Mat-YAML is an interface to LibYAML for Matlab and GNU Octave.

Mat-YAML is intended to make it easy to convert Matlab structs and vectors to
and from YAML mappings and sequences. Because the correspondence between these
types is not perfect (eg, YAML mappings can have arbitrary keys, while Matlab
structs' field names must be valid Matlab identifiers), this is not always
possible. But, in many cases, the correspondence is close enough to be useful.

In dumping Matlab objects to YAML, arrays (of any type except char) become
sequences, scalar structs become mappings, char arrays become strings, numeric
values become floats, and logical values become bools. In loading YAML to
Matlab, sequences become cell arrays, mappings become structs (if possible), and
scalars are resolved according to their type. Because Matlab lacks a true
reference type, documents that include aliases cannot be loaded using the simple
interface. Note also that the Matlab -> YAML -> Matlab round-trip will not
preserve the original data structure exactly (though recovering it should be
fairly trivial in most cases).

In addition to this simple high-level interface, Mat-YAML provides a lower-level
means of representing YAML documents within Matlab which can be used as a basis
for addressing "pathological" mappings (having keys that aren't valid Matlab
identifiers) or aliases. Tinker as necessary!

Note well!
    Reasonable efforts have been made to ensure that this project is without
    serious bugs, but this is an amateur project, and as the license says, it is
    distributed to you "without warranty of any kind." Always back up your data!
    If you do discover a bug, please submit a report on the project's issue
    tracker, or, even better, send in a fix.

INSTALLATION

Linux / Mac OS:
1. Download and install LibYAML (available from http://pyyaml.org/wiki/LibYAML).
   Follow the directions from LibYAML's README, but it should be as simple as
   executing these three commands from the shell in the directory you unpacked
   LibYAML to:
   % ./configure
   % make
   % sudo make install
   (Regarding the libyaml Ubuntu package, see troubleshooting note [1].)

2. Build the yaml_mex mex file for your platform. If LibYAML installed nicely
   and you already have MEX configured properly for your system (see the Matlab
   documentation for that), then you should just need to change to the mex-src
   directory and enter the following command at the Matlab prompt:

   >>> mex yaml_mex.c yaml_mex_util.c -lyaml

   (If this fails and you're running a 64-bit Mac OS system, see troubleshooting
   note [2].)

3. Add mat-yaml/mex-src and mat-yaml/mfiles to your Matlab path, or copy the
   files in them to a directory already on your path. Verify that you can load
   and dump YAML files using yaml_load and yaml_dump.

Windows:
    I'm afraid I haven't got the slightest idea how to build this on Windows,
    and I lack access to a machine (and the time) that would let me figure it
    out. If you build Mat-YAML successfully on Windows, please contact me and
    share your knowledge.

Troubleshooting notes:
[1]  If you are running Linux and installed LibYAML from the Ubuntu repositories
     (perhaps as a dependency of PyYAML) instead of building it yourself, it
     seems that the necessary header file yaml.h is missing, and there are also
     link errors when trying to build yaml_mex. There doesn't seem to be any
     problem with having both the repository and the manual installation, so go
     ahead and install it manually as well.
     This issue was encountered and resolved under Linux Mint 13 Maya.

[2]  If you are on a 64-bit Mac OS system, the mex build may fail, because
     libyaml may build as 32-bit by default, and Matlab can't link to that code
     for a 64-bit MEX file. If this happens, try the following commands instead
     of "make" during the libyaml installation process:

     % make CFLAGS='-g -O2 -arch x86_64'
     % cp src/.libs/libyaml.a <path/to/mat-yaml>/mex-src/libyaml64.a

     And then in Matlab:
     >>> mex yaml_mex.c yaml_mex_util.c libyaml64.a

     This issue was encountered and resolved under Mac OS 10.5.8.