Wolfram/Parser | Paclet Repository

Wolfram`Parser`

MarkdownInlineParse

MarkdownInlineParse [source]
	parses inline markdown source (a String — the span-level content of a single paragraph) into a flat List of inline-atom Association s, each carrying a "Type" discriminator ( "Text" , "Bold" , "Italic" , "Code" , "Link" , "MathInline" , …) plus per-shape payload keys.

Details and Options

▪

The result is a flat

List

Association

s. Prose runs are

"Type""Text","Text"str

; emphasis is

"Bold"

"Italic"

"BoldItalic"

; code is

"Code"

"LiteralCode"

"HtmlCode"

; math is

"MathInline"

"MathDisplay"

; the paired references are

"Link"

and

"Image"

; and

"Sub"

"Sup"

"Strike"

cover the remaining spans.

▪

A span's body — the

"Children"

of an emphasis / sub / sup / strike atom, or the

"Label"

of a link — is itself a

List

of inline atoms, because the captured body is re-parsed recursively. So

**bold$x$**

nests a

"MathInline"

atom inside the

"Bold"

▪

Adjacent

"Text"

atoms are merged, so a contiguous prose run is one atom, not one atom per character.

▪

Overlapping openers resolve by PEG order, longest first:

***

before

, double backtick before single,

before

, and the HTML

<sub>

<sup>

before their single-character Pandoc twins.

▪

Underscore emphasis follows CommonMark word-boundary rules:

_em_

opens emphasis at a word boundary, but an underscore between word characters (

snake_case

) stays literal.

▪

A literal

...

inside a

"Text"

run becomes the Unicode ellipsis character;

...

inside a code span or math span is left verbatim.

▪

A backslash before ASCII punctuation (

) yields that punctuation as literal text rather than opening a span.

▪

An unclosed delimiter is not an error — the opener and its trailing text stay as literal

"Text"

▪

MarkdownInlineParse

handles inline constructs only. Whole-document structure (frontmatter, headings, code fences, thematic breaks, paragraph splitting) is

MarkdownParse

's job; feed a document's

"Prose"

block text through this function to resolve its spans.

▪

MarkdownInlineParse

is the wrapper around the

MarkdownInlineParser

combinator: it runs

Parse

and then merges text runs, applies the ellipsis and underscore-emphasis passes, and re-parses span children. On a parse failure it returns the

Failure

object unchanged.

▪

The grammar and its post-processing passes are walked through in the

Markdown inline parser tutorial

Examples

(15)

Basic Examples

(1)

Plain prose is a single

"Text"

atom:

In[1]:=

MarkdownInlineParse

["plain text"]

Out[1]=

{TypeText,Textplain text}

Emphasis and code spans become their own atoms, with the surrounding prose in

"Text"

runs:

In[2]:=

MarkdownInlineParse

["**bold** and `code`"]

Out[2]=

{TypeBold,Children{TypeText,Textbold},TypeText,Text and ,TypeCode,Codecode}

A link's label is parsed recursively — here it is a code span, not plain text:

In[3]:=

MarkdownInlineParse

["[`Range`](paclet:ref/Range)"]

Out[3]=

{TypeLink,Urlpaclet:ref/Range,Label{TypeCode,CodeRange}}

Scope

(8)



Properties & Relations

(3)



Possible Issues

(2)



Neat Examples

(1)



SeeAlso

▪

▪

▪

▪

RelatedGuides

▪

WolframParser