feat: multiline turtle literals

.changeset/fast-mangos-love.md

@@ -0,0 +1,5 @@
+---
+'@rdfjs-elements/formats-pretty': patch
+---
+
+Preserve line breaks when serialising Turtle and Trig (fixes #139)

packages/formats/package.json

@@ -36,6 +36,7 @@
     "@rdfjs/term-map": "^2.0.0",
     "@tpluscode/rdf-ns-builders": ">=3.0.2",
     "@zazuko/formats-lazy": "^1.0.1",
+    "@zazuko/prefixes": "^2.0.0",
     "readable-stream": ">=3.6.0"
   },
   "devDependencies": {
@@ -45,6 +46,7 @@
     "clownface": "^1.2.0",
     "get-stream": "^6.0.1",
     "mocha": "^10.0.0",
+    "mocha-chai-jest-snapshot": "^1.1.4",
     "rdf-ext": "^2.2.0",
     "rdf-utils-fs": "^2.2.0",
     "string-to-stream": "^3.0.1"

packages/formats/serializers/TransformToConciseHash.js

@@ -2,6 +2,7 @@ import stream from 'readable-stream'
 import { rdf, xsd } from '@tpluscode/rdf-ns-builders'
 import TermMap from '@rdfjs/term-map'
 import graphy from '@graphy/core.data.factory'
+import { LongLiteral } from './graphy/coercions.js'
 
 function isListNode(predicates) {
   return predicates.has(rdf.first) && predicates.has(rdf.rest)
@@ -169,6 +170,10 @@ export class TransformToConciseHash extends stream.Transform {
       }
     }
 
+    if (term.value.includes('\n')) {
+      return new LongLiteral(term)
+    }
+
     return graphy.fromTerm(term).concise(this.prefixes)
   }
 

packages/formats/serializers/graphy.js

@@ -1,7 +1,8 @@
 import { lazySink } from '@zazuko/formats-lazy/LazySink.js'
 import { TransformToConciseHash } from './TransformToConciseHash.js'
+import { coercions } from './graphy/coercions.js'
 
-async function serializer(importScribe, { strict } = {}) {
+async function serializer(importScribe, writerOptions = {}) {
   const create = (await importScribe).default
 
   return class {
@@ -18,13 +19,14 @@ async function serializer(importScribe, { strict } = {}) {
 
       const writer = create({
         prefixes,
+        coercions: writerOptions.coercions,
       })
 
       quadStream
         .pipe(
           new TransformToConciseHash({
             prefixes,
-            strict,
+            strict: writerOptions.strict,
             preserveListNodeProperties,
           })
         )
@@ -36,11 +38,11 @@ async function serializer(importScribe, { strict } = {}) {
 }
 
 export const TurtleSerializer = lazySink(() =>
-  serializer(import('@graphy/content.ttl.write'))
+  serializer(import('@graphy/content.ttl.write'), { coercions })
 )
 export const RdfXmlSerializer = lazySink(() =>
   serializer(import('@graphy/content.xml.scribe'), { strict: true })
 )
 export const TrigSerializer = lazySink(() =>
-  serializer(import('@graphy/content.trig.write'), { strict: true })
+  serializer(import('@graphy/content.trig.write'), { strict: true, coercions })
 )

packages/formats/serializers/graphy/coercions.js

@@ -0,0 +1,36 @@
+import { xsd } from '@tpluscode/rdf-ns-builders'
+import { shrink } from '@zazuko/prefixes'
+
+export class LongLiteral {
+  constructor(term) {
+    this.term = term
+  }
+
+  toTerm() {
+    const raw = `"""${this.term.value.replace(/"$/, '\\"')}"""`
+
+    return {
+      terse: prefixes => raw + this.langOrDatatype(prefixes),
+      verbose: prefixes => raw + this.langOrDatatype(prefixes),
+    }
+  }
+
+  langOrDatatype(prefixes) {
+    if (this.term.language) {
+      return `@${this.term.language}`
+    }
+
+    if (this.term.datatype.equals(xsd.string)) {
+      return ''
+    }
+
+    const shrunk = shrink(this.term.datatype.value, prefixes)
+    if (!shrunk) {
+      return `^^<${this.term.datatype.value}>`
+    }
+
+    return `^^${shrunk}`
+  }
+}
+
+export const coercions = new Map([[LongLiteral, literal => literal.toTerm()]])

packages/formats/test/graphs/multiline-literals.ttl

@@ -0,0 +1,11 @@
+PREFIX ex: <http://example.org/>
+
+ex:a ex:b "The first line\nThe second line\n  more" .
+
+ex:x ex:y "Long\ndescription\nwith datatype"^^ex:multiline .
+
+ex:m ex:n "multiline\nwith language"@en-US .
+
+ex:f ex:g "The first line\n\"The quoted line\"\n  more" .
+
+ex:r ex:q "Quote\nat end\"" .

packages/formats/test/serializers/__snapshots__/index.test.js.snap

@@ -0,0 +1,48 @@
+// Jest Snapshot v1, https://goo.gl/fbAQLP
+
+exports[`@rdfjs-elements/formats-pretty/serializers turtle preserves line breaks in literals 1`] = `
+"
+<http://example.org/a> <http://example.org/b> \\"\\"\\"The first line
+The second line
+  more\\"\\"\\" .
+
+<http://example.org/x> <http://example.org/y> \\"\\"\\"Long
+description
+with datatype\\"\\"\\"^^<http://example.org/multiline> .
+
+<http://example.org/m> <http://example.org/n> \\"\\"\\"multiline
+with language\\"\\"\\"@en-US .
+
+<http://example.org/f> <http://example.org/g> \\"\\"\\"The first line
+\\"The quoted line\\"
+  more\\"\\"\\" .
+
+<http://example.org/r> <http://example.org/q> \\"\\"\\"Quote
+at end\\\\\\"\\"\\"\\" .
+
+"
+`;
+
+exports[`@rdfjs-elements/formats-pretty/serializers turtle preserves line breaks in literals when using prefixes 1`] = `
+"@prefix ex: <http://example.org/> .
+
+ex:a ex:b \\"\\"\\"The first line
+The second line
+  more\\"\\"\\" .
+
+ex:x ex:y \\"\\"\\"Long
+description
+with datatype\\"\\"\\"^^ex:multiline .
+
+ex:m ex:n \\"\\"\\"multiline
+with language\\"\\"\\"@en-US .
+
+ex:f ex:g \\"\\"\\"The first line
+\\"The quoted line\\"
+  more\\"\\"\\" .
+
+ex:r ex:q \\"\\"\\"Quote
+at end\\\\\\"\\"\\"\\" .
+
+"
+`;

packages/formats/test/serializers/index.test.js

@@ -1,4 +1,5 @@
-import { expect } from 'chai'
+import chai, { expect } from 'chai'
+import { jestSnapshotPlugin } from 'mocha-chai-jest-snapshot'
 import prefixes from '@zazuko/prefixes'
 import * as ns from '@tpluscode/rdf-ns-builders'
 import $rdf from 'rdf-ext'
@@ -13,6 +14,8 @@ import { TurtleSerializer } from '../../serializers/graphy.js'
 const __dirname = dirname(fileURLToPath(import.meta.url))
 
 describe('@rdfjs-elements/formats-pretty/serializers', () => {
+  chai.use(jestSnapshotPlugin())
+
   describe('turtle', () => {
     it('combines default and import prefixes', async () => {
       // given
@@ -93,6 +96,38 @@ describe('@rdfjs-elements/formats-pretty/serializers', () => {
       // then
       expect(serialized).to.be.ok
     })
+
+    it('preserves line breaks in literals', async () => {
+      // given
+      const graph = rdfUtil.fromFile(
+        join(__dirname, `../graphs/multiline-literals.ttl`)
+      )
+      const sink = new TurtleSerializer()
+
+      // when
+      const serialized = await getStream(sink.import(graph))
+
+      // then
+      expect(serialized).toMatchSnapshot()
+    })
+
+    it('preserves line breaks in literals when using prefixes', async () => {
+      // given
+      const graph = rdfUtil.fromFile(
+        join(__dirname, `../graphs/multiline-literals.ttl`)
+      )
+      const sink = new TurtleSerializer({
+        prefixes: {
+          ex: 'http://example.org/',
+        },
+      })
+
+      // when
+      const serialized = await getStream(sink.import(graph))
+
+      // then
+      expect(serialized).toMatchSnapshot()
+    })
   })
 
   describe('jsonld', () => {