[Share] Update item metadata

[windingwind]

Description

Update item metadata.

See also https://forums.zotero.org/discussion/75796/how-to-retrieve-and-update-metadata-in-zotero

Thanks @northword: 👍 https://github.com/northword/zotero-format-metadata/blob/main/src/modules/rules/field-retrive.ts

Event

None

Operation

Script

Data

/**
 * Update journal article metadata from DOI
 * @author windingwind
 * @usage 
 * @link https://github.com/windingwind/zotero-actions-tags/discussions/182
 * @see https://forums.zotero.org/discussion/75796/how-to-retrieve-and-update-metadata-in-zotero
 * @see https://github.com/northword/zotero-format-metadata/blob/main/src/modules/rules/field-retrive.ts
 */
if (!item) {
  return;
}

return await updateMetadataByIdentifier(item);

async function updateMetadataByIdentifier(targetItem, mode = "all") {
  let doi = targetItem.getField("DOI");
  // 不存在 DOI 直接结束
  // todo: 若有附件，尝试从附件获取?
  // todo: 弹出 DOI 输入对话框?
  if (!doi) {
    return `DOI not set on item ${targetItem.getField("title")}`;
  }

  if (doi.match(/arxiv/gi)) {
    const arxivID = doi.replace(/10.48550\/arXiv\./gi, "");
    const tmpDOI = await getDOIFromArxiv(arxivID);
    doi = tmpDOI !== false ? tmpDOI : doi;
  }

  const newItem = await translateByDOI(doi);

  const fields = [
    "title",
    "publicationTitle",
    "journalAbbreviation",
    "volume",
    "issue",
    "date",
    "pages",
    "issue",
    "ISSN",
    "url",
    "DOI",
    "abstractNote",
  ];

  // mode == all: 强制更新，无论原值是否为空：mode == "all" ||
  // 对于一个字段，若 mode == "all"，更新
  //              若 mode == "blank"，且 旧值为空，更新
  //              若 mode == "blank"，且 旧值非空，保持
  //              若 mode == "blank"，且 新值为空，？？

  // 更改 ItemType
  if (mode === "all") {
    newItem["itemTypeID"] = Zotero.ItemTypes.getID(newItem["itemType"]);
    if (newItem["itemType"] !== targetItem.itemType && !doi.match(/arxiv/gi)) {
      targetItem.setType(newItem["itemTypeID"]);
    }
  }

  // 更新 creators
  if (mode === "all" || targetItem.getCreators().length == 0) {
    targetItem.setCreators(newItem["creators"]);
  }

  for (const field of fields) {
    const newFieldValue = newItem[field] ?? "",
      oldFieldValue = targetItem.getField(field);

    // 当新条目该字段未空时，结束本次循环
    // 存疑：当新条目该字段为空时，可能是该字段确实为空，用户已有条目字段可能是假值。
    // if (!newFieldValue) continue;

    if (!(mode === "all" || !oldFieldValue)) continue;

    switch (field) {
      // case "publicationTitle":
      //     // 当原条目存在期刊名时，不替换
      //     if (oldFieldValue == "") {
      //         item.setField(field, newFieldValue);
      //     }
      //     break;

      // case "journalAbbreviation":
      //     if (newFieldValue.length == oldFieldValue.toString().length){
      //         //
      //     }

      //     break;
      case "url":
        // 旧的 url 为空、为 WOS 链接时，更新 url
        if (
          oldFieldValue == "" ||
          (typeof oldFieldValue == "string" &&
            oldFieldValue.includes("webofscience"))
        ) {
          targetItem.setField(field, newFieldValue);
        }
        break;

      case "date":
        targetItem.setField(field, Zotero.Date.strToISO(newFieldValue));
        break;

      default:
        targetItem.setField(field, newFieldValue);
        break;
    }
  }
  return `Metadata updated for item ${targetItem.getField("title")}`;
}

async function getDOIFromArxiv(arxivID) {
  const id = arxivID.replace(/arxiv:/gi, arxivID).trim();
  const url = `https://export.arxiv.org/api/query?id_list=${id}`;

  const res = await Zotero.HTTP.request("GET", url);
  const result = res.response;
  if (result == "" || result == null || result == undefined) {
    return false;
  }
  const doc = new DOMParser().parseFromString(result, "text/xml");
  const refDoi = doc.querySelector("doi");
  if (refDoi) {
    return refDoi.innerHTML;
  } else {
    return false;
  }
}


async function translateByDOI(doi) {
  const identifier = {
    itemType: "journalArticle",
    DOI: doi,
  };

  const translate = new Zotero.Translate.Search();
  translate.setIdentifier(identifier);
  const translators = await translate.getTranslators();
  translate.setTranslator(translators);

  // {libraryID: options} 避免条目保存
  // https://github.com/zotero/translate/blob/05755f5051a77737c56458440c79964c7a8874cf/src/translation/translate.js#L1208-L1210
  // 配置这一项后返回的不再是 Zotero.Item[]，而是一个包含字段信息的 Object[]
  const newItems = await translate.translate({ libraryID: false });
  const newItem = newItems[0];
  return newItem;
}

Anything else

Please make sure DOI exists on the target item.

[Zz-dong]

Awesome! Is it possible to support ISBN?

[zararash]

While updating the metadata, the item which has abstract and was added manually gets removed. Is there a way to tell the script not to remove the abstract which is added manually?

[northword]

NOT TESTED!

  for (const field of fields) {
    const newFieldValue = newItem[field] ?? "",
      oldFieldValue = targetItem.getField(field);
    // if (!newFieldValue) continue;

    if (!(mode === "all" || !oldFieldValue)) continue;

    switch (field) {
      default:
+++        if (field == "abstractNote" && newFieldValue == "") break;
        targetItem.setField(field, newFieldValue);
        break;
    }
  }
  return `Metadata updated for item ${targetItem.getField("title")}`;
}

[northword]

@Zz-dong @zararash The linter plugin now supports updating metadata by ISBN, as well as preserving non-empty fields.

If you need to use the Action plugin, you can refer to the following code to modify the script.

https://github.com/northword/zotero-format-metadata/blob/main/src/modules/rules/retrive-metadata.ts

[ljci]

@Zz-dong @zararash The linter plugin now supports updating metadata by ISBN, as well as preserving non-empty fields.

If you need to use the Action plugin, you can refer to the following code to modify the script.

https://github.com/northword/zotero-format-metadata/blob/main/src/modules/rules/retrive-metadata.ts
thanks for your hard work, I am wondering if we can extract the meta data for AI conference paper, conference name, proceeding title, publisher and so on. None of software can do it except paperlib that I just knew but I don't want to switch the software, so it will be perfect if we can extract meta data for conference like, ICCV, ICML, NeurIPS, ECCV and so on. Thanks