scholar

scholar-nulis — literature search and citation verification

Baca dalam bahasa Indonesia

scholar-nulis 0.8.0 · 16 tools, or 21 with an Elsevier key · source

Searches seven open scholarly APIs, verifies citations by DOI, finds legal open-access PDFs, downloads them, and reads their text. Five Scopus/ScienceDirect tools switch on if you have Elsevier credentials.

What it is actually for

Four situations that come up constantly in real writing, and what you do about them. Every output below is real — copied from an actual run, not illustrative.

1. You have a reference list and you are not certain all of it is real

The failure mode is specific. Invented references do not look invented; they look ordinary. A plausible author, a plausible year, a plausible journal. Reading them more carefully will not help, because there is nothing on the page to notice.

"Check every DOI in my reference list against Crossref and tell me which ones do not resolve."

A real one comes back with its registered record:

{ "verified": true, "retracted": false,
  "paper": { "title": "The PRISMA 2020 statement: an updated guideline for reporting
             systematic reviews", "authors": ["Matthew J Page", "Joanne E McKenzie", …] } }

An invented one comes back unambiguously:

{ "verified": false,
  "doi": "10.1016/j.jclinepi.2023.99999",
  "note": "DOI tidak ditemukan di Crossref — tandai [VERIFY]." }

That is the whole point of the tool. It converts "this looks fine" into a yes or a no.

2. You are about to cite a study that has been retracted

Citing retracted work is a serious error, and nothing in the PDF you downloaded will necessarily tell you. nulis_get_paper_by_doi checks retraction status on every lookup — you do not have to ask for it:

{ "verified": true,
  "retracted": true,
  "retraction_evidence": {
    "dari_judul": "RETRACTED: Ileal-lymphoid-nodular hyperplasia, non-specific colitis,
                   and pervasive developmental disorder in children" },
  "peringatan": "STUDI INI DICABUT (retracted) … Jangan disintesis. Bila tetap dibahas
                 karena alasan tertentu, nyatakan statusnya secara eksplisit di teks." }

Note what it does not do: it does not delete the reference for you. A retracted study is sometimes cited deliberately — in a paper about research integrity, for instance. It tells you, and leaves the decision where it belongs.

3. You need the actual PDF, legally

"Find me an open-access PDF for this DOI."

{ "found": true, "via": "unpaywall", "oa_status": "hybrid", "license": "cc-by",
  "pdf_url": "https://www.bmj.com/content/bmj/372/bmj.n71.full.pdf" }

The license field matters more than it looks: it is the difference between "I may read this" and "I may redistribute this to my students". Without CONTACT_EMAIL the same lookup still succeeds through OpenAlex, but returns no licence.

4. You are running a systematic search that a reviewer will re-run

nulis_search_scopus returns the total hit count for the identification box of your PRISMA diagram, and nulis_scopus_export_csv pages through the whole result set into a screening-ready file. Because the query is passed through untouched, the number you report and the string you report agree with each other — see below.

Search — no key required

Tool	Purpose
nulis_search_arxiv	Search arXiv
nulis_search_openalex	Search OpenAlex (year filter)
nulis_search_crossref	Search Crossref
nulis_search_semantic_scholar	Search Semantic Scholar
nulis_search_pubmed	Search PubMed
nulis_search_europepmc	Search Europe PMC
nulis_search_doaj	Search open-access journal articles in DOAJ

All seven work without registration. Setting CONTACT_EMAIL puts you in the Crossref and OpenAlex polite pool — a more generous quota, and your requests are no longer treated as anonymous traffic.

Verification and access

Tool	Purpose
nulis_get_paper_by_doi	Verify and retrieve paper metadata by DOI (Crossref)
nulis_get_open_access_pdf	Find a legal open-access PDF link for a DOI
nulis_download_pdf	Download a PDF from a URL
nulis_download_arxiv	Download an arXiv paper's PDF
nulis_read_arxiv_paper	Read the full text of an arXiv paper
nulis_read_pdf	Extract text from a PDF (URL or local file)
nulis_pdf_to_text	Extract a PDF's text and save it as .md under fulltext/
nulis_batch_acquire_pdfs	Attempt PDF downloads for a whole set of studies at once

nulis_get_paper_by_doi is the most important tool here. It resolves a DOI against Crossref and returns what is actually registered — authors, title, journal, year. This is the only way to tell a real citation from a fabricated one: a plausible-looking author–year–journal combination is the signature pattern of invented references, and no amount of careful reading substitutes for asking the registrar.

nulis_get_open_access_pdf only points at copies that are genuinely, legally open. It does not look for pirated copies. It works without CONTACT_EMAIL by going through OpenAlex; setting the email adds the Unpaywall route, which also reports the licence of the copy it found — worth having when you need to know whether you may redistribute it.

Scopus & ScienceDirect — Elsevier key required

The five tools below are not registered at all when SCOPUS_API_KEY and SCIENCEDIRECT_API_KEY are empty. They do not appear and then fail when called; they simply are not there — tools/list returns 16 instead of 21.

Tool	Purpose
nulis_search_scopus	Search Scopus using its own query syntax, passed through verbatim
nulis_scopus_abstract	Full abstract, author keywords, citation count (by DOI or Scopus ID)
nulis_sciencedirect_fulltext	ScienceDirect full text by DOI
nulis_scopus_export_csv	Run a query, collect every page, save as screening-ready CSV
nulis_elsevier_status	Check that keys are present, accepted, and what quota remains

Queries are passed through verbatim

nulis_search_scopus accepts native Scopus syntax — TITLE-ABS-KEY, AND/OR/NOT, W/n, PUBYEAR, DOCTYPE, LANGUAGE, SRCTYPE — and does not translate it:

TITLE-ABS-KEY("islamic contract" W/3 freedom) AND PUBYEAR > 2014 AND DOCTYPE(ar)

This is a design decision, not a limitation. The search string you report in Methods must be identical to the one actually executed. The moment some layer silently normalises or "fixes" your query, the reproducibility claim in your manuscript stops being true — and a reviewer re-running your query gets a different number with nothing to explain it.

For the same reason, nulis_search_scopus returns total hits — the number that goes in the identification box of a PRISMA flow diagram.

Run nulis_elsevier_status before you start

Elsevier quota is tied to an institutional subscription and can run out mid-way. nulis_elsevier_status reports whether the key is read, whether it is actually accepted, and how much quota is left — far cheaper than discovering it in the middle of a systematic search.

If off-campus access is refused with 401/403, ask your librarian or licence admin for an insttoken and set ELSEVIER_INSTTOKEN.

Keys never leak through error messages

Elsevier errors routinely echo the full URL with apiKey= in it. Every error message here passes through scrub(), which replaces the key and any apiKey=… pattern with «redacted» before it reaches the caller.

Diagnostics

nulis_server_status reports the running version, the folder PDFs will be saved to, and which optional features are active. It is the right first step when something behaves unexpectedly — including confirming that Claude Desktop really loaded the version you just installed.

Download folder

The order used: DOWNLOAD_DIR if set to a valid value → ~/Downloads → the system temp folder. .md files produced by nulis_pdf_to_text go into a fulltext/ subfolder.

---

← Back · Installation · zotero · scr-toolkit · FAQ