[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: How to represent mixed content in JSON and JSON Schema?

• From: doug.duboulay@gmail.com
• To: xml-dev@lists.xml.org
• Date: Thu, 12 Jul 2018 14:17:48 +0800

I recently used pandoc to do a Latex to DocBook conversion.
    http://pandoc.org/
Turns out, IIUC, that they use a JSON representation of the document (mixed) 
content as the intermediate format for conversion between all of the target 
specifications. 
I see there is a proto schema here:
    https://gist.github.com/tarleb/bcb4b31e35263fcbc6cf7e533670c1cf

I don't know that they have an explicit slot for a bidirectional attribute, 
but since pandoc also reads DocBook  and DocBook has a "dir" attribute on some 
elements, it might be worth a  look under the hood...



On Wednesday, 11 July 2018 3:57:22 PM AWST Costello, Roger L. wrote:
> Hi Folks,
> 
> I am converting an XML Schema to a JSON Schema.
> 
> The XML Schema declares an <example> element that has bidirectional-override
> (BDO) elements intermingled with text, e.g.,
> 
> <example>Some text <bdo dir="rtl">reverse this text <bdo dir="ltr">not this
> text</bdo></bdo> finishing text</example>
> 
> rtl = right-to-left
> ltr = left-to-right
> 
> As you can see, the <example> element has mixed content, as does the <bdo>
> element.
> 
> At the bottom of this message is the XML Schema.
> 
> JSON and JSON Schema do not support mixed content, nor do they support
> attributes.
> 
> How to represent the XML in JSON and JSON Schema?
> 
> Here's what I am thinking: Mixed content is essentially an ordered sequence
> of text, element, text, element, ...
> 
> JSON arrays are intended to represent ordered sequences. So, I am thinking
> that mixed content can be represented in JSON with an array. The above
> <example> element can be expressed this way in JSON:
> 
> {
>   "example": [
>       {"text": "Some text "},
>       {"bdo":
>           [
>               {"dir": "rtl"},
>               {"text": "reverse this text "},
>               {"bdo":
>                   [
>                      {"dir": "ltr"},
>                      {"text": "not this text"}
>                   ]
>               }
>           ]
>       },
>       {"text": " finishing text"}
>   ]
> }
> 
> Notice that the content of the items with mixed content (example and bdo)
> are represented with arrays.
> 
> At the bottom of this message is the JSON Schema.
> 
> Is there a better way to represent mixed content in JSON? What makes one way
> better than another? One way can be processed easier, more efficiently than
> another?  /Roger
> 
> XML Schema
> 
> <xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
> 
>     <xs:element name="example" type="BDOMixedContent" />
> 
>     <xs:complexType name="BDOMixedContent" mixed="true">
>         <xs:choice minOccurs="0" maxOccurs="unbounded">
>             <xs:element name="bdo" type="BDOMixedContent" />
>         </xs:choice>
>         <xs:attribute name="dir">
>             <xs:simpleType>
>                 <xs:restriction base="xs:string">
>                     <xs:enumeration value="ltr" />
>                     <xs:enumeration value="rtl" />
>                 </xs:restriction>
>             </xs:simpleType>
>         </xs:attribute>
>     </xs:complexType>
> 
> </xs:schema>
> 
> 
> JSON Schema
> 
> {
>     "$schema": "http://json-schema.org/draft-07/schema#",
>     "definitions": {
>         "BDOMixedContent": {
>             "type": "array",
>             "items": {
>               "anyOf": [
>                        {
>                            "type": "object",
>                            "properties": {
>                                  "text": {"type": "string"}
>                            },
>                            "additionalProperties": false
>                        },
>                        {
>                            "type": "object",
>                            "properties": {
>                                  "bdo": {"$ref":
> "#/definitions/BDOMixedContent"} },
>                            "additionalProperties": false
>                        },
>                        {
>                            "type": "object",
>                            "properties": {
>                                  "dir": {"type": "string", "enum": ["ltr",
> "rtl"]} },
>                            "additionalProperties": false
>                        }
>                 ]
>             }
>         }
>     },
>     "type": "object",
>     "properties": {
>           "example": {"$ref": "#/definitions/BDOMixedContent"}
>     },
>     "required": [ "example" ],
>     "additionalProperties": false
> }

• References:
  • How to represent mixed content in JSON and JSON Schema?
    • From: "Costello, Roger L." <costello@mitre.org>