mdld-parse 0.1.0 → 0.2.1

package/README.md

@@ -1,28 +1,36 @@
-# MD-LD Parser
+# MD-LD Parse
 
-A standards-compliant parser for **MD-LD (Markdown-Linked Data)** — a human-friendly RDF authoring format that extends Markdown with semantic annotations.
+**Markdown-Linked Data (MD-LD)** — a human-friendly RDF authoring format that extends Markdown with semantic annotations.
+
+[NPM](https://www.npmjs.com/package/mdld-parse)
+
+[Website](https://mdld.js.org)
 
 ## What is MD-LD?
 
 MD-LD allows you to author RDF graphs directly in Markdown using familiar syntax:
 
 ```markdown
----
-"@context":
-  "@vocab": "http://schema.org/"
-"@id": "#doc"
-"@type": Article
----
+# My Note {=urn:mdld:my-note-20251231 .NoteDigitalDocument}
 
-# My Article {#article typeof="Article"}
+[ex]{: http://example.org/}
 
-Written by [Alice Johnson](#alice){property="author" typeof="Person"}
+Written by [Alice Johnson](=ex:alice){author .Person}
 
-[Alice](#alice) works at [Tech Corp](#company){rel="worksFor" typeof="Organization"}
+## Alice's biography {=ex:alice}
+
+[Alice](ex:alice){name} works at [Tech Corp](=ex:tech-corp){worksFor .Organization}
 ```
 
 This generates valid RDF triples while remaining readable as plain Markdown.
 
+```n-quads
+<urn:mdld:my-note-20251231> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://schema.org/NoteDigitalDocument> .
+<urn:mdld:my-note-20251231> <http://schema.org/author> <http://example.org/alice> .
+<http://example.org/alice> <http://schema.org/name> "Alice" .
+<http://example.org/alice> <http://schema.org/worksFor> <http://example.org/tech-corp> .
+```
+
 ## Architecture
 
 ### Design Principles
@@ -32,6 +40,8 @@ This generates valid RDF triples while remaining readable as plain Markdown.
 3. **Standards Compliant** — Outputs RDF quads compatible with RDFa semantics
 4. **Markdown Native** — Plain Markdown yields minimal but valid RDF
 5. **Progressive Enhancement** — Add semantics incrementally via attributes
+6. **BaseIRI Inference** — Automatically infers baseIRI from document structure
+7. **Default Vocabulary** — Provides default vocabulary for common properties, extensible via options
 
 ### Stack Choices
 
@@ -40,8 +50,7 @@ This generates valid RDF triples while remaining readable as plain Markdown.
 We implement a **minimal, purpose-built parser** for maximum control and zero dependencies:
 
 - **Custom Markdown tokenizer** — Line-by-line parsing of headings, lists, paragraphs, code blocks
-- **Inline attribute parser** — Pandoc-style `{#id .class key="value"}` attribute extraction
-- **YAML-LD frontmatter parser** — Minimal YAML subset for `@context` and `@id` parsing
+- **Inline attribute parser** — Pandoc-style `{=iri .class key="value"}` attribute extraction
 - **RDF quad generator** — Direct mapping from tokens to RDF/JS quads
 
 **Why custom?**
@@ -79,9 +88,7 @@ Markdown Text
     ↓
 [Custom Tokenizer] — Extract headings, lists, paragraphs, code blocks
     ↓
-[YAML-LD Parser] — Extract frontmatter @context and @id
-    ↓
-[Attribute Parser] — Parse {#id property="value"} from tokens
+[Attribute Parser] — Parse {=iri .class key="value"} from tokens
     ↓
 [Inline Parser] — Extract [text](url){attrs} spans
     ↓
@@ -101,17 +108,6 @@ The zero-dependency design provides:
 3. **Predictable performance** — Linear time complexity, bounded memory
 4. **Easy integration** — Works in Node.js, browsers, and edge runtimes
 
-### Performance Profile
-
-| Document Size | Peak Memory | Parse Time |
-| ------------- | ----------- | ---------- |
-| 10 KB         | ~100 KB     | <2ms       |
-| 100 KB        | ~500 KB     | <20ms      |
-| 1 MB          | ~2 MB       | <100ms     |
-| 10 MB         | ~10 MB      | <1s        |
-
-_Measured on modern JavaScript engines. Actual performance depends on document structure._
-
 ## Installation
 
 ### Node.js
@@ -121,12 +117,11 @@ npm install mdld-parse
 ```
 
 ```javascript
-import { parseMDLD } from "mdld-parse";
+import { parse } from "mdld-parse";
 
-const markdown = `# Hello\n{#doc typeof="Article"}`;
-const quads = parseMDLD(markdown, {
-	baseIRI: "http://example.org/doc",
-});
+const markdown = `# Hello {=urn:mdld:hello .Article}`;
+const result = parse(markdown);
+const quads = result.quads;
 ```
 
 ### Browser (via CDN)
@@ -141,42 +136,63 @@ const quads = parseMDLD(markdown, {
 </script>
 
 <script type="module">
-	import { parseMDLD } from "mdld-parse";
-	// use parseMDLD...
+	import { parse } from "mdld-parse";
+	// use parse...
 </script>
 ```
 
 ## API
 
-### `parseMDLD(markdown, options)`
+### `parse(markdown, options)`
 
-Parse MD-LD markdown and return RDF quads.
+Parse MD-LD markdown and return parsing result.
 
 **Parameters:**
 
 - `markdown` (string) — MD-LD formatted text
 - `options` (object, optional):
-  - `baseIRI` (string) — Base IRI for relative references (default: `''`)
-  - `defaultVocab` (string) — Default vocabulary (default: `'http://schema.org/'`)
+  - `baseIRI` (string) — Base IRI for relative references
+  - `context` (object) — Additional context to merge with default context
   - `dataFactory` (object) — Custom RDF/JS DataFactory (default: built-in)
 
-**Returns:** Array of RDF/JS Quads
+**Returns:** Object containing:
+- `quads` — Array of RDF/JS Quads
+- `origin` — Object with `blocks` and `quadIndex` for serialization
+- `context` — Final context used for parsing
+
+### `serialize({ text, diff, origin, options })`
+
+Serialize RDF changes back to markdown with proper positioning.
+
+**Parameters:**
+
+- `text` (string) — Original markdown text
+- `diff` (object) — Changes to apply:
+  - `add` — Array of quads to add
+  - `delete` — Array of quads to remove
+- `origin` (object) — Origin object from parse result
+- `options` (object, optional) — Additional options
+
+**Returns:** Object containing:
+- `text` — Updated markdown text
+- `origin` — Updated origin object
 
 ```javascript
-const quads = parseMDLD(
+const result = parse(
 	`
-# Article Title
-{#article typeof="Article"}
+# Article Title {=ex:article .Article}
 
-Written by [Alice](#alice){property="author"}
+Written by [Alice](ex:alice) {ex:author}
 `,
 	{
 		baseIRI: "http://example.org/doc",
-		defaultVocab: "http://schema.org/",
+		context: {
+      '@vocab': 'http://schema.org/',
+    },
 	}
 );
 
-// quads[0] = {
+// result.quads[0] = {
 //   subject: { termType: 'NamedNode', value: 'http://example.org/doc#article' },
 //   predicate: { termType: 'NamedNode', value: 'http://www.w3.org/1999/02/22-rdf-syntax-ns#type' },
 //   object: { termType: 'NamedNode', value: 'http://schema.org/Article' },
@@ -184,78 +200,51 @@ Written by [Alice](#alice){property="author"}
 // }
 ```
 
-### Batch Processing
-
-For multiple documents, process them sequentially:
-
-```javascript
-const documents = [markdown1, markdown2, markdown3];
-const allQuads = documents.flatMap((md) =>
-	parseMDLD(md, { baseIRI: "http://example.org/" })
-);
-```
-
 ## Implementation Details
 
 ### Subject Resolution
 
 MD-LD follows a clear subject inheritance model:
 
-1. **Root subject** — Declared in YAML-LD `@id` field
-2. **Heading subjects** — `## Title {#id typeof="Type"}`
-3. **Inline subjects** — `[text](#id){typeof="Type"}`
+1. **Root subject** — Declared in the first heading of the document or inferred it's text content
+2. **Heading subjects** — `## Title {=ex:title .Type}`
+3. **Inline subjects** — `[text](=ex:text) {.Type}`
 4. **Blank nodes** — Generated for incomplete triples
 
 ```markdown
-# Document
-
-{#doc typeof="Article"}
+# Document {=urn:mdld:doc .Article}
 
-## Section
+## Section 1 {=urn:mdld:sec1 .Section} 
 
-{#sec1 typeof="Section"}
+[Text] {name} ← property of sec1
 
-[Text]{property="name"} ← property of #sec1
+Back to [doc](=urn:mdld:doc) {hasPart}
 ```
 
-### Property Mapping
-
-| Markdown                | RDF Predicate                                                                   |
-| ----------------------- | ------------------------------------------------------------------------------- |
-| Top-level H1 (no `#id`) | `rdfs:label` on root subject                                                    |
-| Heading with `{#id}`    | `rdfs:label` on subject                                                         |
-| First paragraph         | `dct:description` on root                                                       |
-| `{property="name"}`     | Resolved via `@vocab` (e.g., `schema:name`)                                     |
-| `{rel="author"}`        | Resolved via `@vocab` (e.g., `schema:author`)                                   |
-| Code block              | `schema:SoftwareSourceCode` with `schema:programmingLanguage` and `schema:text` |
-
 ### List Handling
 
-```markdown
-- [Item 1]{property="item"}
-- [Item 2]{property="item"}
+```markdown {item}
+- Item 1
+- Item 2
 ```
 
 Creates **multiple triples** with same predicate (not RDF lists):
 
 ```turtle
-<#doc> schema:item "Item 1" .
-<#doc> schema:item "Item 2" .
+<subject> schema:item "Item 1" .
+<subject> schema:item "Item 2" .
 ```
 
-For RDF lists (`rdf:List`), use `@inlist` in generated HTML.
-
 ### Code Block Semantics
 
-Fenced code blocks are automatically mapped to `schema:SoftwareSourceCode`:
-
 ```markdown
-\`\`\`sparql {#query-1}
-SELECT * WHERE { ?s ?p ?o }
+\`\`\`sparql {=ex:query-1 .SoftwareSourceCode}
+SELECT \* WHERE { ?s ?p ?o }
 \`\`\`
 ```
 
 Creates:
+
 - A `schema:SoftwareSourceCode` resource (or custom type via `typeof`)
 - `schema:programmingLanguage` from the info string (`sparql`)
 - `schema:text` with the raw source code
@@ -263,112 +252,32 @@ Creates:
 
 This enables semantic queries like "find all SPARQL queries in my notes."
 
-### Blank Node Strategy
-
-Blank nodes are created for:
-
-1. Task list items without explicit `#id`
-2. Code blocks without explicit `#id`
-3. Inline `typeof` without `id` when used with `rel`
-
-## Testing
-
-```bash
-npm test
-````
-
-Tests cover:
-
-- ✅ YAML-LD frontmatter parsing
-- ✅ Subject inheritance via headings
-- ✅ Property literals and datatypes (`property`, `datatype`)
-- ✅ Object relationships (`rel` on links)
-- ✅ Blank node generation (tasks, code blocks)
-- ✅ List mappings (repeated properties)
-- ✅ Code block semantics (`SoftwareSourceCode`)
-- ✅ Semantic links in lists (`hasPart` TOC)
-- ✅ Cross-references via fragment IDs
-- ✅ Minimal Markdown → RDF (headings, paragraphs)
-
 ## Syntax Overview
 
 ### Core Features
 
-**YAML-LD Frontmatter** — Define context and root subject:
-
-```yaml
----
-"@context":
-  "@vocab": "http://schema.org/"
-"@id": "#doc"
-"@type": Article
----
-```
-
 **Subject Declaration** — Headings create typed subjects:
 
 ```markdown
-## Alice Johnson {#alice typeof="Person"}
+## Alice Johnson {=ex:alice .Person}
 ```
 
 **Literal Properties** — Inline spans create properties:
 
 ```markdown
-[Alice Johnson]{property="name"}
-[30]{property="age" datatype="xsd:integer"}
+[Alice Johnson] {name}
+[30] {age ^^xsd:integer}
 ```
 
 **Object Properties** — Links create relationships:
 
 ```markdown
-[Tech Corp](#company){rel="worksFor"}
+[Tech Corp](=ex:company) {worksFor}
 ```
 
 **Lists** — Repeated properties:
 
-```markdown
-- [Item 1]{property="tag"}
-- [Item 2]{property="tag"}
+```markdown {tag}
+- Item 1
+- Item 2
 ```
-
-**Code Blocks** — Automatic `SoftwareSourceCode` mapping:
-
-````markdown
-```sparql
-SELECT * WHERE { ?s ?p ?o }
-```
-````
-
-````
-
-**Tasks** — Markdown checklists become `schema:Action`:
-```markdown
-- [x] Completed task
-- [ ] Pending task
-````
-
-### Optimization Tips
-
-1. **Reuse DataFactory** — Pass custom factory instance to avoid allocations
-2. **Minimize frontmatter** — Keep `@context` simple for faster parsing
-3. **Batch processing** — Process multiple documents sequentially
-4. **Fragment IDs** — Use `#id` on headings for efficient cross-references
-
-## Future Work
-
-- [ ] Streaming API for large documents
-- [ ] Tables → CSVW integration
-- [ ] Math blocks → MathML + RDF
-- [ ] Image syntax → `schema:ImageObject`
-- [ ] Bare URL links → `dct:references`
-- [ ] Language tags (`lang` attribute)
-- [ ] Source maps for debugging
-
-## Standards Compliance
-
-This parser implements:
-
-- [MD-LD v0.1 Specification](./mdld_spec_dogfood.md)
-- [RDF/JS Data Model](https://rdf.js.org/data-model-spec/)
-- [RDFa Core 1.1](https://www.w3.org/TR/rdfa-core/) (subset)
-- [JSON-LD 1.1](https://www.w3.org/TR/json-ld11/) (frontmatter)