xs2jIter

XML Stream to JSON Iterator

Convert XML to JSON and iterate over its top-level elements.

Índex

• Abstract
• Usage
  • As a command-line tool
    • Install
    • Console Usage
    • Usage from other languages
  • As a library
    • Install
    • Syntax
      • Parameters:
      • Return value
        • header
      • Array-Like methods:
    • Example
• Contributing

Abstract

XML is a polymorphic data format able to mix multiple kinds of data with different structures in a single file (or stream). But, in fact, it is often used to share simple arrays of objects with the same or similar structure which are more easily handled in simpler formats like JSON.

Converting XML into JSON directly is almost impossible because XML is actually a document format and has too many intrincated semantics (attributes, text chunks...) which doesn't fit fine in simpler formats like JSON.

There are many approaches to do that over the net, bust most of them are structure-agnostic so they doesn't return any data until the whole XML document is fully parsed.

Despite the annoying delay, that can also be a problem while parsing huge files of abovementioned array-like files due to its rapidly increase of memory requirements or even, in worst cases, ending up in out of memory issues when, in fact, you only were trying to convert long list of items to JSON array.

With xs2jIter you can stream those array-like files and igerate over each asynchronously JSON encoded item not bothering on memory issues bacause conversion is done asynchronously.

To do that, xs2jIter expects an XML as its only parameter which can be provided both as string or thought a stream and returns an iterable object you can simply iterate (in a synchronous-like way) while input data is actually parsed asynchronously.

Parsing is done though node-expat and is expected to have an optional DTD and a single top-level tag containing the array of tags (with fully free structure) over which we will iterate.

That is:

• 0-Level: Single container (document) tag.
  • Tag name doesn't care.
  • Tag attributes are threated as heading information and can be accessed tought header attribute of the returned iterable.
  • If there are more top-level elements, they will be ignored.
• 1-Level: Array of tags:
  • Tag name:
    • Doesn't care (but usually will be the same).
    • Can be read thought "@" property (only for 1-Level tags).

Conversion to JSON is done following the specification of converting-between-xml-and-json.html

NOTE: The only change done to that specification was adding a "@" attribute at top level of each object with the tag name. This way, even XMLs with mixed tag types can be parsed distinguishing each tag type. In case of arrays, it is created as non enumerable property to not broke its behaviour.

Usage

As a command-line tool

Even designed as a javascript library to handle large XML files asynchronously (see usage as a library) xs2jiter can also be used as command line tool for data analysis and inspection purposes or as input for non javascript languages.

Install

npm install -g xs2jiter

Console Usage

$ xs2jiter.js -h

  Usage: xs2jiter [options] [ inputFile ] [, outputFile ]

  XML Stream to JSON Iterator - Convert XML to JSON and iterate over its top-level elements.

  Options:

    -h, --help              output usage information
    -V, --version           output the version number
    -p, --pretty            Output prettyfied JSON chunks (default)
    -r, --raw               Output raw JSON chunks
    -b, --base64            Output base64-encoded JSON chunks
    -n, --noExtraNewline    Don't output extra newline characters
    -a, --Array             Generate valid JSON-Array output.
    -i --inspect            Inspect data structure
    -D --iDeep <deep>       Maximum sample values per item (default 5)
    -A --iPick <addresses>  Adresses to pick whole distinct values (ie: foo.bar,foo.baz) on inspection

Usage from other languages

As you can figure out, to take advantage of xs2jiter from other languages, you can simply invoke its cli tool and parse its output.

But... How do you handle asincrony?

You actually don't need to do that: xs2jiter console tool outputs a blank line after each item to make easier to visually detect boundarys. For automated parsing you can disable it with -n modifier and to avoid fake positives because of possible newlines in json data, use base64 output to ensure the only newlines are actual register separators.

Example written in PHP:

<?php
$p = popen('xs2jiter -bn /path/to/file.xml', 'r');
while (false !== $str = fgets($p)) {
    $data = json_decode(base64_decode($str), true);
    // Do something with $data
};
?>

As a library

Install

npm install --save xs2jiter

Syntax

var x2j = require("xs2jiter");
var data = x2j(xml [, maxBufferLength]);
    // data        -> Iterable (once) over the whole items.
    // data.header -> Header information (container tag attributes).

Parameters:

• xml: XML stream or string.
• maxBufferLength: Specify the internal elements buffer length (Default is 50).

NOTE: Default value for maxBufferLength will be usually fine. But lower values may help to limit memory usage when having big objects.

On the other hand, greater values may increase speed if your parsing time varies (typically because too different object lengths)

Return value

Singleton iterable with header property.

for (var j of data) {...};

...but it has also additional interesting properties and methods:

header

Header information (attributes of the document container tag).

Array-Like methods:

Following array-like methods are also supprted. They work like its Array equivalents but returns an iterator instead of an array. For more information see abuffer documentation.

• buff.map(cbk [, thisArg])

• buff.filter(cbk [,thisArg])

Example

var x2j = require("xs2jiter");
var Fs = require("fs");
var xml = Fs.createReadStream("path/to/file.xml");

var data = x2j(xml);

console.log("=================================");
console.log(data.header);
console.log("=================================");

for (var j of data) {
    console.log("---------------------------------");
    console.log(JSON.stringify(j, null, 4));
};

Contributing

If you are interested in contributing with this project, you can do it in many ways:

• Creating and/or mantainig documentation.

• Implementing new features or improving code implementation.

• Reporting bugs and/or fixing it.

• Sending me any other feedback.

• Whatever you like...

Please, contact-me, open issues or send pull-requests thought this project GIT repository