test-equiv-*.R equivalence suites (vs
bibliometrix and biblionetwork) have been
moved out of the package into a local-only
local_testing_and_equivalence/ directory. These developer
checks pulled in data.table/biblionetwork,
whose OpenMP parallelism caused the “CPU time 4 times elapsed time” NOTE
on the Debian r-devel pre-test. They remain runnable locally but are no
longer part of R CMD check.bibliometrix, biblionetwork, and
data.table removed from Suggests — they were
used only by the relocated equivalence tests.tests/testthat.R keeps a 2-thread BLAS/OpenMP cap as
defence for crossprod() / tcrossprod() in
multiply_bipartite().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().