require! fs

lang = process.argv.filter((in <[ a t h c ]>)).0

unless lang in <[ a t h c ]>

console.log "Please invoke this program with a single-letter argument, one of <[ a t h c ]>."

process.exit!

fs.mkdir-sync lang unless fs.exists-sync lang

PUA2UNI = {

\⿰𧾷百 : \𬦀

\⿸疒哥 : \󿗧

\⿰亻恩 : \𫣆

\⿰虫念 : \𬠖

\⿺皮卜 : \󿕅

}

grok = -> JSON.parse fs.read-file-sync(it, \utf8).replace(/[⿰⿸⿺](?:𧾷|.)./g -> PUA2UNI[it])

dump = (file, data) ->

console.log "Writing: #file"

fs.write-file-sync file, JSON.stringify data

entries = switch lang

| \t => grok(\dict-twblg.json) ++ grok(\dict-twblg-ext.json)

| \a => grok(\dict-revised.pua.json)

| \h => grok(\dict-hakka.json)

| \c => grok(\dict-csld.json)

prefix = {}

defs = {}

buckets = {}

i = 0

for {title, heteronyms}:entry in entries

continue if title is /\{\[[0-9a-f]{4}\]\}/ # Unsubstituted

continue if title is /\uDB40[\uDD00-\uDD0F]/ # Variant

continue if title is /[⿰⿸⿺]/

pre = title.slice(0, 1)

code = pre.charCodeAt(0)

if 0xD800 <= code <= 0xDBFF

pre = title.slice(0, 2)

code = pre.charCodeAt(1) - 0xDC00

post = title.slice(2)

else

post = title.slice(1)

prefix[pre] ?= ''

prefix[pre] += "|#post" if post.length

# throw "Impossible: #title" if defs[title]

defs[title] = entry

fs.write-file-sync \prefix.json JSON.stringify prefix

codepoints-of = -> it.length - it.split( /[\uD800-\uDBFF][\uDC00-\uDFFF]/g ).length + 1

trie = prefix

abbrevToTitle = {}

lenToTitles = {}

lenToRegex = {}

lens = []

for k, v of trie

prefix-length = codepoints-of k

for suffix in v / '|'

abbrevIndex = suffix.indexOf '('

if abbrevIndex >= 0

orig = suffix

suffix.=slice(0, abbrevIndex)

abbrevToTitle["#k#suffix"] = "#k#orig"

(lenToTitles[prefix-length + suffix.length] ?= []).push "#k#suffix"

for len, titles of lenToTitles

lens.push len

titles.sort!

lenToRegex[len] = (titles * \|).replace(/[-[\]{}()*+?.,\\#\s]/g, "\\$&")

fs.write-file-sync "lenToRegex.#len.json" JSON.stringify {"#len": (titles * \|).replace(/[-[\]{}()*+?.,\\#\s]/g, "\\$&")}

lens.sort (a, b) -> b - a

for len in [2 3 4] | titles = lenToTitles[len]

cur = ''

re = ''

for t in titles

one = t.slice(0, 1)

two = t.slice(1)

code = one.charCodeAt(0)

if 0xD800 <= code <= 0xDBFF

one = t.slice(0, 2)

two = t.slice(2)

if one is cur

re += "|" if len isnt 2

re += two

else

re += "]|#one[#two" if len is 2

re += ")|#one(#two" if len isnt 2

cur = one

re = re.replace(/\[(.|[\uD800-\uDBFF].)\]/g '$1') if len is 2

re = re.replace(/\(([^|]+)\)/g '$1') if len isnt 2

re = re.slice(2).replace(/[-{}*+?.,\\#\s]/g, "\\$&")

re += "]" if len is 2

re += ")" if len isnt 2

dump "#lang/lenToRegex.#len.json" {"#len": re}

lenToRegex[len] = re

dump "#lang/precomputed.json" { abbrevToTitle }

dump "#lang/lenToRegex.json" { lenToRegex }