[Release] JSONLab v2.9.8 contains many new features and updates

[fangq]

JSONLab v2.9.8 - code named "Micronus Prime - beta" - is the beta-release of the next milestone (v3.0), containing a number of key feature enhancements and bug fixes. The major new features include

1. exporting JSON Memory-Map (jsonget,jsonset) for rapid disk-map like reading/writing of JSON/binary JSON files and writing, implementing JSON-Mmap spec v1 Draft 1
2. supporting JSONPath query (jsonpath) to MATLAB data and JSON/binary JSON file and streams, including deep-scan operators,
3. (breaking) upgrading the supported BJData spec to V1 Draft 2 where the default numerical data byte order changed from Big-Endian to Little-Endian,
4. adding initial support to JData _DataLink_ decoding to link multiple JSON/binary JSON files
5. dynamically cache linked data files (jsoncache, jdlink) to permit on-demand download and processing of complex JSON-encoded datasets such as neuroimaging datasets hosted on https://neurojson.io
6. support high-performance Blosc2 meta-compressor for storing large N-D array data,
7. savejson/loadjson can use MATLAB/Octave built-in jsonencode/jsondecode using the BuiltinJSON option
8. automatically switch from struct to containers.Map when encoded key-length exceeds 63
9. provide fall-back zlib/gzip compression/decompression function (octavezmat) on Octave when ZMat is not installed
10. include built-in .nii/.nii.gz/.jnii/.h5/.snirf/.tsv/.csv parsers to allow loadjd.m to read wide range of files
11. include json2couch from jbids (https://github.com/NeuroJSON/jbids) to allow uploading json files to CouchDB server

the package is super compact (199KB zipped), highly portable and supports all versions of MATLAB as old as R2008, and all versions of GNU Octave as old as 3.8. We spent a huge effort to make sure the code is compatible with both MATLAB and Octave so that you can write a single code and expect it to run robustly across all MATLAB/Octave versions.

Click here to register and download and fill out a short survey (if you haven't) to help us on future integration of JSONLab with the large JSON data resources on https://neurojson.io