After pii_redact_files() has written a redacted copy beside each flagged
file, adds each original to its repository's .gitignore, so the clean
copy can be committed and the original cannot be by accident. Each rule
names one file. Every rule added is checked with git check-ignore, and a
redacted copy that an existing rule would also hide is reported.
Arguments
- report
The result of
pii_redact_files().
Value
A tibble with one row per redacted original: source, the
gitignore file changed, the action taken (added, already ignored
or not in a git repository), whether the original is_tracked, and
detail.
Details
It does not rewrite history. An original already committed stays in the
repository's past, and stays tracked until removed with
git rm --cached; such files are reported with is_tracked.
Examples
project <- file.path(tempdir(), "survey_project")
dir.create(project)
system2("git", c("init", "-q", project))
write.csv(
data.frame(email = paste0("person", 1:20, "@example.org")),
file.path(project, "survey.csv"),
row.names = FALSE
)
report <- pii_redact_files(project)
#> Wrote 1 redacted copy.
#> ℹ The originals are unchanged and still hold the personal data.
pii_ignore_originals(report)
#> Added 1 original to .gitignore.
#> # A tibble: 1 × 5
#> source gitignore action is_tracked detail
#> <chr> <chr> <chr> <lgl> <chr>
#> 1 /tmp/RtmpeoiXV6/survey_project/survey.csv /tmp/Rtmpe… added FALSE NA
readLines(file.path(project, ".gitignore"))
#> [1] "# Originals holding personal data; redacted copies sit beside them (piinata)"
#> [2] "/survey.csv"
unlink(project, recursive = TRUE)