Fetch works from Semantic Scholar

Query the Semantic Scholar Academic Graph API for papers and return the results as an sm_corpus.

Supports both keyword search (query) and batch paper lookup (paper_ids). Optionally fetches pre-computed SPECTER embeddings.

Usage

sm_fetch_semantic_scholar(
  query = NULL,
  paper_ids = NULL,
  n_max = 100L,
  api_key = Sys.getenv("SEMANTIC_SCHOLAR_API_KEY"),
  include_embeddings = FALSE,
  verbose = TRUE,
  call = rlang::caller_env()
)

Arguments

query

Free-text search query. If NULL, paper_ids must be supplied.

paper_ids

Character vector of Semantic Scholar paper IDs, DOIs (prefixed DOI:), arXiv IDs (prefixed ARXIV:), or PMIDs (prefixed PMID:). If NULL, query must be supplied.

n_max

Maximum number of papers to return (default 100).

api_key

Semantic Scholar API key. Read from SEMANTIC_SCHOLAR_API_KEY env var by default.

include_embeddings

Logical; if TRUE, request SPECTER embeddings for each paper and attach as an embedding matrix.

verbose

Print progress messages?

call

Caller environment for error reporting.

Value

An sm_corpus object.

See also

Other fetchers: sm_fetch_arxiv(), sm_fetch_biorxiv(), sm_fetch_crossref(), sm_fetch_openalex(), sm_fetch_orcid(), sm_fetch_overton(), sm_fetch_pubmed()

Examples

if (FALSE) { # \dontrun{
corpus <- sm_fetch_semantic_scholar(query = "bibliometrics", n_max = 10)
print(corpus)
} # }