Copies every local data.txt of a metadata tibble byte for byte to
<path>/<name>.json, named by a column of the metadata (the study result
id by default, or worker_id, query_prolific_pid, or a field from
jatos_extract_fields()), which is the layout labs share raw JSON in.
Nothing is parsed or rewritten. A study result with several files (one
per component) gets <name>_<component_result_id>.json for each. When
two different study results share a name, for example a participant who
ran twice, nothing is written and the error lists the name and the study
result ids, so that duplicates are resolved on purpose rather than
silently. Each copy is written under a temporary name in path and
renamed into place, so a copy that fails never sits under the real
name.
Arguments
- metadata
A metadata tibble with the
filecolumn, fromjatos_download_results()orjatos_read_metadata(). Rows without a local file are left out with a message.- path
Directory to write into; created when missing.
- name_by
Name of the metadata column whose value names each file.
NAvalues are refused, and so is a value that could leavepath: one holding a slash or a backslash, or equal to.or... The values may come from the participants themselves (a Prolific id in the URL, a field of the result data), so they are never trusted as paths.- overwrite
If
TRUE, existing files inpathare replaced.
Value
A tibble with study_result_id, component_result_id, file
(the source) and path (the copy), invisibly; a message reports the
count.
See also
jatos_export_archive() for the server's own zip of the same
files.
Examples
cache <- system.file("extdata", "JATOS_data", package = "jatosr")
meta <- jatos_read_metadata(cache)
out <- tempfile("jatosr-example-")
jatos_write_raw(meta, out, name_by = "worker_id")
#> ℹ 2 rows have no local file.
#> ✔ Wrote 2 files to /tmp/RtmpHTC4Nw/jatosr-example-19043d2d7b22.
list.files(out)
#> [1] "501.json" "502.json"
unlink(out, recursive = TRUE)