sm_snapshot() serializes an sm_corpus to disk as a compressed RDS file,
embedding the corpus hash as part of the filename for traceability.
sm_snapshot_load() reads a previously saved snapshot and validates that
the loaded object is a well-formed sm_corpus.
The snapshot stores the corpus hash (sm_hash_corpus()) in
metadata$corpus_hash and the version of the hash algorithm in
metadata$hash_version. sm_snapshot_load() recomputes the hash with
that version, or with version 1 for snapshots written before scimapR
0.5.0, and warns when it differs.
Usage
sm_snapshot(
corpus,
path = NULL,
compress = c("xz", "gzip", "bzip2", "none"),
call = rlang::caller_env()
)
sm_snapshot_load(path, call = rlang::caller_env())Arguments
- corpus
An
sm_corpusobject.- path
Character. File path for the snapshot. If
NULL, a path is generated in the current working directory using the corpus hash.- compress
Character. Compression method passed to
saveRDS(). One of"xz"(default, best compression),"gzip","bzip2", or"none".- call
Caller environment for error reporting.
See also
Other reproducibility:
sm_certificate(),
sm_cite_corpus(),
sm_diff_corpora(),
sm_hash_corpus(),
sm_provenance(),
sm_register_input()
Examples
corpus <- sm_example_corpus()
path <- tempfile(fileext = ".rds")
sm_snapshot(corpus, path = path)
#> ✔ Corpus snapshot saved to /tmp/RtmpicujyO/file1c542b920891.rds.
#> ℹ Size: 122K | Hash: 61c0637eede7
loaded <- sm_snapshot_load(path)
#> ✔ Loaded corpus from /tmp/RtmpicujyO/file1c542b920891.rds.
#> ℹ 200 works, 80 authors.
identical(nrow(corpus$works), nrow(loaded$works))
#> [1] TRUE