read_scopus, read_wos, read_dimensions, read_lens) now use
small bundled fixtures under inst/extdata/, reached via
system.file(). API-wrapper readers (read_openalex, read_crossref)
use an inline data frame matching the upstream column shape so the
conversion path runs without a network call. read_biblio examples
now demonstrate multi-file, directory, and generic-CSV modes against
the bundled fixtures.inst/extdata/scopus_sample.csv, wos_sample.txt,
dimensions_sample.csv, lens_sample.csv (2 records each).read_lens() no longer inflates output to n^2 rows when neither
Lens ID nor ID columns are present.read_openalex() no longer inflates output to n^2 rows when the id
column is absent.read_scopus() now normalises empty-string DOIs to NA, so
is.na(doi) deduplication checks behave as expected.read_wos() empty-file return now includes the keywords_plus
list-column to match the non-empty schema.read_crossref() no longer crashes with "row names contain missing
values" when the issued column has NA entries.to_igraph(), to_tbl_graph(), to_cograph()) now
use @examplesIf requireNamespace(...) so they execute when the
suggested package is installed instead of being silently skipped.read_biblio(), read_bibtex(), and read_ris() now ship runnable
examples backed by either the bundled extdata/openalex_works.csv
fixture or a tempfile()-based minimal record.read_scopus(), read_wos(),
read_ris(), read_lens(), read_dimensions(), read_crossref(),
read_biblio(), read_openalex(), plus dedicated coverage for
R/edgelist.R and build_bipartite_long().temporal_network() — builds time-windowed networks with fixed, sliding, or
cumulative strategies. Results include a window column for easy stacking.historiograph() — Garfield-style chronological citation network among the
most locally cited documents.local_citations() — counts within-dataset citations (Local Citation Score).backbone() — disparity filter for extracting statistically significant
edges from dense weighted networks.prune() — threshold and top-n edge pruning.read_biblio() — universal reader with auto-format detection (Scopus, WoS,
BibTeX, RIS, Dimensions, Lens.org).read_dimensions() — Dimensions CSV export reader.read_crossref() — converter for rcrossref::cr_works() output.to_gephi() — exports node and edge tables in Gephi CSV format; writes
nodes.csv + edges.csv when a directory path is supplied.to_graphml() — pure base-R GraphML writer; no XML package required.to_cograph() — converts edge list to a cograph_network object with
optional node metadata for direct use with cograph::splot().weight descending and reset
row names.local_citations() canonical column order: id, lcs, gcs, year,
title, journal, doi.historiograph() empty-result schema matches non-empty schema.id, title, year, journal,
doi, cited_by_count, abstract, type, authors, references,
keywords, then source-specific extras.backbone() and prune() use single-pass O(m) node statistics via
tapply() / split() — faster on large networks.temporal_network() converted from for loop to lapply.read_dimensions() / read_crossref() now apply standardize_authors()
and standardize_refs() for consistency with other readers.count renamed to counting; measure renamed to similarity
across all network functions.co_network() renamed to conetwork().read_openalex() — reads OpenAlex JSON export.filter_top() — keeps only the top-n most connected nodes.normalize() — post-hoc normalisation of any edge list.Initial release.
author_network(), document_network(),
reference_network(), keyword_network(), institution_network(),
country_network(), source_network(), conetwork().to_igraph(), to_tbl_graph(), to_matrix().