cruftlesser

Yet another simple way to parse and generate XML


License
MIT
Install
npm install cruftlesser@1.0.5

Documentation

! This is a fork of the great cruftless library !

This fork fixed a few vulnerabilities, since the original maintainer of the library does not seem to be active anymore. The original can still be found under:

https://github.com/wspringer/cruftless

README

An XML builder / parser that tries to ease the common cases, allowing you to quickly build a model from your document structure and get a builder / parser for free.

Yet another XML binding framework?

I hate to say this, but: 'yes'. Or, perhaps: 'no'. Because Cruftless is not really an XML binding framework as you know it. It's almost more like Handlebars. But where Handlebars allows you to only generate documents, Cruftless also allows you to extract data from documents.

Building XML documents

Cruftless builds a simplified metamodel of your XML document, and it's not based on a DOM API. So, if this is the XML document:

<person>
  <name>John Doe</name>
  <age>16</age>
</person>

Then, using the builder API, Cruftless allows you to build a model of your document like this:

const { element, attr, text } = require('cruftlesser')();

let el = element("person").content(
  element("name").content(text().value("John Doe")),
  element("age").content(text().value(16))
);

… and then to turn it back into XML, you'd use the toXML() operation:

el.toXML(); // ⇨ '<person>\r\n  <name>John Doe</name>\r\n  <age>16</age>\r\n</person>'

… or the toDOM() operation instead to return a DOM representation of the document:

el.toDOM();

Binding

Now, this itself doesn't seem all that useful. Where it gets useful is when you start adding references to your document model:

el = element("person").content(
  element("name").content(text().bind("name")),
  element("age").content(text().bind("age"))
);

Now, if you want to generate different versions of your XML document for different persons, you can simply pass in an object with name and age properties:

let xml = el.toXML({ name: "John Doe", age: "16" }); // ⇨ '<person>\r\n  <name>John Doe</name>\r\n  <age>16</age>\r\n</person>'

But the beauty is, it also works the other way around. If you have your model with binding expressions, then you're able to extract data from XML like this:

el.fromXML(xml); // ⇨ { name: 'John Doe', age: '16' }

Less tedious, please

I hope you can see how this is useful. However, I also hope you can see that this is perhaps not ideal. I mean, it's nice that you're able to build a model of XML, but in many cases, you already have the snippets of XML that you need to populate with data. So, the question is if there is an easier way to achieve the same, if you already have snippets of XML. Perhaps not surprisingly, there is:

let template = `<person>
  <name>{{name}}</name>
  <age>{{age}}</age>
</person>`;

let { parse } = require('cruftlesser')();

el = parse(template);
console.log(el.toXML({ name: "Jane Doe", age: "18" }));
β‡’ <person>
β‡’   <name>Jane Doe</name>
β‡’   <age>18</age>
β‡’ </person>

Additional metadata

The example above is rather simple. However, Cruftless allows you also deal with more complex cases. And not only that, it also allows you to set additional metadata on binding expressions, using the pipe symbol. In the template below, we're binding <person/> elements inside a <persons/> element to a property persons, and we're inserting every occurence of it into the persons array. The processing instruction annotation might be feel a little awkward at first. There are other ways to define the binding, including one that requires using attributes of particular namespace. Check the test files for examples.

template = parse(`<persons>
  <person><?bind persons|array?>
    <name>{{name|required}}</name>
    <age>{{age|integer|required}}</age>
  </person>
</persons>`);

// Note that because of the 'integer' modifier, integer values are
// now automatically getting transfered from and to strings.

console.log(
  template.toXML({
    persons: [
      { name: "John Doe", age: 16 },
      { name: "Jane Doe", age: 18 },
    ],
  })
);
β‡’ <persons>
β‡’   <person>
β‡’     <name>John Doe</name>
β‡’     <age>16</age>
β‡’   </person>
β‡’   <person>
β‡’     <name>Jane Doe</name>
β‡’     <age>18</age>
β‡’   </person>
β‡’ </persons>

You can add your own value types to convert from and to the string literals included in the XML representation.

const { element, attr, text, parse } = require('cruftlesser')({
  types: {
    zeroOrOne: {
      type: "boolean",
      from: (str) => str == "1",
      to: (value) => (value ? "1" : "0"),
    },
  },
});

template = parse(`<foo>{{value|zeroOrOne}})</foo>`);
console.log(template.toXML({ value: true }));
console.log(template.toXML({ value: false }));
β‡’ <foo>1</foo>
β‡’ <foo>0</foo>

The same works with attributes as well:

template = parse(`<foo bar="{{value|zeroOrOne}}"/>`);
console.log(template.toXML({ value: true }));
console.log(template.toXML({ value: false }));
β‡’ <foo bar="1"/>
β‡’ <foo bar="0"/>

Sometimes, it's still useful to be able to access the raw field values, ignoring the type annotations.

To get the actual data:

// The second argument defaults to false, so might as well leave it out
console.log(template.fromXML("<foo bar='1'/>", false));
β‡’ { value: true }

To get the raw data:

console.log(template.fromXML("<foo bar='1'/>", true));
β‡’ { value: '1' }

Alternative notation

The <!--persons|array--> way of annotating an element is not the only way you are able to add metadata. Another way to add metadata to elements is by using one of the reserved attributes prefixed with c-.

template = parse(`<persons>
  <person c-bind="persons|array">
    <name>{{name|required}}</name>
    <age>{{age|integer|required}}</age>
  </person>
</persons>`);

console.log(
  template.toXML({
    persons: [
      { name: "John Doe", age: 16 },
      { name: "Jane Doe", age: 18 },
    ],
  })
);
β‡’ <persons>
β‡’   <person>
β‡’     <name>John Doe</name>
β‡’     <age>16</age>
β‡’   </person>
β‡’   <person>
β‡’     <name>Jane Doe</name>
β‡’     <age>18</age>
β‡’   </person>
β‡’ </persons>

If you hate the magic c- prefixed attributes, then you can also a slightly less readable but admittedly more correct XML namespace:

template = parse(`<persons>
  <person xmlns:c="https://github.com/wspringer/cruftless" c:bind="persons|array">
    <name>{{name|required}}</name>
    <age>{{age|integer|required}}</age>
  </person>
</persons>`);

console.log(
  template.toXML({
    persons: [
      { name: "John Doe", age: 16 },
      { name: "Jane Doe", age: 18 },
    ],
  })
);
β‡’ <persons>
β‡’   <person>
β‡’     <name>John Doe</name>
β‡’     <age>16</age>
β‡’   </person>
β‡’   <person>
β‡’     <name>Jane Doe</name>
β‡’     <age>18</age>
β‡’   </person>
β‡’ </persons>

Conditionals

There may be times when you want to exclude entire sections of an XML structure if a particular condition is met. Cruftless has some basic support for that, albeit limited. You can set conditions on elements, using the c-if attribute. In that case, the element will only be included in case the expression of the c-if attribute is evaluating to something else than undefined or null.

template = parse(`<foo><bar c-if="a">text</bar></foo>`);

template.toXML({}); // ⇨ '<foo/>'
template.toXML({ a: null }); // ⇨ '<foo/>'
template.toXML({ a: void 0 }); // ⇨ '<foo/>'
template.toXML({ a: 3 }); // ⇨ '<foo>\r\n  <bar>text</bar>\r\n</foo>'

If your template contains variable references, and the data structure you are passing in does not contain these references, then β€” instead of generating the value undefined, Cruftless will drop the entire element. In fact, if a deeply nested element contains references to variable, and that variable is not defined, then it will not only drop that element, but all elements that included that element referring to a non-existing variable.

template = parse(`<level1>
  <level2 b="{{b}}">
    <level3>{{a}}</level3>
  </level2>
</level1>`);

console.log(template.toXML({ b: 2 }));
β‡’ <level1>
β‡’   <level2 b="2"/>
β‡’ </level1>
console.log(template.toXML({ b: 2, a: 3 }));
β‡’ <level1>
β‡’   <level2 b="2">
β‡’     <level3>3</level3>
β‡’   </level2>
β‡’ </level1>
console.log(template.toXML({ a: 3 }));
β‡’ <level1>
β‡’   <level2>
β‡’     <level3>3</level3>
β‡’   </level2>
β‡’ </level1>

CDATA

Your XML documents might contain CDATA sections. Cruftless will treat those like ordinary text nodes. That is, if you have an element that has a text node bound to a variable, then it will resolve those values regardless of the fact if the incoming XML document has a text node or a CDATA node.

template = parse(`<person>{{name}}</person>`);
console.log(template.fromXML(`<person>Alice</person>`));
β‡’ { name: 'Alice' }
console.log(template.fromXML(`<person><![CDATA[Alice]]></person>`));
β‡’ { name: 'Alice' }

However, if you would produce XML, then β€” by default β€”Β it will always produce a text node:

console.log(template.toXML({ name: "Alice" }));
β‡’ <person>Alice</person>

That is, unless you specifiy a cdata option in your binding:

template = parse(`<person>{{name|cdata}}</person>`);
console.log(template.toXML({ name: "Alice" }));
β‡’ <person>
β‡’   <![CDATA[Alice]]>
β‡’ </person>

JSON-ish Schema (incomplete, subject to change)

Since Cruftless has all of the metadata of your XML document and how it binds to your data structures at its disposal, it also allows you to generate a 'schema' of the data structure it expects.

let schema = template.descriptor();
console.log(JSON.stringify(schema, null, 2));
β‡’ {
β‡’   "type": "object",
β‡’   "keys": {
β‡’     "name": {
β‡’       "type": "string"
β‡’     }
β‡’   }
β‡’ }

The schema will include additional metadata you attached to expressions:

template = parse(`<person>
  <name>{{name|sample:Wilfred}}</name>
  <age>{{age|integer|sample:45}}</age>
</person>`);

schema = template.descriptor();
console.log(JSON.stringify(schema, null, 2));
β‡’ {
β‡’   "type": "object",
β‡’   "keys": {
β‡’     "name": {
β‡’       "type": "string",
β‡’       "sample": "Wilfred"
β‡’     },
β‡’     "age": {
β‡’       "type": "integer",
β‡’       "sample": 45
β‡’     }
β‡’   }
β‡’ }

RelaxNG Schema

Since Cruftless captures the structure of the XML document, it's also able to generate an XML Schema representation of the document structure. Only, it's not relying on XML Schema. It's using RelaxNG instead. If you never heard of RelaxNG before: think of it as a more readable better version of XML Schema, without the craziness.

So based on the template above, this would give you the RelaxNG schema:

const { relaxng } = require('cruftlesser')();

console.log(relaxng(template));
β‡’ <grammar datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes" xmlns="http://relaxng.org/ns/structure/1.0">
β‡’   <start>
β‡’     <element name="person">
β‡’       <optional>
β‡’         <element name="name">
β‡’           <data type="string"/>
β‡’         </element>
β‡’       </optional>
β‡’       <optional>
β‡’         <element name="age">
β‡’           <data type="integer"/>
β‡’         </element>
β‡’       </optional>
β‡’     </element>
β‡’   </start>
β‡’ </grammar>

Markdown generated from ./README.js.md by RunMD Logo