feat: support for bytes and num inline formatting

NEWS.md

@@ -3,6 +3,9 @@
 * `code_highlight()` supports long strings and symbols
   (#727 @moodymudskipper).
 
+* New `{.num}` and `{.bytes}` inline styles to format numbers
+  and bytes (@m-muecke, #644, #588, #643).
+
 # cli 3.6.4
 
 * Pluralization now handles edge cases (`NA`, `NaN`, `Inf` and `-Inf`)

R/numbers.R

@@ -0,0 +1,128 @@
+
+format_num <- local({
+
+  pretty_num <- function(number, style = c("default", "nopad", "6")) {
+
+    style <- switch(
+      match.arg(style),
+      "default" = pretty_num_default,
+      "nopad" = pretty_num_nopad,
+      "6" = pretty_num_6
+    )
+
+    style(number)
+  }
+
+  compute_num <- function(number, smallest_prefix = "y") {
+    prefixes0 <- c("y","z","a","f","p","n","u","m","", "k", "M", "G", "T", "P", "E", "Z", "Y")
+    zeroshif0 <- 9L
+
+    stopifnot(
+      is.numeric(number),
+      is.character(smallest_prefix),
+      length(smallest_prefix) == 1,
+      !is.na(smallest_prefix),
+      smallest_prefix %in% prefixes0
+    )
+
+    limits <- c(999950 * 1000 ^ (seq_len(length(prefixes0)) - (zeroshif0 + 1L)))
+    nrow <- length(limits)
+    low <- match(smallest_prefix, prefixes0)
+    zeroshift <- zeroshif0 + 1L - low
+    prefixes <- prefixes0[low:length(prefixes0)]
+    limits <- limits[low:nrow]
+    nrow <- nrow - low + 1
+
+    neg <- number < 0 & !is.na(number)
+    number <- abs(number)
+    mat <- matrix(
+      rep(number, each = nrow),
+      nrow = nrow,
+      ncol = length(number)
+    )
+    mat2 <- matrix(mat < limits, nrow  = nrow, ncol = length(number))
+    exponent <- nrow - colSums(mat2) - (zeroshift - 1L)
+    in_range <- function(exponent) {
+      max(min(exponent, nrow - zeroshift, na.rm = FALSE), 1L - zeroshift, na.rm = TRUE)
+    }
+    if (length(exponent)) {
+      exponent <- sapply(exponent, in_range)
+    }
+    res <- number / 1000 ^ exponent
+    prefix <- prefixes[exponent + zeroshift]
+
+    ## Zero number
+    res[number == 0] <- 0
+    prefix[number == 0] <- prefixes[zeroshift]
+
+    ## NA and NaN number
+    res[is.na(number)] <- NA_real_
+    res[is.nan(number)] <- NaN
+    prefix[is.na(number)] <- "" # prefixes0[low] is meaningless    # Includes NaN as well
+
+    data.frame(
+      stringsAsFactors = FALSE,
+      amount = res,
+      prefix = prefix,
+      negative = neg
+    )
+  }
+
+  pretty_num_default <- function(number) {
+    szs <- compute_num(number)
+    amt <- szs$amount
+    sep <- " "
+
+    ## String. For fractions we always show two fraction digits
+    res <- character(length(amt))
+    int <- is.na(amt) | abs(amt - as.integer(amt)) <= .Machine$double.eps
+    res[int] <- format(
+      ifelse(szs$negative[int], -1, 1) * amt[int],
+      scientific = FALSE
+    )
+    res[!int] <- sprintf("%.2f", ifelse(szs$negative[!int], -1, 1) * amt[!int])
+
+    format(paste(res, szs$prefix, sep = sep), justify = "right")
+  }
+
+  pretty_num_nopad <- function(number) {
+    sub("^\\s+", "", pretty_num_default(number))
+  }
+
+  pretty_num_6 <- function(number) {
+    szs <- compute_num(number, smallest_prefix = "y")
+    amt <- round(szs$amount, 2)
+    sep <- " "
+
+    na   <- is.na(amt)
+    nan  <- is.nan(amt)
+    neg  <- !na & !nan & szs$negative
+    l10p  <- !na & !nan & !neg & amt < 10
+    l100p <- !na & !nan & !neg & amt >= 10 & amt < 100
+    b100p <- !na & !nan & !neg & amt >= 100
+    l10n  <- !na & !nan & neg & amt < 10
+    l100n <- !na & !nan & neg & amt >= 10 & amt < 100
+    b100n <- !na & !nan & neg & amt >= 100
+
+    famt <- character(length(amt))
+    famt[na] <- "  NA"
+    famt[nan] <- " NaN"
+    famt[l10p] <- sprintf("%.2f", amt[l10p])
+    famt[l100p] <- sprintf("%.1f", amt[l100p])
+    famt[b100p] <- sprintf(" %.0f", amt[b100p])
+    famt[l10n] <- sprintf("-%.1f", amt[l10n])
+    famt[l100n] <- sprintf(" -%.0f", amt[l100n])
+    famt[b100n] <- sprintf("-%.0f", amt[b100n])
+
+    sub(" $", "  ", paste0(famt, sep, szs$prefix))
+  }
+
+  structure(
+    list(
+      .internal     = environment(),
+      pretty_num  = pretty_num,
+      compute_num = compute_num
+    ),
+    class = c("standalone_num", "standalone")
+  )
+})

R/themes.R

@@ -252,7 +252,13 @@ builtin_theme <- function(dark = getOption("cli.theme_dark", "auto")) {
       transform = function(x) format_inline(typename(x))
     ),
     span.or = list("vec-sep2" = " or ", "vec-last" = ", or "),
-    span.timestamp = list(before = "[", after = "]", color = "grey")
+    span.timestamp = list(before = "[", after = "]", color = "grey"),
+    span.bytes = list(
+      transform = function(x) format_bytes$pretty_bytes(as.numeric(x))
+    ),
+    span.num = list(
+      transform = function(x) format_num$pretty_num(as.numeric(x))
+    )
   )
 }
 

tests/testthat/_snaps/inline-2.md

@@ -436,3 +436,25 @@
     Output
       [1] "\nfoo\n\nbar\n"
 
+# .bytes
+
+    Code
+      format_inline("--- {.bytes 123123123} ---")
+    Output
+      [1] "--- 123.12 MB ---"
+    Code
+      format_inline("--- {.bytes {1:4 * 10000}} ---")
+    Output
+      [1] "--- 10 kB, 20 kB, 30 kB, and 40 kB ---"
+
+# .num
+
+    Code
+      format_inline("--- {.num 123123123} ---")
+    Output
+      [1] "--- 123.12 M ---"
+    Code
+      format_inline("--- {.num {1:4 * 10000}} ---")
+    Output
+      [1] "--- 10 k, 20 k, 30 k, and 40 k ---"
+

tests/testthat/test-inline-2.R

@@ -197,3 +197,17 @@ test_that("format_inline and newlines", {
     format_inline("\ffoo\f\fbar\f", keep_whitespace = FALSE)
   })
 })
+
+test_that(".bytes", {
+  expect_snapshot({
+    format_inline("--- {.bytes 123123123} ---")
+    format_inline("--- {.bytes {1:4 * 10000}} ---")
+  })
+})
+
+test_that(".num", {
+  expect_snapshot({
+    format_inline("--- {.num 123123123} ---")
+    format_inline("--- {.num {1:4 * 10000}} ---")
+  })
+})