be modified in one of the files to include the humdrum-tools bin directories.
|Shell||Single-user setup file||System-wide setup file|
|bash||~/.bash_profile, else ~/.bash_login, else ~/.profile||/etc/profile|
Bash is the most common shell. For single-user installations, the choice of setup file is complicated: If the file ~/.bash_profile exists (“~” is unix shorthand for your home directory), then the bash shell will read that file. If ~/.bash_profile does not exist, then bash will instead try to read ~/.bash_login. If ~/.bash_login does not exist, then bash will try to read ~/.profile. Only one of those files will be read, and ~/.bash_profile is the first one bash that will try to read. Note that if there is currently a ~/.profile file but no ~/.bash_profile, the settings in ~/.profile will be ignored if you create the ~/.bash_profile file.
For system-wide installations, super-users will have to add the PATH= lines to the correct file within the /etc/ directory. Super-users can usually install for all users on a computer system in any common shell by running these commands:
cd /usr/local/humdrum-tools # or wherever humdrum-tools was downloaded. # bash, sh, ksh, and zsh sudo echo "export PATH=`pwd`/humdrum/bin:$PATH" >> /etc/profile sudo echo "export PATH=`pwd`/humextra/bin:$PATH" >> /etc/profile # csh and tcsh: sudo echo "set PATH=`pwd`/humdrum/bin:$PATH" >> /etc/csh.cshrc sudo echo "set PATH=`pwd`/humextra/bin:$PATH" >> /etc/csh.cshrc
Many linux systems have a directory called /etc/profile.d into which package-specific shell settings are placed rather than altering /etc/profile. If so, then for bash-like shells, create a file in that directory with the PATH variable updates rather than editing the /etc/profile file:
if [ -e /etc/profile.d ] then cd /usr/local/humdrum-tools # or wherever humdrum-tools is installed. echo "export PATH=`pwd`/humdrum/bin:$PATH" >> /etc/profile.d/humdrum.sh echo "export PATH=`pwd`/humextra/bin:$PATH" >> /etc/profile.d/humdrum.sh else echo "/etc/profile.d directory does not exist." fi
To verify that the PATH lines were added to the correct file, try opening a new terminal window and type:
echo $PATH | tr : '\n' | grep humdrum
The computer should reply with the full path names of the bin directories for the Humdrum Toolkit and Humdrum Extras. With a system-wide installation, the above command should display something like this:
You can also use the which command to see where a command is located. If the humdrum-tools command paths are set up correctly, then the following commands:
which key which keycor
should reply something like this:
This implies that the PATH variable contains the correct two bin directories for using humdrum-tools.
The command make regression will run regression tests for the Humdrum Toolkit and Humdrum Extras. The make target will run the commands on files and compare the results to the expected output from the programs for the given options. Here is a sample of the regression test display:
(cd humdrum; make regression) TEST 01 for accent: OK TEST 01 for assemble: OK TEST 01 for barks: OK TEST 01 for cbr: OK ... (cd humextra; make regression) bin/run-command-tests autostem test 001 OK Add stems to notes in the treble clef autostem test 005 OK Add stems to notes, overwriting any which already exist. barnum test 007 OK Numbering measures with repeat bars in the middle of the measure. harm2kern test 006 OK Seventh chords and their inversions in C Major. harm2kern test 007 OK Chord qualities with a root on C. myank test 009 OK Extract a measure, not including ending barline. prange test 003 OK Count pitches with duration weighting rscale test 010 OK Recover original rhythms from exotic rhythmic values. sample test 001 OK Sample the music every quarter-note.
If you only want to see failed tests, run this command:
make regression | grep -v OK
Software (and data if installed with make data) periodically can be updated to the most recent versions by typing these commands:
cd `which mint | sed 's/humdrum\/bin\/mint$//'` make update # Download any updates from GitHub. make # Re-compile the programs.
make update will download any new changes, and make will recompile the programs. If you make changes to the files in the humdrum-tools directory, the above commands may complain if the same file has been updated in the repository. Type git status to see what files have been locally modified or added since you did the last update or download.
If you want to keep local changes but still update, try these commands:
cd `which mint | sed 's/humdrum\/bin\/mint$//'` git stash make update make git stash apply
If you want to undo any local changes before updating, you can run this command:
cd `which mint | sed 's/humdrum\/bin\/mint$//'` git reset --hard HEAD^ make update make