Skip to content

Commit

Permalink
Start fixing MathSciNet
Browse files Browse the repository at this point in the history
  • Loading branch information
adam3smith committed Jul 5, 2023
1 parent a6e270e commit 1ede28e
Showing 1 changed file with 55 additions and 72 deletions.
127 changes: 55 additions & 72 deletions AMS MathSciNet.js
Original file line number Diff line number Diff line change
Expand Up @@ -2,44 +2,40 @@
"translatorID": "a354331-981b-43de-a61-bc26dd1be3a9",
"label": "AMS MathSciNet",
"creator": "Simon Kornblith",
"target": "^https?://(mathscinet\\.)?ams\\.[^/]*/mathscinet(\\-getitem\\?|/search/(publications\\.html|publdoc\\.html))",
"target": "^https?://(mathscinet.)?ams.[^/]*/mathscinet(/article\\?|/publications-search\\?)",
"minVersion": "3.0",
"maxVersion": "",
"priority": 100,
"inRepository": true,
"translatorType": 4,
"browserSupport": "gcsibv",
"lastUpdated": "2017-10-20 16:08:00"
"lastUpdated": "2023-07-05 03:32:36"
}

function detectWeb(doc, url) {

var itemType;
if (getSearchResults(doc, true)) {
return "multiple"
} else if (itemType = ZU.xpathText(doc, '//div[@class="headlineMenu"]/*[last()-1]')) {
switch (itemType.trim().toLowerCase()) {
case 'article':
return "journalArticle";
case 'book':
return "book";
case 'chapter':
} else if (url.includes("article?mr=")) {
if (doc.querySelector('div[data-testid="ap-book-isbn"]')) {
if (doc.querySelector('div[data-testid="ap-book-collection"]'))
return "bookSection";
}
else return "book";
}
else return "journalArticle";

}
else return false;
}


function getSearchResults(doc, checkOnly) {
var items = {};
var found = false;
//TODO: adjust the xpath
var rows = ZU.xpath(doc, '//div[@class="headlineText"]');
for (var i=0; i<rows.length; i++) {
//TODO: check and maybe adjust
var href = ZU.xpathText(rows[i], './a[@class="mrnum"]/@href')
//TODO: check and maybe adjust
var title = ZU.xpathText(rows[i], './span[@class="title"]')
// TODO: adjust the CSS selector
var rows = doc.querySelectorAll('.results div.font-weight-bold');
for (let row of rows) {
let href = attr(row, 'a', 'href');
let title = ZU.trimInternal(row.textContent);
if (!href || !title) continue;
if (checkOnly) return true;
found = true;
Expand All @@ -48,64 +44,51 @@ function getSearchResults(doc, checkOnly) {
return found ? items : false;
}


function doWeb(doc, url) {
if (detectWeb(doc, url) == "multiple") {
Zotero.selectItems(getSearchResults(doc, false), function (items) {
if (!items) {
return true;
}
var articles = [];
for (var i in items) {
articles.push(i);
}
ZU.processDocuments(articles, scrape);
});
} else {
scrape(doc, url);
async function doWeb(doc, url) {
if (detectWeb(doc, url) == 'multiple') {
let items = await Zotero.selectItems(getSearchResults(doc, false));
if (!items) return;
for (let url of Object.keys(items)) {
await scrape(await requestDocument(url));
}
}
else {
await scrape(doc, url);
}
}

function scrape(doc, url) {
var host = url.match(/^(.+)\/mathscinet/)
var pub = host[0] + "/search/publications.html?fmt=bibtex";
var MR = ZU.xpathText(doc, '//div[@id="content"]/div[@class="doc"]/div[@class="headline"]/strong[1]');
pub += "&pg1=MR&s1="+MR.replace(/^MR0*/, "");

ZU.doGet(pub, function(text) {
var preRE = /<pre>\s*([\s\S]*?)\s*<\/pre>/g;
var bibTeXString = "";

var m;
while (m = preRE.exec(text)) {
bibTeXString += m[1] + '\n';
}

// import using BibTeX
var translator = Zotero.loadTranslator("import");
translator.setTranslator("9cb70025-a888-4a29-a210-93ec52da40d4");
translator.setString(bibTeXString);
translator.setHandler("itemDone", function(obj, item) {
// Fix/fetch MR number
var mrnumber;
if (item.extra) {
item.extra = item.extra.replace(/^MR:\s*(?:MR)?(\d+).*/gm,
function(m, mr) {
mrnumber = mr;
return 'MR: ' + mr;
});
}

if (mrnumber) {
url = 'https://mathscinet.ams.org/mathscinet-getitem?mr=' + mrnumber;
}
item.attachments.push({title: "MathSciNet Snapshot", document: doc});
item.url = url;
async function scrape(doc, url = doc.location.href) {
Z.debug(url);
let id = url.match(/\?mr=(\d+)/);
if (!id) {
throw("No MR ID, can't proceed")
}
let bibJSONUrl = '/mathscinet/api/publications/format?formats=bib&ids=' + id[1];
Z.debug(bibJSONUrl)
let bibJSON = await requestText(bibJSONUrl);
Z.debug(bibJSON)
// the JSON parser doesn't like newlines or backslashes in the bibtex
bibJSON = bibJSON.replace(/\\n/g , "").replace(/\\/g, "");
bibJSON = JSON.parse(bibJSON);
let bibTex = bibJSON[0].bib;
Z.debug(bibTex)
let translator = Zotero.loadTranslator("import");
translator.setTranslator('9cb70025-a888-4a29-a210-93ec52da40d4');
translator.setString(bibTex);
translator.setHandler('itemDone', (_obj, item) => {
item.attachments.push({
title: 'Snapshot',
document: doc
});
item.complete();
});
translator.translate();
});
await translator.translate();
}





/** BEGIN TEST CASES **/
var testCases = [
{
Expand Down Expand Up @@ -206,7 +189,7 @@ var testCases = [
"seeAlso": []
}
]
},
},
{
"type": "web",
"url": "https://mathscinet.ams.org/mathscinet-getitem?mr=2663710",
Expand Down

0 comments on commit 1ede28e

Please sign in to comment.