Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improvements to find dates #405

Merged
merged 1 commit into from
Oct 18, 2024
Merged

Improvements to find dates #405

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions src/utils/extractMetaData.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,7 @@

import { DOMParser } from 'linkedom'
import extractLdSchema from './extractLdSchema.js'
import findDate from './findDate.js'

/**
* @param {Element} node
Expand Down Expand Up @@ -143,5 +144,10 @@ export default (html) => {
})

const entries = extractLdSchema(doc, entry)

if (!entries.published) {
entries.published = findDate(doc)
}

return entries
}
26 changes: 26 additions & 0 deletions src/utils/extractMetaData.test.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,12 @@ import extractMetaData from './extractMetaData.js'

const keys = 'url shortlink amphtml canonical title description image author source published favicon type'.split(' ')

function isDateString (date) {
if (typeof date !== 'string') return false
const d = new Date(date)
return !isNaN(d.getTime())
}

describe('test extractMetaData', () => {
it('test extractMetaData(good content)', async () => {
const html = readFileSync('./test-data/regular-article.html', 'utf8')
Expand All @@ -28,4 +34,24 @@ describe('test extractMetaData', () => {
assert.ok(hasProperty(result, k))
})
})

it('test extractMetaData(find date)', async () => {
const html1 = readFileSync('./test-data/regular-article-date-time.html', 'utf8')
const html2 = readFileSync('./test-data/regular-article-date-itemprop.html', 'utf8')
const html3 = readFileSync('./test-data/regular-article-date-span.html', 'utf8')
const result1 = extractMetaData(html1)
const result2 = extractMetaData(html2)
const result3 = extractMetaData(html3)
assert.ok(isObject(result1))
assert.ok(isObject(result2))
assert.ok(isObject(result3))
keys.forEach((k) => {
assert.ok(hasProperty(result1, k))
assert.ok(hasProperty(result3, k))
assert.ok(hasProperty(result3, k))
})
assert.ok(isDateString(result1.published))
assert.ok(isDateString(result2.published))
assert.ok(isDateString(result3.published))
})
})
57 changes: 57 additions & 0 deletions src/utils/findDate.js
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@

/**
* Convert date format to YYYY-MM-DD
*
* @param {string} dateString
* @returns {string} YYYY-MM-DD
*/
function convertDateFormat (dateString) {
const parts = dateString.split('/')
if (parts.length !== 3) return dateString

let year, month, day

if (parseInt(parts[0]) > 12) {
[day, month, year] = parts
} else {
[month, day, year] = parts
}

year = year.length === 2 ? '20' + year : year
return `${year}-${month.padStart(2, '0')}-${day.padStart(2, '0')}T00:00:00`
}

/**
* Look for the publication date in the body of the content.
*
* @param {Document} document - The HTML Document
* @returns {string} The date string
*/
export default function (doc) {
const datePatterns = [
/\d{4}-\d{2}-\d{2}/,
/\d{1,2}\/\d{1,2}\/\d{2,4}/,
]

const findDate = (element) => {
for (const pattern of datePatterns) {
const match = element.textContent.match(pattern)
if (match) return convertDateFormat(match[0])
}
return null
}

const priorityElements = doc.querySelectorAll('time, [datetime], [itemprop~=datePublished], [itemprop~=dateCreated]')
for (const el of priorityElements) {
const date = el.getAttribute('datetime') || el.getAttribute('content') || findDate(el)
if (date) return date
}

const secondaryElements = doc.querySelectorAll('p, span, div')
for (const el of secondaryElements) {
const date = findDate(el)
if (date) return date
}

return null
}
57 changes: 57 additions & 0 deletions test-data/regular-article-date-itemprop.html
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@
<!doctype html>
<html>
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Article title here - ArticleParser</title>
<meta name="author" content="Alice">
<meta name="description" content="Few words about this article">
<meta name="keywords" content="alpha, beta, gamma">
<meta name="twitter:site" content="@ArticleParser">
<meta name="twitter:url" content="https://somewhere.com/path/to/article-title-here">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:image" content="https://somewhere.com/path/to/image.jpg">
<meta name="twitter:creator" content="@alice">
<meta property="og:title" content="Article title here">
<meta property="og:type" content="article">
<meta property="og:url" content="https://somewhere.com/path/to/article-title-here">
<meta property="og:description" content="Navigation here Few can name a rational peach that isn't a conscientious goldfish! One cannot separate snakes from plucky pomegranates? Draped neatly on a hanger, the melons could be said to resemble knowledgeable pigs.">
<meta property="og:image" content="https://somewhere.com/path/to/image.jpg">

<link rel="stylesheet" href="/path/to/cssfile.css">
<link rel="canonical" href="https://somewhere.com/another/path/to/article-title-here">
<link rel="amphtml" href="https://m.somewhere.com/another/path/to/article-title-here.amp">
<link rel="shortlink" href="https://sw.re/419283">
<link rel="icon" href="https://somewhere.com/favicon.ico">

<link rel="alternate" title="ArticleParser" type="application/atom+xml" href="https://somewhere.com/atom.xml">

<link rel="manifest" href="/manifest.json">
</head>
<body>
<header>Page header here</header>
<main>
<section>
<nav>Navigation here</nav>
</section>
<section>
<h1>Article title here</h1>
<article>
<div itemprop="datePublished" datetime='2024-10-16T07:33+03:00' class='published'> Published 11/09/2024 07h33min</div>

<div class="contentdetail">Few can name a <a href="https://otherwhere.com/descriptions/rational-peach">rational peach</a> that isn't a conscientious goldfish! One cannot separate snakes from plucky pomegranates? Draped neatly on a hanger, the melons could be said to resemble knowledgeable pigs. Some posit the enchanting tiger to be less than confident. The literature would have us believe that an impartial turtle is not but a hippopotamus. Unfortunately, that is wrong; on the contrary, those cows are nothing more than pandas! The chicken is a shark; A turtle can hardly be considered a kind horse without also being a pomegranate. Zebras are witty persimmons.</div>
<p class="contentdetail">
Those cheetahs are nothing more than dogs. A <a href="/dict/watermelon">watermelon</a> is an exuberant kangaroo. An octopus is the tangerine of a grapes? The cherry is a shark. Recent controversy aside, they were lost without the cheerful plum that composed their fox. As far as we can estimate, one cannot separate camels from dynamic hamsters. Those tigers are nothing more than cows! A cow is a squirrel from the right perspective. Their banana was, in this moment, a helpful bear.</p>
<p>The first fair dog is, in its own way, a lemon.</p>
<address>4746 Kelly Drive, West Virginia</address>
<img src="./orange.png" style="border: solid 1px #000">
</article>
</section>
<section class="sidebar-widget">
<widget>Some widget here</widget>
<widget>Some widget here</widget>
</section>
</main>
<footer>Page footer here</footer>
</body>
</html>
57 changes: 57 additions & 0 deletions test-data/regular-article-date-span.html
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@
<!doctype html>
<html>
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Article title here - ArticleParser</title>
<meta name="author" content="Alice">
<meta name="description" content="Few words about this article">
<meta name="keywords" content="alpha, beta, gamma">
<meta name="twitter:site" content="@ArticleParser">
<meta name="twitter:url" content="https://somewhere.com/path/to/article-title-here">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:image" content="https://somewhere.com/path/to/image.jpg">
<meta name="twitter:creator" content="@alice">
<meta property="og:title" content="Article title here">
<meta property="og:type" content="article">
<meta property="og:url" content="https://somewhere.com/path/to/article-title-here">
<meta property="og:description" content="Navigation here Few can name a rational peach that isn't a conscientious goldfish! One cannot separate snakes from plucky pomegranates? Draped neatly on a hanger, the melons could be said to resemble knowledgeable pigs.">
<meta property="og:image" content="https://somewhere.com/path/to/image.jpg">

<link rel="stylesheet" href="/path/to/cssfile.css">
<link rel="canonical" href="https://somewhere.com/another/path/to/article-title-here">
<link rel="amphtml" href="https://m.somewhere.com/another/path/to/article-title-here.amp">
<link rel="shortlink" href="https://sw.re/419283">
<link rel="icon" href="https://somewhere.com/favicon.ico">

<link rel="alternate" title="ArticleParser" type="application/atom+xml" href="https://somewhere.com/atom.xml">

<link rel="manifest" href="/manifest.json">
</head>
<body>
<header>Page header here</header>
<main>
<section>
<nav>Navigation here</nav>
</section>
<section>
<h1>Article title here</h1>
<article>
<span class='published'> Published at 11/09/2024 07h33 am</span>

<div class="contentdetail">Few can name a <a href="https://otherwhere.com/descriptions/rational-peach">rational peach</a> that isn't a conscientious goldfish! One cannot separate snakes from plucky pomegranates? Draped neatly on a hanger, the melons could be said to resemble knowledgeable pigs. Some posit the enchanting tiger to be less than confident. The literature would have us believe that an impartial turtle is not but a hippopotamus. Unfortunately, that is wrong; on the contrary, those cows are nothing more than pandas! The chicken is a shark; A turtle can hardly be considered a kind horse without also being a pomegranate. Zebras are witty persimmons.</div>
<p class="contentdetail">
Those cheetahs are nothing more than dogs. A <a href="/dict/watermelon">watermelon</a> is an exuberant kangaroo. An octopus is the tangerine of a grapes? The cherry is a shark. Recent controversy aside, they were lost without the cheerful plum that composed their fox. As far as we can estimate, one cannot separate camels from dynamic hamsters. Those tigers are nothing more than cows! A cow is a squirrel from the right perspective. Their banana was, in this moment, a helpful bear.</p>
<p>The first fair dog is, in its own way, a lemon.</p>
<address>4746 Kelly Drive, West Virginia</address>
<img src="./orange.png" style="border: solid 1px #000">
</article>
</section>
<section class="sidebar-widget">
<widget>Some widget here</widget>
<widget>Some widget here</widget>
</section>
</main>
<footer>Page footer here</footer>
</body>
</html>
57 changes: 57 additions & 0 deletions test-data/regular-article-date-time.html
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@
<!doctype html>
<html>
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Article title here - ArticleParser</title>
<meta name="author" content="Alice">
<meta name="description" content="Few words about this article">
<meta name="keywords" content="alpha, beta, gamma">
<meta name="twitter:site" content="@ArticleParser">
<meta name="twitter:url" content="https://somewhere.com/path/to/article-title-here">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:image" content="https://somewhere.com/path/to/image.jpg">
<meta name="twitter:creator" content="@alice">
<meta property="og:title" content="Article title here">
<meta property="og:type" content="article">
<meta property="og:url" content="https://somewhere.com/path/to/article-title-here">
<meta property="og:description" content="Navigation here Few can name a rational peach that isn't a conscientious goldfish! One cannot separate snakes from plucky pomegranates? Draped neatly on a hanger, the melons could be said to resemble knowledgeable pigs.">
<meta property="og:image" content="https://somewhere.com/path/to/image.jpg">

<link rel="stylesheet" href="/path/to/cssfile.css">
<link rel="canonical" href="https://somewhere.com/another/path/to/article-title-here">
<link rel="amphtml" href="https://m.somewhere.com/another/path/to/article-title-here.amp">
<link rel="shortlink" href="https://sw.re/419283">
<link rel="icon" href="https://somewhere.com/favicon.ico">

<link rel="alternate" title="ArticleParser" type="application/atom+xml" href="https://somewhere.com/atom.xml">

<link rel="manifest" href="/manifest.json">
</head>
<body>
<header>Page header here</header>
<main>
<section>
<nav>Navigation here</nav>
</section>
<section>
<h1>Article title here</h1>
<article>
<time datetime='2024-10-16T07:33+03:00' class='published'> Published 11/09/2024 07h33min</time>

<div class="contentdetail">Few can name a <a href="https://otherwhere.com/descriptions/rational-peach">rational peach</a> that isn't a conscientious goldfish! One cannot separate snakes from plucky pomegranates? Draped neatly on a hanger, the melons could be said to resemble knowledgeable pigs. Some posit the enchanting tiger to be less than confident. The literature would have us believe that an impartial turtle is not but a hippopotamus. Unfortunately, that is wrong; on the contrary, those cows are nothing more than pandas! The chicken is a shark; A turtle can hardly be considered a kind horse without also being a pomegranate. Zebras are witty persimmons.</div>
<p class="contentdetail">
Those cheetahs are nothing more than dogs. A <a href="/dict/watermelon">watermelon</a> is an exuberant kangaroo. An octopus is the tangerine of a grapes? The cherry is a shark. Recent controversy aside, they were lost without the cheerful plum that composed their fox. As far as we can estimate, one cannot separate camels from dynamic hamsters. Those tigers are nothing more than cows! A cow is a squirrel from the right perspective. Their banana was, in this moment, a helpful bear.</p>
<p>The first fair dog is, in its own way, a lemon.</p>
<address>4746 Kelly Drive, West Virginia</address>
<img src="./orange.png" style="border: solid 1px #000">
</article>
</section>
<section class="sidebar-widget">
<widget>Some widget here</widget>
<widget>Some widget here</widget>
</section>
</main>
<footer>Page footer here</footer>
</body>
</html>
Loading