Data update

This commit is contained in:
Ingy döt Net 2026-04-30 12:34:36 -04:00
parent 4bb20c9b71
commit cbaf4c4b64
12390 changed files with 318560 additions and 27248 deletions

View file

@ -0,0 +1,53 @@
local fmt = require "fmt"
$define DIGITS = "22233344455566677778889999"
local word_list = "unixdict.txt"
local map = {}
local count_valid = 0
local nl = os.platform == "windows" ? "\r\n" : "\n"
local words = io.contents(word_list):rstrip():split(nl)
for words as word do
local valid = true
local sb = ""
local lword = word:lower()
for i = 1, #lword do
local c = lword[i]
if not("a" <= c <= "z") then
valid = false
break
end
sb ..= DIGITS[c:byte() - 96]
end
if valid then
count_valid += 1
if map[sb] then
map[sb]:insert(word)
else
map[sb] = {word}
end
end
end
local t = map:filtered(|v| -> #v > 1)
local textonyms = t:keys():map(|k| -> {k, t[k]})
local report = $"There are {count_valid} words in '{word_list}' " ..
"which can be represented by the digit key mapping.\n" ..
$"They require {#map:keys()} digit combinations to represent them.\n" ..
$"{#textonyms} digit combinations represent Textonyms.\n"
print(report)
local longest = textonyms:sorted(|i, j| -> #j[1] < #i[1])
local ambiguous = longest:sorted(|i, j| -> #j[2] < #i[2])
print("Top 8 in ambiguity: \n")
print("Count Textonym Words")
print("====== ======== =====")
local f = "%4d %-8s %s"
for ambiguous:slice(1, 8) as a do fmt.print(f, #a[2], a[1], a[2]:concat(" ")) end
f = f:replace("8", "14")
print("\nTop 6 in length:\n")
print("Length Textonym Words")
print("====== ============== =====")
for longest:slice(1, 6) as l do fmt.print(f, #l[1], l[1], l[2]:concat(" ")) end