Attachment 174126 Details for Bug 39667 – R script to process and visualise file sizes

R script to process and visualise file sizes

process.R (text/x-r-source), 2.35 KB, created by Stéphane Guillou (stragu) on 2021-08-07 11:27:54 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Stéphane Guillou (stragu)

Created: 2021-08-07 11:27:54 UTC

Size: 2.35 KB

patch

obsolete

>#### Get sizes of PDFs ####
>
># function to get file size and file name in a dataframe
>get_sizes <- function(directory) {
>  size <- file.size(dir(directory, full.names = TRUE))
>  data.frame(file = dir(directory),
>             size)
>}
>
># get names, sizes and stages
>library(dplyr)
>source_files <- get_sizes("source_files") |>  
>  mutate(what = "source")
>pdf_export <- get_sizes("pdf_export") |> 
>  mutate(what = "pdf")
>pdfTagged_export <- get_sizes("pdfTagged_export") |> 
>  mutate(what = "pdfTagged")
>pdfUA_export <- get_sizes("pdfUA_export") |> 
>  mutate(what = "pdfUA")
>
># bind into single dataframe
>library(stringr)
>long <- bind_rows(source_files, pdf_export, pdfTagged_export, pdfUA_export) |>
>  # remove file extensions
>  mutate(file = str_extract(file, "^[^\\.]+"))
># names are unique as an ID number was appended before converting to PDF
>
># make dataframe wide, calculate rate of change
>library(tidyr)
>wide <- long |> 
>  pivot_wider(names_from = what,
>              values_from = size) |> 
>  mutate(change_tagged = pdfTagged / pdf,
>         change_UA = pdfUA / pdf)
>
>#### Visualise ####
>
>library(ggplot2)
># make dataframe tidy (i.e. long) for ggplot2
>tidied <- wide |> 
>  select(file, change_tagged, change_UA) |> 
>  pivot_longer(-file,
>               names_to = "type", values_to = "change")
># separate dataframe for each facet's median
>medians <- tidied |>
>  group_by(type) |>
>  summarise(median = median(change, na.rm = TRUE))
>
>tidied |> 
>  ggplot(aes(x = change)) +
>  geom_histogram(binwidth = 0.5, boundary = 1) +
>  geom_vline(mapping = aes(xintercept = 1),
>             colour = "blue") +
>  geom_vline(data = medians,
>             mapping = aes(xintercept = median),
>             colour = "red", linetype = "dashed") +
>  geom_label(data = medians,
>             aes(x = median, y = 0, label = median |> round(2)),
>             colour = "red") +
>  scale_x_continuous(breaks = seq(1, 21, 2)) +
>  facet_wrap(vars(type)) +
>  labs(title = "Rate of change in PDF size",
>       x = "Size change rate",
>       y = "Number of files",
>       caption = paste("Blue line is no change (rate of change = 1).
>       Red line is median change rate.
>                       Number of files in test sample:",
>                       nrow(wide)))
>
>ggsave("tagged_vs_UA_histograms.svg",
>       width = 10, height = 6, dpi = 600)
>
>#### Export data ####
>
># export to ODS
>library(readODS)
>write_ods(wide, "results.ods")
>

Actions: View

Attachments on bug 39667: 173642 | 173711 | 174124 | 174125 | 174126