Move frontmatter extraction from renderer to redux (#1413)

src/components/common/note-frontmatter/extract-frontmatter.test.ts

@@ -0,0 +1,94 @@
+/*
+ * SPDX-FileCopyrightText: 2021 The HedgeDoc developers (see AUTHORS file)
+ *
+ * SPDX-License-Identifier: AGPL-3.0-only
+ */
+
+import { extractFrontmatter } from './extract-frontmatter'
+import { PresentFrontmatterExtractionResult } from './types'
+
+describe('frontmatter extraction', () => {
+  describe('frontmatterPresent property', () => {
+    it('is false when note does not contain three dashes at all', () => {
+      const testNote = 'abcdef\nmore text'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(false)
+    })
+    it('is false when note does not start with three dashes', () => {
+      const testNote = '\n---\nthis is not frontmatter'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(false)
+    })
+    it('is false when note start with less than three dashes', () => {
+      const testNote = '--\nthis is not frontmatter'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(false)
+    })
+    it('is false when note starts with three dashes but contains other characters in the same line', () => {
+      const testNote = '--- a\nthis is not frontmatter'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(false)
+    })
+    it('is false when note has no ending marker for frontmatter', () => {
+      const testNote = '---\nthis is not frontmatter\nbecause\nthere is no\nend marker'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(false)
+    })
+    it('is false when note end marker is present but with not the same amount of dashes as start marker', () => {
+      const testNote = '---\nthis is not frontmatter\n----\ncontent'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(false)
+    })
+    it('is true when note end marker is present with the same amount of dashes as start marker', () => {
+      const testNote = '---\nthis is frontmatter\n---\ncontent'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(true)
+    })
+    it('is true when note end marker is present with the same amount of dashes as start marker but without content', () => {
+      const testNote = '---\nthis is frontmatter\n---'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(true)
+    })
+    it('is true when note end marker is present with the same amount of dots as start marker', () => {
+      const testNote = '---\nthis is frontmatter\n...\ncontent'
+      const extraction = extractFrontmatter(testNote)
+      expect(extraction.frontmatterPresent).toBe(true)
+    })
+  })
+
+  describe('frontmatterLines property', () => {
+    it('is correct for single line frontmatter without content', () => {
+      const testNote = '---\nsingle line frontmatter\n...'
+      const extraction = extractFrontmatter(testNote) as PresentFrontmatterExtractionResult
+      expect(extraction.frontmatterLines).toEqual(3)
+    })
+    it('is correct for single line frontmatter with content', () => {
+      const testNote = '---\nsingle line frontmatter\n...\ncontent'
+      const extraction = extractFrontmatter(testNote) as PresentFrontmatterExtractionResult
+      expect(extraction.frontmatterLines).toEqual(3)
+    })
+    it('is correct for multi-line frontmatter without content', () => {
+      const testNote = '---\nabc\n123\ndef\n...'
+      const extraction = extractFrontmatter(testNote) as PresentFrontmatterExtractionResult
+      expect(extraction.frontmatterLines).toEqual(5)
+    })
+    it('is correct for multi-line frontmatter with content', () => {
+      const testNote = '---\nabc\n123\ndef\n...\ncontent'
+      const extraction = extractFrontmatter(testNote) as PresentFrontmatterExtractionResult
+      expect(extraction.frontmatterLines).toEqual(5)
+    })
+  })
+
+  describe('rawFrontmatterText property', () => {
+    it('contains single-line frontmatter text', () => {
+      const testNote = '---\nsingle-line\n...\ncontent'
+      const extraction = extractFrontmatter(testNote) as PresentFrontmatterExtractionResult
+      expect(extraction.rawFrontmatterText).toEqual('single-line')
+    })
+    it('contains multi-line frontmatter text', () => {
+      const testNote = '---\nmulti\nline\n...\ncontent'
+      const extraction = extractFrontmatter(testNote) as PresentFrontmatterExtractionResult
+      expect(extraction.rawFrontmatterText).toEqual('multi\nline')
+    })
+  })
+})

src/components/common/note-frontmatter/extract-frontmatter.ts

@@ -0,0 +1,40 @@
+/*
+ * SPDX-FileCopyrightText: 2021 The HedgeDoc developers (see AUTHORS file)
+ *
+ * SPDX-License-Identifier: AGPL-3.0-only
+ */
+import { FrontmatterExtractionResult } from './types'
+
+const FRONTMATTER_BEGIN_REGEX = /^-{3,}$/
+const FRONTMATTER_END_REGEX = /^(?:-{3,}|\.{3,})$/
+
+/**
+ * Extracts a frontmatter block from a given multiline string.
+ * A valid frontmatter block requires the content to start with a line containing at least three dashes.
+ * The block is terminated by a line containing the same amount of dashes or dots as the first line.
+ * @param content The multiline string from which the frontmatter should be extracted.
+ * @return { frontmatterPresent } false if no frontmatter block could be found, true if a block was found.
+ *         { rawFrontmatterText } if a block was found, this property contains the extracted text without the fencing.
+ *         { frontmatterLines }   if a block was found, this property contains the number of lines to skip from the
+ *                                given multiline string for retrieving the non-frontmatter content.
+ */
+export const extractFrontmatter = (content: string): FrontmatterExtractionResult => {
+  const lines = content.split('\n')
+  if (lines.length < 2 || !FRONTMATTER_BEGIN_REGEX.test(lines[0])) {
+    return {
+      frontmatterPresent: false
+    }
+  }
+  for (let i = 1; i < lines.length; i++) {
+    if (lines[i].length === lines[0].length && FRONTMATTER_END_REGEX.test(lines[i])) {
+      return {
+        frontmatterPresent: true,
+        rawFrontmatterText: lines.slice(1, i).join('\n'),
+        frontmatterLines: i + 1
+      }
+    }
+  }
+  return {
+    frontmatterPresent: false
+  }
+}

src/components/common/note-frontmatter/note-frontmatter.test.ts

@@ -0,0 +1,68 @@
+/*
+ * SPDX-FileCopyrightText: 2021 The HedgeDoc developers (see AUTHORS file)
+ *
+ * SPDX-License-Identifier: AGPL-3.0-only
+ */
+
+import { NoteFrontmatter } from './note-frontmatter'
+
+describe('yaml frontmatter', () => {
+  it('should parse "title"', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml('title: test')
+    expect(noteFrontmatter.title).toEqual('test')
+  })
+
+  it('should parse "robots"', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml('robots: index, follow')
+    expect(noteFrontmatter.robots).toEqual('index, follow')
+  })
+
+  it('should parse the deprecated tags syntax', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml('tags: test123, abc')
+    expect(noteFrontmatter.tags).toEqual(['test123', 'abc'])
+    expect(noteFrontmatter.deprecatedTagsSyntax).toEqual(true)
+  })
+
+  it('should parse the tags list syntax', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml(`tags:
+      - test123
+      - abc
+    `)
+    expect(noteFrontmatter.tags).toEqual(['test123', 'abc'])
+    expect(noteFrontmatter.deprecatedTagsSyntax).toEqual(false)
+  })
+
+  it('should parse the tag inline-list syntax', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml("tags: ['test123', 'abc']")
+    expect(noteFrontmatter.tags).toEqual(['test123', 'abc'])
+    expect(noteFrontmatter.deprecatedTagsSyntax).toEqual(false)
+  })
+
+  it('should parse "breaks"', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml('breaks: false')
+    expect(noteFrontmatter.breaks).toEqual(false)
+  })
+
+  it('should parse an empty opengraph object', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml('opengraph:')
+    expect(noteFrontmatter.opengraph).toEqual(new Map<string, string>())
+  })
+
+  it('should parse an opengraph title', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml(`opengraph:
+      title: Testtitle
+    `)
+    expect(noteFrontmatter.opengraph.get('title')).toEqual('Testtitle')
+  })
+
+  it('should parse multiple opengraph values', () => {
+    const noteFrontmatter = NoteFrontmatter.createFromYaml(`opengraph:
+      title: Testtitle
+      image: https://dummyimage.com/48.png
+      image:type: image/png
+    `)
+    expect(noteFrontmatter.opengraph.get('title')).toEqual('Testtitle')
+    expect(noteFrontmatter.opengraph.get('image')).toEqual('https://dummyimage.com/48.png')
+    expect(noteFrontmatter.opengraph.get('image:type')).toEqual('image/png')
+  })
+})

src/components/common/note-frontmatter/note-frontmatter.ts

@@ -0,0 +1,70 @@
+/*
+ * SPDX-FileCopyrightText: 2021 The HedgeDoc developers (see AUTHORS file)
+ *
+ * SPDX-License-Identifier: AGPL-3.0-only
+ */
+
+// import { RevealOptions } from 'reveal.js'
+import { load } from 'js-yaml'
+import { ISO6391, NoteTextDirection, NoteType, RawNoteFrontmatter } from './types'
+
+/**
+ * Class that represents the parsed frontmatter metadata of a note.
+ */
+export class NoteFrontmatter {
+  title: string
+  description: string
+  tags: string[]
+  deprecatedTagsSyntax: boolean
+  robots: string
+  lang: typeof ISO6391[number]
+  dir: NoteTextDirection
+  breaks: boolean
+  GA: string
+  disqus: string
+  type: NoteType
+  opengraph: Map<string, string>
+
+  /**
+   * Creates a new frontmatter metadata instance based on the given raw metadata properties.
+   * @param rawData A {@link RawNoteFrontmatter} object containing the properties of the parsed yaml frontmatter.
+   */
+  constructor(rawData: RawNoteFrontmatter) {
+    this.title = rawData.title ?? ''
+    this.description = rawData.description ?? ''
+    this.robots = rawData.robots ?? ''
+    this.breaks = rawData.breaks ?? true
+    this.GA = rawData.GA ?? ''
+    this.disqus = rawData.disqus ?? ''
+    this.lang = (rawData.lang ? ISO6391.find((lang) => lang === rawData.lang) : undefined) ?? 'en'
+    this.type =
+      (rawData.type ? Object.values(NoteType).find((type) => type === rawData.type) : undefined) ?? NoteType.DOCUMENT
+    this.dir =
+      (rawData.dir ? Object.values(NoteTextDirection).find((dir) => dir === rawData.dir) : undefined) ??
+      NoteTextDirection.LTR
+    if (typeof rawData?.tags === 'string') {
+      this.tags = rawData?.tags?.split(',').map((entry) => entry.trim()) ?? []
+      this.deprecatedTagsSyntax = true
+    } else if (typeof rawData?.tags === 'object') {
+      this.tags = rawData?.tags?.filter((tag) => tag !== null) ?? []
+      this.deprecatedTagsSyntax = false
+    } else {
+      this.tags = []
+      this.deprecatedTagsSyntax = false
+    }
+    this.opengraph = rawData?.opengraph
+      ? new Map<string, string>(Object.entries(rawData.opengraph))
+      : new Map<string, string>()
+  }
+
+  /**
+   * Creates a new frontmatter metadata instance based on a raw yaml string.
+   * @param rawYaml The frontmatter content in yaml format.
+   * @throws Error when the content string is invalid yaml.
+   * @return Frontmatter metadata instance containing the parsed properties from the yaml content.
+   */
+  static createFromYaml(rawYaml: string): NoteFrontmatter {
+    const rawNoteFrontmatter = load(rawYaml) as RawNoteFrontmatter
+    return new NoteFrontmatter(rawNoteFrontmatter)
+  }
+}

src/components/common/note-frontmatter/types.ts

@@ -0,0 +1,253 @@
+/*
+ * SPDX-FileCopyrightText: 2021 The HedgeDoc developers (see AUTHORS file)
+ *
+ * SPDX-License-Identifier: AGPL-3.0-only
+ */
+
+export type FrontmatterExtractionResult = PresentFrontmatterExtractionResult | NonPresentFrontmatterExtractionResult
+
+export interface RendererFrontmatterInfo {
+  offsetLines: number
+  frontmatterInvalid: boolean
+  deprecatedSyntax: boolean
+}
+
+export interface PresentFrontmatterExtractionResult {
+  frontmatterPresent: true
+  rawFrontmatterText: string
+  frontmatterLines: number
+}
+
+interface NonPresentFrontmatterExtractionResult {
+  frontmatterPresent: false
+}
+
+export interface RawNoteFrontmatter {
+  title: string | undefined
+  description: string | undefined
+  tags: string | string[] | undefined
+  robots: string | undefined
+  lang: string | undefined
+  dir: string | undefined
+  breaks: boolean | undefined
+  GA: string | undefined
+  disqus: string | undefined
+  type: string | undefined
+  slideOptions: unknown
+  opengraph: { [key: string]: string } | null
+}
+
+export const ISO6391 = [
+  'aa',
+  'ab',
+  'af',
+  'am',
+  'ar',
+  'ar-ae',
+  'ar-bh',
+  'ar-dz',
+  'ar-eg',
+  'ar-iq',
+  'ar-jo',
+  'ar-kw',
+  'ar-lb',
+  'ar-ly',
+  'ar-ma',
+  'ar-om',
+  'ar-qa',
+  'ar-sa',
+  'ar-sy',
+  'ar-tn',
+  'ar-ye',
+  'as',
+  'ay',
+  'de-at',
+  'de-ch',
+  'de-li',
+  'de-lu',
+  'div',
+  'dz',
+  'el',
+  'en',
+  'en-au',
+  'en-bz',
+  'en-ca',
+  'en-gb',
+  'en-ie',
+  'en-jm',
+  'en-nz',
+  'en-ph',
+  'en-tt',
+  'en-us',
+  'en-za',
+  'en-zw',
+  'eo',
+  'es',
+  'es-ar',
+  'es-bo',
+  'es-cl',
+  'es-co',
+  'es-cr',
+  'es-do',
+  'es-ec',
+  'es-es',
+  'es-gt',
+  'es-hn',
+  'es-mx',
+  'es-ni',
+  'es-pa',
+  'es-pe',
+  'es-pr',
+  'es-py',
+  'es-sv',
+  'es-us',
+  'es-uy',
+  'es-ve',
+  'et',
+  'eu',
+  'fa',
+  'fi',
+  'fj',
+  'fo',
+  'fr',
+  'fr-be',
+  'fr-ca',
+  'fr-ch',
+  'fr-lu',
+  'fr-mc',
+  'fy',
+  'ga',
+  'gd',
+  'gl',
+  'gn',
+  'gu',
+  'ha',
+  'he',
+  'hi',
+  'hr',
+  'hu',
+  'hy',
+  'ia',
+  'id',
+  'ie',
+  'ik',
+  'in',
+  'is',
+  'it',
+  'it-ch',
+  'iw',
+  'ja',
+  'ji',
+  'jw',
+  'ka',
+  'kk',
+  'kl',
+  'km',
+  'kn',
+  'ko',
+  'kok',
+  'ks',
+  'ku',
+  'ky',
+  'kz',
+  'la',
+  'ln',
+  'lo',
+  'ls',
+  'lt',
+  'lv',
+  'mg',
+  'mi',
+  'mk',
+  'ml',
+  'mn',
+  'mo',
+  'mr',
+  'ms',
+  'mt',
+  'my',
+  'na',
+  'nb-no',
+  'ne',
+  'nl',
+  'nl-be',
+  'nn-no',
+  'no',
+  'oc',
+  'om',
+  'or',
+  'pa',
+  'pl',
+  'ps',
+  'pt',
+  'pt-br',
+  'qu',
+  'rm',
+  'rn',
+  'ro',
+  'ro-md',
+  'ru',
+  'ru-md',
+  'rw',
+  'sa',
+  'sb',
+  'sd',
+  'sg',
+  'sh',
+  'si',
+  'sk',
+  'sl',
+  'sm',
+  'sn',
+  'so',
+  'sq',
+  'sr',
+  'ss',
+  'st',
+  'su',
+  'sv',
+  'sv-fi',
+  'sw',
+  'sx',
+  'syr',
+  'ta',
+  'te',
+  'tg',
+  'th',
+  'ti',
+  'tk',
+  'tl',
+  'tn',
+  'to',
+  'tr',
+  'ts',
+  'tt',
+  'tw',
+  'uk',
+  'ur',
+  'us',
+  'uz',
+  'vi',
+  'vo',
+  'wo',
+  'xh',
+  'yi',
+  'yo',
+  'zh',
+  'zh-cn',
+  'zh-hk',
+  'zh-mo',
+  'zh-sg',
+  'zh-tw',
+  'zu'
+] as const
+
+export enum NoteType {
+  DOCUMENT = '',
+  SLIDE = 'slide'
+}
+
+export enum NoteTextDirection {
+  LTR = 'ltr',
+  RTL = 'rtl'
+}