[Share] Extract Abstract Keywords

[MuiseDestiny]

Description

PDF's Abstract -> Item's Abstract Note field.
PDF's Keywords -> Item's Tags.

Event

Creat Item

Operation

Script

Data

/**
 * Extract Abstract Keywords
 * @author Polygon
 * @link https://github.com/windingwind/zotero-actions-tags/discussions/136
 */
const ZoteroPane = require("ZoteroPane")
const window = require("window")
const console = window.console
let n = 0
await Promise.all(ZoteroPane.getSelectedItems()
	.map(async (item) => {
	if ((await item.getBestAttachmentState()).exists == false) { return }
	const pdfItem = await item.getBestAttachment()
	if (!pdfItem) { return }
	const text = (await Zotero.PDFWorker.getFullText(pdfItem.id, 2, true)).text
	const res = text.match(/\n(Abstract[\s\S]+?)\nKeywords(.+)/i)
	if (res && res.length == 3) {
		const data = {
			abstract: res[1].replace(/^abstract:?\s*/i, "").replace("\n", " ").replace(/\s+/, " "),
			keywords: res[2].replace(/^\s*:\s*/, "").trim().split(/[;,]\s*/).map(i=>i.trim())
		}
		console.log(data)
		await Promise.all(data.keywords.map(async(tag)=>{
			item.addTag(tag)
		}))
		item.setField("abstractNote", data.abstract)
		await item.saveTx();
		n += 1;
	}
}))
return `Completed ${n}`

Anything else

No response

[windingwind]

The script will be executed for each selected item, and saving item is always called by plugin. So maybe just:

/**
 * Extract Abstract Keywords
 * @author Polygon
 * @link https://github.com/windingwind/zotero-actions-tags/discussions/136
 */
const window = require("window");
const console = window.console;
let n = 0;
if ((await item.getBestAttachmentState()).exists == false) {
  return;
}
const pdfItem = await item.getBestAttachment();
if (!pdfItem) {
  return;
}
const text = (await Zotero.PDFWorker.getFullText(pdfItem.id, 2, true)).text;
const res = text.match(/\n(Abstract[\s\S]+?)\nKeywords(.+)/i);
if (res && res.length == 3) {
  const data = {
    abstract: res[1]
      .replace(/^abstract:?\s*/i, "")
      .replace("\n", " ")
      .replace(/\s+/, " "),
    keywords: res[2]
      .replace(/^\s*:\s*/, "")
      .trim()
      .split(/[;,]\s*/)
      .map((i) => i.trim()),
  };
  console.log(data);
  await Promise.all(
    data.keywords.map(async (tag) => {
      item.addTag(tag);
    })
  );
  item.setField("abstractNote", data.abstract);
}

return "Extracted item keywords and abstract.";

[MediCoderHZ]

你好，运行该脚本没有报错，但是也没有提取出关键词是什么原因呢？PDF文件是使用Zotfile另存的，不是存在Zotero内部的，会有影响吗

[LiWeo]

一样，没有反应😂

[wang-zhu20]

我也是

[CoffeeDrinkRabbit]

我也是

[HuimingPan]

跟PDF格式有关，const res = text.match(/\n(Abstract[\s\S]+?)\nKeywords(.+)/i); 是正则匹配，如果没有提取到 Keywords: 这个字符串就没有办法提取关键字。不同出版商的PDF格式不同，IEEE很多的都是 Index Items-，需要改成：
const res = text.match(/\n(Abstract[\s\S]+?)\nIndex Terms—(.+)/i);

[KagaJiankui]

Zotero版本7.0.0-beta.82+79fa160e4(64-bit), 运行该脚本遇到报错

Script Error: item is null

在脚本中加入debug语句后发现未获取到条目, 但其他需要获取条目的脚本工作正常. 请问一下有没有解决办法? 另外, 能否将添加摘要的功能改成仅在没有摘要时添加来自PDF的摘要, 大概类似下面这个JS片段:

  if(item.getField("abstractNote") === null && data.abstract !== undefined){
	item.setField("abstractNote", data.abstract);
  }

[OESNES]

I would like to follow-up on the Script Error: item is null for copying tags to see if there are any updates? I simply want a way to copy the comma separated list of keywords from my research papers and paste them into the tags section in the required \r\n or \n format.

[dpowless517]

How do I use this template? How do I run an action script?

[nkh0472]

会提取到一些“Biochips · CAD tools · Euler tour · Graph matching · Testing”这样的关键词，需要对·做进一步分词处理