Skip to contents

Scan Project Dependencies (Robust AST Method)

scan_dependencies.Rd

Scans a specified project directory to detect which R packages are actually being used in the code (via library(), ::, or Roxygen tags) and compares them against the DESCRIPTION file.

This helps identify:

  • Ghosts: Packages used in the code but missing from DESCRIPTION.

  • Unused: Packages declared in DESCRIPTION but never used in code.

  • Bloat: A frequency table of how often each package is called.

Usage

scan_dependencies(dir_path)

Arguments

dir_path

A character string specifying the path to the project root. Must contain R files and optionally a DESCRIPTION file.

Value

A named list containing:

  • usage_stats: A data frame of package usage counts.

  • undeclared_ghosts: A character vector of packages used but not listed in DESCRIPTION.

  • unused_declarations: A character vector of packages listed in DESCRIPTION but not found in code.

Examples

# 1. Create a temporary "Fake Project" for testing
tmp_proj <- tempfile("test_project")
dir.create(tmp_proj)

# 2. Create a dummy DESCRIPTION file
desc_content <- c(
  "Package: TestProj",
  "Imports: dplyr, fs"
)
writeLines(desc_content, file.path(tmp_proj, "DESCRIPTION"))

# 3. Create a dummy R script
script_content <- c(
  "library(dplyr)",
  "x <- tidyr::pivot_longer(mtcars)"
)
writeLines(script_content, file.path(tmp_proj, "script.R"))

# 4. Run the scanner
# (We wrap in if/try in case internal dependencies aren't loaded in this example context)
if (requireNamespace("fs", quietly = TRUE)) {
  try({
    result <- scan_dependencies(tmp_proj)
    print(result$undeclared_ghosts)
  })
}
#> [1] "tidyr"

# 5. Cleanup
unlink(tmp_proj, recursive = TRUE)