JData for Python - lightweight and serializable data annotations for Python

• Copyright: (C) Qianqian Fang (2019-2024) <q.fang at neu.edu>
• License: Apache License, Version 2.0
• Version: 0.6.0
• URL: https://github.com/NeuroJSON/pyjdata
• Acknowledgement: This project is supported by US National Institute of Health (NIH) grant U24-NS124027

The JData Specification defines a lightweight language-independent data annotation interface enabling easy storing and sharing of complex data structures across different programming languages such as MATLAB, JavaScript, Python etc. Using JData formats, a complex Python data structure, including numpy objects, can be encoded as a simple dict object that is easily serialized as a JSON/binary JSON file and share such data between programs of different languages.

Since 2021, the development of PyJData module and the underlying data format specificaitons JData and BJData have been funded by the US National Institute of Health (NIH) as part of the NeuroJSON project (https://neurojson.org and https://neurojson.io).

The goal of the NeuroJSON project is to develop scalable, searchable, and reusable neuroimaging data formats and data sharing platforms. All data produced from the NeuroJSON project will be using JSON/Binary JData formats as the underlying serialization standards and the lightweight JData specification as language-independent data annotation standard.

How to install

• Github: download from https://github.com/NeuroJSON/pyjdata
• PIP: run pip install jdata see https://pypi.org/project/jdata/

This package can also be installed on Ubuntu 21.04 or Debian Bullseye via

sudo apt-get install python3-jdata

On older Ubuntu or Debian releases, you may install jdata via the below PPA:

sudo add-apt-repository ppa:fangq/ppa
sudo apt-get update
sudo apt-get install python3-jdata

Dependencies:

• numpy: PIP: run pip install numpy or sudo apt-get install python3-numpy
• (optional) bjdata: PIP: run pip install bjdata or sudo apt-get install python3-bjdata, see https://pypi.org/project/bjdata/, only needed to read/write BJData/UBJSON files
• (optional) lz4: PIP: run pip install lz4, only needed when encoding/decoding lz4-compressed data
• (optional) backports.lzma: PIP: run sudo apt-get install liblzma-dev and pip install backports.lzma (needed for Python 2.7), only needed when encoding/decoding lzma-compressed data
• (optional) blosc2: PIP: run pip install blosc2, only needed when encoding/decoding blosc2-compressed data

Replacing pip by pip3 if you are using Python 3.x. If either pip or pip3 does not exist on your system, please run

    sudo apt-get install python3-pip

Please note that in some OS releases (such as Ubuntu 20.04), python2.x and python-pip are no longer supported.

One can also install this module from the source code. To do this, you first check out a copy of the latest code from Github by

    git clone https://github.com/NeuroJSON/pyjdata.git
    cd pyjdata

then install the module to your local user folder by

    python3 setup.py install --user

or, if you prefer, install to the system folder for all users by

    sudo python3 setup.py install

Please replace python by python3 if you want to install it for Python 3.x instead of 2.x.

Instead of installing the module, you can also import the jdata module directly from your local copy by cd the root folder of the unzipped pyjdata package, and run

   import jdata as jd

How to use

The PyJData module is easy to use. You can use the encode()/decode() functions to encode Python data into JData annotation format, or decode JData structures into native Python data, for example

import jdata as jd
import numpy as np

a={'str':'test','num':1.2,'list':[1.1,[2.1]],'nan':float('nan'),'np':np.arange(1,5,dtype=np.uint8)}
jd.encode(a)
jd.decode(jd.encode(a))
d1=jd.encode(a,{'compression':'zlib','base64':1})
d1
jd.decode(d1,{'base64':1})

One can further save the JData annotated data into JSON or binary JSON (UBJSON) files using the jdata.save function, or loading JData-formatted data to Python using jdata.load

import jdata as jd
import numpy as np

a={'str':'test','num':1.2,'list':[1.1,[2.1]],'nan':float('nan'),'np':np.arange(1,5,dtype=np.uint8)}
jd.save(a,'test.json')
newdata=jd.load('test.json')
newdata

One can use loadt or savet to read/write JSON-based data files and loadb and saveb to read/write binary-JSON based data files. By default, JData annotations are automatically decoded after loading and encoded before saving. One can set {'encode': False} in the save functions or {'decode': False} in the load functions as the opt to disable further processing of JData annotations. We also provide loadts and loadbs for parsing a string-buffer made of text-based JSON or binary JSON stream.

PyJData supports multiple N-D array data compression/decompression methods (i.e. codecs), similar to HDF5 filters. Currently supported codecs include zlib, gzip, lz4, lzma, base64 and various blosc2 compression methods, including blosc2blosclz, blosc2lz4, blosc2lz4hc, blosc2zlib, blosc2zstd. To apply a selected compression method, one simply set {'compression':'method'} as the option to jdata.encode or jdata.save function; jdata.load or jdata.decode automatically decompress the data based on the _ArrayZipType_ annotation present in the data. Only blosc2 compression methods support multi-threading. To set the thread number, one should define an nthread value in the option (opt) for both encoding and decoding.

Reading JSON via REST-API

If a REST-API (URL) is given as the first input of load, it reads the JSON data directly from the URL and parse the content to native Python data structures. To avoid repetitive download, load automatically cache the downloaded file so that future calls directly load the locally cached file. If one prefers to always load from the URL without local cache, one should use loadurl() instead. Here is an example

import jdata as jd
data = jd.load('https://neurojson.io:7777/openneuro/ds000001');
data.keys()

Using JSONPath to access and query complex datasets

Starting from v0.6.0, PyJData provides a lightweight implementation JSONPath, a widely used format for query and access a hierarchical dict/list structure, such as those parsed by load or loadurl. Here is an example

import jdata as jd

data = jd.loadurl('https://raw.githubusercontent.com/fangq/jsonlab/master/examples/example1.json');
jd.jsonpath(data, '$.age')
jd.jsonpath(data, '$.address.city')
jd.jsonpath(data, '$.phoneNumber')
jd.jsonpath(data, '$.phoneNumber[0]')
jd.jsonpath(data, '$.phoneNumber[0].type')
jd.jsonpath(data, '$.phoneNumber[-1]')
jd.jsonpath(data, '$.phoneNumber..number')
jd.jsonpath(data, '$[phoneNumber][type]')
jd.jsonpath(data, '$[phoneNumber][type][1]')

The jd.jsonpath function does not support all JSONPath features. If more complex JSONPath queries are needed, one should install jsonpath_ng or other more advanced JSONPath support. Here is an example using jsonpath_ng

import jdata as jd
from jsonpath_ng.ext import parse

data = jd.loadurl('https://raw.githubusercontent.com/fangq/jsonlab/master/examples/example1.json');

val = [match.value for match in parse('$.address.city').find(data)]
val = [match.value for match in parse('$.phoneNumber').find(data)]

Downloading and caching _DataLink_ referenced external data files

Similarly to JSONLab, PyJData also provides similar external data file downloading/caching capability.

The _DataLink_ annotation in the JData specification permits linking of external data files in a JSON file - to make downloading/parsing externally linked data files efficient, such as processing large neuroimaging datasets hosted on http://neurojson.io, we have developed a system to download files on-demand and cache those locally. jsoncache.m is responsible of searching the local cache folders, if found the requested file, it returns the path to the local cache; if not found, it returns a SHA-256 hash of the URL as the file name, and the possible cache folders

When loading a file from URL, below is the order of cache file search paths, ranking in search order

   global-variable NEUROJSON_CACHE | if defined, this path will be searched first
   [pwd '/.neurojson']  	   | on all OSes
   /home/USERNAME/.neurojson	   | on all OSes (per-user)
   /home/USERNAME/.cache/neurojson | if on Linux (per-user)
   /var/cache/neurojson 	   | if on Linux (system wide)
   /home/USERNAME/Library/neurojson| if on MacOS (per-user)
   /Library/neurojson		   | if on MacOS (system wide)
   C:\ProgramData\neurojson	   | if on Windows (system wide)

When saving a file from a URL, under the root cache folder, subfolders can be created; if the URL is one of a standard NeuroJSON.io URLs as below

   https://neurojson.org/io/stat.cgi?action=get&db=DBNAME&doc=DOCNAME&file=sub-01/anat/datafile.nii.gz
   https://neurojson.io:7777/DBNAME/DOCNAME
   https://neurojson.io:7777/DBNAME/DOCNAME/datafile.suffix

the file datafile.nii.gz will be downloaded to /home/USERNAME/.neurojson/io/DBNAME/DOCNAME/sub-01/anat/ folder if a URL does not follow the neurojson.io format, the cache folder has the below form

   CACHEFOLDER{i}/domainname.com/XX/YY/XXYYZZZZ...

where XXYYZZZZ.. is the SHA-256 hash of the full URL, XX is the first two digit, YY is the 3-4 digits

In PyJData, we provide jdata.jdlink() function to dynamically download and locally cache externally linked data files. jdata.jdlink() only parse files with JSON/binary JSON suffixes that load supports. Here is a example

import jdata as jd

data = jd.load('https://neurojson.io:7777/openneuro/ds000001');
extlinks = jd.jsonpath(data, '$..anat.._DataLink_')  # deep-scan of all anatomical folders and find all linked NIfTI files
jd.jdlink(extlinks, {'regex': 'sub-0[12]_.*nii'})  # download only the nii files for sub-01 and sub-02
jd.jdlink(extlinks)                                # download all links

Utility

One can convert from JSON based data files (.json, .jdt, .jnii, .jmsh, .jnirs) to binary-JData based binary files (.bjd, .jdb, .bnii, .bmsh, .bnirs) and vice versa using command

python3 -mjdata /path/to/text/json/file.json # convert to /path/to/text/json/file.jdb
python3 -mjdata /path/to/text/json/file.jdb  # convert to /path/to/text/json/file.json
python3 -mjdata -h                           # show help info

Test

To see additional data type support, please run the built-in test using below command

python3 -m unittest discover -v test