Crate paper_resolver

Concurrent PDF resolver for academic papers — 9 open-access sources.

What it does

Given a DOI, URL, or title, queries up to 9 academic sources in parallel and returns the best downloadable PDF URL. No Zotero, no reference manager dependency — just (doi, url, title) → Option<ResolvedPdf>.

Sources (by priority)

#	Source	Coverage	Method
1	arXiv	arXiv papers	Instant DOI/URL pattern match (no network)
2	OpenAlex	250M+ works	Structured OA location data
3	CORE	300M+ OA works	Institutional repositories
4	Google Scholar	Widest coverage	HTML scraping (rate-limit risk)
5	Unpaywall	30M+ OA articles	DOI lookup (requires email)
6	Crossref	Publisher links	DOI metadata
7	Zenodo	Cross-disciplinary	CERN preprint repository
8	SSRN	Finance/economics	Preprint abstracts
9	Semantic Scholar	CS/bio/med	OA PDFs + disclaimer parsing

Quick start

// Resolve a PDF by DOI (sync — creates its own tokio runtime):
let result = paper_resolver::resolve_pdf(
    Some("10.48550/arXiv.1706.03762"), // DOI
    None,                              // URL
    Some("Attention Is All You Need"), // title (fallback)
);

if let Some(pdf) = result {
    println!("Found: {} (via {})", pdf.url, pdf.source);
    println!("Downloadable: {}", pdf.downloadable);
}

Detailed reporting

Use resolve_pdf_with_report to see why each source succeeded or failed:

let report = paper_resolver::resolve_pdf_with_report(
    Some("10.1109/TSE.2010.62"), None,
    Some("mutation testing"),
    &paper_resolver::ResolverConfig::default(),
);
println!("{}", report.summary());
// PDF found via google_scholar (downloadable: true)
//   https://mutationtesting.uni.lu/TR-09-06.pdf
//
// Sources queried:
//   openalex: no OA location found
//   google_scholar: found https://...
//   unpaywall: skipped — configure resolver.email
//   ...

Custom configuration

use paper_resolver::{ResolverConfig, SourceEntry};

let mut config = ResolverConfig::default();
config.email = "researcher@university.edu".into();
config.timeout_secs = 10;
config.sources = vec![
    SourceEntry::new("arxiv", true),
    SourceEntry::new("openalex", true),
    SourceEntry::new("unpaywall", true),
    // disable the rest
];

let result = paper_resolver::resolve_pdf_with_config(
    Some("10.1234/example"), None, None, &config,
);

Design decisions

• Concurrent by default: all enabled sources are queried simultaneously via futures::future::join_all. First-to-return wins by priority.
• Blocked domains: publisher paywalls (IEEE, Springer, Elsevier, etc.) are detected and marked downloadable: false rather than silently failing.
• No file I/O: this crate has zero filesystem dependency. Configuration is passed as a struct — the caller owns config file parsing.
• Standalone: no Zotero dependency. Usable in any Rust project that needs academic PDF resolution.

Structs

Endpoints

Base URLs for each source — overridable for testing.

ResolveReport

Detailed resolution result — includes per-source failure reasons.

ResolvedPdf

Result of PDF URL resolution.

ResolverConfig

Configuration for the paper resolver.

SourceEntry

A source entry — name + enabled flag.

Constants

SOURCE_NAMES

All available source names.

Functions

resolve_pdf

Resolve a PDF URL using all available sources (default config).

resolve_pdf_async

Async version with configuration — caller owns the tokio runtime.

resolve_pdf_async_with_report

Async version with detailed per-source reporting.

resolve_pdf_with_config

Resolve a PDF URL with custom configuration.

resolve_pdf_with_report

Resolve with detailed per-source reporting.