Medium: Update selectors and blog title extraction (#2565)

这个提交包含在:
Abe Jellinek 2021-07-26 10:27:41 -06:00 提交者 GitHub
父节点 115b9b3dbd
当前提交 611ccc3613
找不到此签名对应的密钥
GPG 密钥 ID: 4AEE18F83AFDEB23

106
Medium.js
查看文件

@ -2,20 +2,20 @@
"translatorID": "6c957d6b-a554-474f-81a9-91c178fef65d",
"label": "Medium",
"creator": "Philipp Zumstein",
"target": "^https?://medium\\.com",
"target": "^https?://([^.]+\\.)?medium\\.com",
"minVersion": "3.0",
"maxVersion": "",
"priority": 100,
"inRepository": true,
"translatorType": 4,
"browserSupport": "gcsibv",
"lastUpdated": "2016-12-27 14:13:31"
"lastUpdated": "2021-07-20 01:55:16"
}
/*
***** BEGIN LICENSE BLOCK *****
Copyright © 2016 Philipp Zumstein
Copyright © 2016-2021 Philipp Zumstein and Abe Jellinek
This file is part of Zotero.
@ -37,7 +37,7 @@
function detectWeb(doc, url) {
if (ZU.xpathText(doc, '//article[contains(@class, "postArticle")]')) {
if (doc.querySelector('article')) {
return "blogPost";
} else if (getSearchResults(doc, true)) {
return "multiple";
@ -47,10 +47,11 @@ function detectWeb(doc, url) {
function getSearchResults(doc, checkOnly) {
var items = {};
var found = false;
var rows = ZU.xpath(doc, '//a[@data-post-id and h3]|//div[contains(@class, "postArticle-content")]/a[section][1]|//a[div[contains(@class, "postArticle-content")]]');
for (var i=0; i<rows.length; i++) {
var href = rows[i].href;
var title = ZU.xpathText(rows[i], './/h2|.//h3');
var rows = ZU.xpath(doc, '//a[@data-post-id and h3]|//div[contains(@class, "postArticle-content")]/a[section][1]|//a[div[contains(@class, "postArticle-content")]]|//h1//a[@rel="noopener"]');
for (let row of rows) {
var href = row.href;
var title = ZU.xpathText(row, './/h2|.//h3');
if (!title) title = row.textContent;
if (!href || !title) continue;
if (checkOnly) return true;
found = true;
@ -66,11 +67,7 @@ function doWeb(doc, url) {
if (!items) {
return true;
}
var articles = [];
for (var i in items) {
articles.push(i);
}
ZU.processDocuments(articles, scrape);
ZU.processDocuments(Object.keys(items), scrape);
});
} else {
scrape(doc, url);
@ -81,16 +78,12 @@ function scrape(doc, url) {
var translator = Zotero.loadTranslator('web');
// Embedded Metadata
translator.setTranslator('951c027d-74ac-47d4-a107-9c3069ab7b48');
//translator.setDocument(doc);
translator.setDocument(doc);
translator.setHandler('itemDone', function (obj, item) {
var parts = item.title.split('–');
if (parts.length == 2) {
item.title = parts[0];
item.blogTitle = parts[1];
delete item.publicationTitle;
}
delete item.publicationTitle;
item.blogTitle = JSON.parse(text(doc, 'script[type="application/ld+json"]'))
.publisher.name;
item.complete();
});
@ -132,14 +125,77 @@ var testCases = [
"creatorType": "author"
}
],
"date": "2014-04-22T16:02:32.005Z",
"date": "2014-04-24T12:59:00.600Z",
"abstractNote": "Election monitoring, new power of social media and old power of structural power",
"blogTitle": "The Message",
"shortTitle": "What If the Feature Is the Bug?",
"language": "en",
"url": "https://medium.com/message/what-if-the-feature-is-the-bug-7957d8e685c4",
"attachments": [
{
"title": "Snapshot"
"title": "Snapshot",
"mimeType": "text/html"
}
],
"tags": [],
"notes": [],
"seeAlso": []
}
]
},
{
"type": "web",
"url": "https://susanorlean.medium.com/lessons-from-an-old-dog-about-creaky-bones-and-graying-hair-9539ca79e49",
"items": [
{
"itemType": "blogPost",
"title": "Lessons from an Old Dog About Creaky Bones and Graying Hair",
"creators": [
{
"firstName": "Susan",
"lastName": "Orlean",
"creatorType": "author"
}
],
"date": "2021-07-15T16:37:56.246Z",
"abstractNote": "The terrible truth about pets aging faster than you",
"blogTitle": "Medium",
"language": "en",
"url": "https://susanorlean.medium.com/lessons-from-an-old-dog-about-creaky-bones-and-graying-hair-9539ca79e49",
"attachments": [
{
"title": "Snapshot",
"mimeType": "text/html"
}
],
"tags": [],
"notes": [],
"seeAlso": []
}
]
},
{
"type": "web",
"url": "https://medium.com/illumination-curated/4-solid-facts-to-help-you-overcome-your-fear-of-selling-49be17f87255",
"items": [
{
"itemType": "blogPost",
"title": "4 Solid Facts To Help You Overcome Your Fear of Selling",
"creators": [
{
"firstName": "Dennis De",
"lastName": "Silva",
"creatorType": "author"
}
],
"date": "2021-07-19T19:18:52.609Z",
"abstractNote": "There is still hope introverts, do not despair",
"blogTitle": "ILLUMINATION-Curated",
"language": "en",
"url": "https://medium.com/illumination-curated/4-solid-facts-to-help-you-overcome-your-fear-of-selling-49be17f87255",
"attachments": [
{
"title": "Snapshot",
"mimeType": "text/html"
}
],
"tags": [],
@ -149,4 +205,4 @@ var testCases = [
]
}
]
/** END TEST CASES **/
/** END TEST CASES **/