Citeseerx API

https://citeseerx.ist.psu.edu/api

# Keyword search
curl "https://citeseerx.ist.psu.edu/api/search?q=graph+neural+networks&start=0&rows=20"

# Search by title
curl "https://citeseerx.ist.psu.edu/api/search?q=title:attention+is+all+you+need"

# Search by author
curl "https://citeseerx.ist.psu.edu/api/search?q=author:hinton&rows=25"

# Filter by year
curl "https://citeseerx.ist.psu.edu/api/search?q=federated+learning&year=2024"

# Sort by citation count
curl "https://citeseerx.ist.psu.edu/api/search?q=reinforcement+learning&sort=citationCount+desc"

# Get document metadata
curl "https://citeseerx.ist.psu.edu/api/document?doi=10.1.1.123.456"

# Get citations for a document
curl "https://citeseerx.ist.psu.edu/api/citations?doi=10.1.1.123.456"

# Get citing documents
curl "https://citeseerx.ist.psu.edu/api/citedby?doi=10.1.1.123.456"

{
  "response": {
    "numFound": 5200,
    "docs": [
      {
        "id": "10.1.1.123.456",
        "title": "Graph Neural Networks: A Review",
        "authors": ["Zhou, Jie", "Cui, Ganqu"],
        "year": 2020,
        "abstract": "Graph neural networks have been widely applied...",
        "venue": "AI Open",
        "citationCount": 3500,
        "url": "https://citeseerx.ist.psu.edu/doc/10.1.1.123.456"
      }
    ]
  }
}

import requests

BASE_URL = "https://citeseerx.ist.psu.edu/api"


def search_citeseerx(query: str, rows: int = 20,
                     sort_by_citations: bool = False) -> list:
    """Search CiteSeerX computer science literature."""
    params = {
        "q": query,
        "rows": rows,
        "start": 0,
    }
    if sort_by_citations:
        params["sort"] = "citationCount desc"

    resp = requests.get(f"{BASE_URL}/search", params=params, timeout=30)
    resp.raise_for_status()
    data = resp.json()

    results = []
    for doc in data.get("response", {}).get("docs", []):
        results.append({
            "id": doc.get("id"),
            "title": doc.get("title"),
            "authors": doc.get("authors", []),
            "year": doc.get("year"),
            "venue": doc.get("venue"),
            "citations": doc.get("citationCount", 0),
            "abstract": doc.get("abstract", "")[:300],
            "url": doc.get("url"),
        })
    return results


def get_citations(doc_id: str) -> list:
    """Get papers cited by a document."""
    resp = requests.get(
        f"{BASE_URL}/citations",
        params={"doi": doc_id},
        timeout=30,
    )
    resp.raise_for_status()
    return resp.json().get("citations", [])


def get_cited_by(doc_id: str) -> list:
    """Get papers that cite a document."""
    resp = requests.get(
        f"{BASE_URL}/citedby",
        params={"doi": doc_id},
        timeout=30,
    )
    resp.raise_for_status()
    return resp.json().get("citedby", [])


# Example: find most-cited CS papers on a topic
papers = search_citeseerx("knowledge distillation",
                          rows=10, sort_by_citations=True)
for p in papers:
    print(f"[{p['year']}] {p['title']} (cited: {p['citations']})")

# Example: citation chain analysis
if papers:
    refs = get_citations(papers[0]["id"])
    print(f"\nReferences of top paper ({len(refs)} citations):")
    for r in refs[:5]:
        print(f"  -> {r.get('title', 'Unknown')}")