Write a Salmon Data Package from preassembled metadata — write_salmon

Advanced/manual writer for cases where you already have the canonical Salmon Data Package (SDP) metadata tables assembled. It writes the SDP CSV metadata files under metadata/ (dataset.csv, tables.csv, column_dictionary.csv, and optional codes.csv) plus the data resource files themselves under data/. For interoperability with Frictionless-style tooling, the function also emits a derived datapackage.json descriptor at the package root.

Usage

write_salmon_datapackage(
  resources,
  dataset_meta,
  table_meta,
  dict,
  codes = NULL,
  path,
  format = "csv",
  overwrite = FALSE,
  write_datapackage = TRUE,
  prune = FALSE
)

Arguments

resources: Named list of data frames/tibbles (one per resource)
dataset_meta: Tibble with dataset-level metadata (one row)
table_meta: Tibble with table-level metadata (one row per table)
dict: Dictionary tibble with column definitions
codes: Optional tibble with code lists
path: Character; directory path where package will be written
format: Character; resource format: "csv" (default, only format supported)
overwrite: Logical; if FALSE (default), errors when path is a directory that already holds something. An existing but completely empty directory is written into without overwrite — there is nothing there to destroy — while a dot-file, a stale .sdp-package ownership sentinel, or an empty data/ subdirectory all count as content and still require it. If TRUE, the package is updated in place — see prune. Replacement is only allowed for a directory recognised as a package: one holding the shared .sdp-package ownership sentinel, or its SDP metadata. An older .metasalmon-package sentinel on its own is not recognised.
write_datapackage: Logical; if TRUE (default), write a root datapackage.json descriptor declaring the SDP Frictionless profile after package validation passes. Use FALSE for draft authoring output.
prune: Logical; if FALSE (default), only files this writer owns are replaced: the metadata/ SDP CSVs, the data/ resources declared in tables.csv (including any a previous write declared and this one does not), datapackage.json, and the ownership sentinel. Everything else is preserved — reviewed SSSOM mappings and measurement decompositions under metadata/semantic/, EML and EDH XML, eml-mapping.yml, review notes, and publication/ artifacts. If TRUE, every entry in the directory is deleted first (the pre-0.2.0 behaviour). Requires overwrite = TRUE.

Value

Invisibly returns the path to the created package

Details

The SDP CSV files remain the canonical package metadata. datapackage.json is a convenience export, not the source of truth.

The write is transactional over the files it owns: every output is fully rendered before anything on disk is deleted or replaced, and a failure while installing the rendered files restores the previous package. An error partway through therefore leaves an existing package intact rather than destroyed. The one exception is prune = TRUE, which deletes files the writer does not own and so cannot restore: the wipe still happens only after every input-dependent step has succeeded, but a filesystem failure (disk full, permissions) between the wipe and the install can lose the deleted files.

Examples

# Create a simple package
resources <- list(main_table = mtcars)
dataset_meta <- tibble::tibble(
  dataset_id = "test-1",
  title = "Test Dataset",
  description = "A test dataset"
)
table_meta <- tibble::tibble(
  dataset_id = "test-1",
  table_id = "main_table",
  file_name = "data/main_table.csv",
  table_label = "Main Table"
)
dict <- infer_dictionary(mtcars, dataset_id = "test-1", table_id = "main_table")
write_salmon_datapackage(
  resources, dataset_meta, table_meta, dict,
  path = tempfile("sdp-example-")
)
#> ✔ Dictionary validation passed
#> ✔ Created Salmon Data Package at /var/folders/pm/twz8_z1j6_zb996w0b17bz2r0000gn/T//RtmpP4K6i1/sdp-example-1101160f75d73