This lib parses markdown into a simplified Abstract Syntax tree.
npm install docs-and-graphsThis lib parses markdown into a simplified Abstract Syntax tree.
Several Note-Taking apps are 'node-based,' 'markdown based,' etc.
I use Markdown, and I cannot get used to outlines. However, I recognize the benefits of having node-based systems, where
you can reference a specific node from any other node.
So my question was: How can I have these nodes, and still use Markdown?
After considering this, I realized that Markdown has some structure. It has headers that can be inside other headings
and lists that can be inside other lists. These are the nodes this library generates.
Say you have the following markdown
``markdown
---
hello: world
---
Some text under Heading 1
Text that has (inline::variables)
- Tana and logseq likes
- embedded nodes
`
The lib
`js
import { simpleAst } from 'docs-and-graphs'
const json = simpleAst(yourMarkdownString)
// With options
const json = simpleAst(yourMarkdownString, {
normalize: false, // Remove prefixes like # from tags, ^ from block IDs
inlineAsArray: false, // Return inline fields as arrays vs nested objects
includePosition: false, // Include source position metadata
maxDepth: null // Flatten headers deeper than this level
})
`
will produce the following Json
`json
{
"type": "root",
"depth": 0,
"data": [
{
"hello": "world"
}
],
"children": [
{
"type": "block",
"value": "# Heading 1",
"depth": 1,
"children": [
{
"type": "text",
"value": "Some text under Heading 1"
},
{
"type": "text",
"data": [
{
"inline": "variables"
}
],
"value": "Text that has (inline::variables)"
},
{
"type": "block",
"value": "## Inline elements",
"depth": 2,
"children": [
{
"type": "outline",
"ordered": false,
"children": [
{
"type": "outline",
"value": "Tana and logseq likes "
},
{
"type": "outline",
"ordered": false,
"children": [
{
"type": "outline",
"value": "embedded nodes"
}
]
}
]
}
]
}
]
}
]
}
`
, removes prefixes from parsed elements:
- Tags: #tag becomes tag
- Block IDs: ^block-id becomes block-id
- Also applies text normalization (trimming, etc.)$3
Controls how inline fields like subject :: inline :: field are parsed:
- false: Creates nested objects {subject: {inline: "field"}}
- true: Returns arrays ["subject", "inline", "field"]$3
When true, includes source position metadata (line/column numbers) from the markdown parser. Useful for debugging or source mapping.$3
Limits header nesting depth by flattening deeper headers:
- null: No limit (default behavior)
- 2: Headers deeper than level 2 become level 2 siblings
- Content and inline fields stay with their original headers when flattenedExample with
maxDepth: 2:
`markdown
Level 1
Level 2
$3
#### Level 4 ← becomes level 2
``I use this structure to later produce RDF using
a vault-triplifier, but you can use it for whatever you want.