4  PDF (.pdf) Documents

Author

Daniel Manrique-Castano

Published

December 18, 2025

4.1 Overview

PDF is considered the standard for fixed-layout documents. However, not all PDFs are created in the same manner.

NoteCuration Goal

Validate the archival suitability of PDF documents. Our objective is to distinguish between high-quality PDF/A files and “Image-Only” scans, ensuring content is searchable, accessible, and free from encryption that would lock data away.

WarningPreservation Risk

Not all PDFs are equal. Encrypted files, image-only scans without OCR, and documents with external font dependencies pose significant risks for long-term accessibility and automated text mining.

This notebook evaluates PDF files on three critical dimensions:

  1. Security: Detecting Encryption or password protection.
  2. Accessibility: Categorizing files as Searchable or Image-Only documents.
  3. Metadata Quality: Checking for embedded Title and Author fields.

4.2 Setup

We use pdftools for PDF analysis and tidyverse for reporting.

4.2.1 R Packages

Code 
# install.packages(c("tidyverse", "pdftools", "rstudioapi"))

4.2.2 Load libraries

Code 
library(tidyverse)
library(pdftools)
library(rstudioapi)

4.3 Select Target Directory

Code 
if (interactive() && .Platform$OS.type == "windows") { 
  selected_dir <- rstudioapi::selectDirectory(caption = "Select PDF Directory")
} else {
  selected_dir <- NULL
}

if (!is.null(selected_dir)) {
  target_dir <- selected_dir
} else {
  target_dir <- params$target_dir
}

print(paste("Analyzing directory:", target_dir))
[1] "Analyzing directory: data/Inspect_pdf/"

4.4 Find PDF Files

Code 
pdf_files <- list.files(
  path = target_dir,
  pattern = "\\.pdf$", 
  recursive = TRUE, 
  full.names = TRUE, 
  ignore.case = TRUE
)

print(paste("Found", length(pdf_files), "PDF files."))
[1] "Found 6 PDF files."

4.5 File inventory and inspection

We iterate through the files to extract metadata, check for encryption, and sample the text content to determine its accessibility status.

Code 
message("Generating PDF Report...")

results_list <- lapply(pdf_files, function(file_path) {
  fname <- basename(file_path)
  
  tryCatch({
    # 1. Metadata & Security
    info <- pdftools::pdf_info(file_path)
    
    # NEW: Capture Encryption Status explicitly
    is_encrypted <- isTRUE(info$encrypted)
    
    # 2. Text Extraction
    # If encrypted, text extraction might fail, so we wrap it or handle it
    first_page_text <- tryCatch(pdftools::pdf_text(file_path)[1], error = function(e) "")
    
    # Handle NA/NULL text
    if(is.na(first_page_text) || is.null(first_page_text)) first_page_text <- ""
    
    char_count <- nchar(trimws(first_page_text))
    has_text <- char_count > 10 
    
    # 3. Fonts
    fonts <- tryCatch(pdftools::pdf_fonts(file_path), error = function(e) NULL)
    font_names <- if (!is.null(fonts)) paste(head(unique(fonts$name), 5), collapse = ", ") else "Unknown"
    
    # Helper for metadata
    get_meta <- function(x) {
      if (is.null(x) || length(x) == 0 || x == "") return("Unknown")
      return(as.character(x))
    }

    # Return valid tibble with explicit types
    tibble(
      FileName = fname,
      Pages = as.integer(info$pages),
      Encrypted = as.logical(is_encrypted), # <--- NEW COLUMN
      Author = get_meta(info$keys$Author),
      Title = get_meta(info$keys$Title),
      Creator = get_meta(info$keys$Creator),
      Created = as.character(info$created),
      HasText = as.logical(has_text),
      FirstPageChars = as.integer(char_count),
      Fonts = as.character(font_names),
      Status = "Success"
    )
    
  }, error = function(e) {
    tibble(
      FileName = fname,
      Pages = as.integer(NA),
      Encrypted = as.logical(NA),
      Author = as.character(NA),
      Title = as.character(NA),
      Creator = as.character(NA),
      Created = as.character(NA),
      HasText = as.logical(NA),
      FirstPageChars = as.integer(NA),
      Fonts = as.character(NA),
      Status = paste("Failed:", e$message)
    )
  })
})

report <- bind_rows(results_list)

# Display preview
print("--- PDF Report Preview ---")
[1] "--- PDF Report Preview ---"
Code 
print(head(report))
# A tibble: 6 × 11
  FileName   Pages Encrypted Author Title Creator Created HasText FirstPageChars
  <chr>      <int> <lgl>     <chr>  <chr> <chr>   <chr>   <lgl>            <int>
1 20230724_…     7 FALSE     Unkno… Unkn… Unknown 2024-0… TRUE                57
2 example_1…     1 FALSE     Unkno… Unkn… Matplo… 2024-0… FALSE                0
3 Oil_capac…     4 FALSE     Unkno… Oil_… TextEd… 2025-0… TRUE              2883
4 Oil_Trans…     1 FALSE     Unkno… Oil_… TextEd… 2025-0… TRUE              1982
5 Past_Spil…     2 FALSE     Unkno… Past… TextEd… 2025-0… TRUE              2819
6 Spill_eve…     1 FALSE     Unkno… Spil… TextEd… 2025-0… TRUE               814
# ℹ 2 more variables: Fonts <chr>, Status <chr>

4.6 Save Results

Code 
output_dir <- file.path("Results", "Inspect_pdf")
dir.create(output_dir, recursive = TRUE, showWarnings = FALSE)

output_file <- file.path(output_dir, paste0("PDF_Report_", Sys.Date(), ".csv"))

write.csv(report, output_file, row.names = FALSE)

print(paste("Report saved to:", output_file))
[1] "Report saved to: Results/Inspect_pdf/PDF_Report_2026-05-15.csv"

4.7 Curation Insights

Use the generated CSV report to guide your preservation actions:

  • Encryption (Encrypted = TRUE): Encrypted files cannot be migrated, indexed, or arguably “preserved.” Curators must contact the depositor immediately to remove the password. If the password is lost, the file cannot be reused.

  • Accessibility (TextStatus = “Image-Only”): These are scanned images. They are readable by humans but invisible to search engines and screen readers (accessibility violation). You can run these files through an OCR tool (like Tesseract or Adobe Acrobat) to generate a text layer.

  • Font Embedding (Fonts): If Fonts is empty or lists generic names like “Helvetica” without embedding, the layout may break on future systems. the curator can convert the file to PDF/A-1b or PDF/A-2b, which forces font embedding.

4.8 Additional Tools

  • VeraPDF: This is the standard open-source validator for PDF/A compliance. It checks if fonts are embedded and if multimedia is excluded (see https://verapdf.org).

  • Tesseract OCR: Is an open-source engine for optical character recognition . It can take an “Image-Only” PDF and output a searchable text file or a new PDF with a text layer.

  • Google AI: Extractd text and data from images and documents using AI capabilities. See a use case in another chapter of this book.

  • Ghostscript: This is a command-line interpreter often used to repair corrupted PDFs or convert standard PDFs into PDF/A format.

4.9 Using the Non-Interactive R Script

For users who want to run this analysis on a server, in a batch job, or from the command line, here is a pure R script that performs the same process.

Download the R Script: Inspect_PDF_Script.R

4.9.1 Example HPC Submission Script

#!/bin/bash
#SBATCH --job-name=pdf_inspect
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --time=00:30:00
#SBATCH --mem=8G
#SBATCH --output=logs/pdf_inspect_%j.log

module load R

# Define target directory
TARGET_DIR="/scratch/user/project_data/documents"

# Prepare output folders
mkdir -p Results/Inspect_pdf
mkdir -p logs

# Run
echo "Starting PDF Inspection on $TARGET_DIR"
Rscript Inspect_PDF_Script.R "$TARGET_DIR"

4.10 References