wadamesh/scripts/build/gen-lang-builtin.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""Bake deploy/apps/lang/*.lang into a compiled-in fallback table.

Boards that can reach the Lua Store download their language as a file. Boards
that cannot (the V4: ~14 KB of free internal heap, so the net worker and the
store are unreliable there) still need translations, and before this they were
stuck on English.

This generates src/ui-touch/i18n_builtin.h from the SAME .lang files the store
serves, so the two can never drift: translators edit deploy/apps/lang/*.lang,
and both paths follow. Rows are emitted sorted by key so TR() can binary-search
them, and the table lives in .rodata — flash, not RAM.

Run from the repo root (a PlatformIO pre-build step does this automatically):
    python3 scripts/build/gen-lang-builtin.py
"""
import glob, os, sys

ROOT = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
LANGDIR = os.path.join(ROOT, 'deploy/apps/lang')
OUT = os.path.join(ROOT, 'src/ui-touch/i18n_builtin.h')

# Index order must match the UiLang enum in i18n.h.
CODES = ["en", "hu", "nl", "de", "fr", "es", "it",
         "ru", "uk", "bg", "sr", "el", "pt-br", "ro"]


def c_escape(s):
    out = []
    for ch in s:
        if ch == '\\':   out.append('\\\\')
        elif ch == '"':  out.append('\\"')
        elif ch == '\n': out.append('\\n')
        elif ch == '\t': out.append('\\t')
        elif ch == '\r': out.append('\\r')
        elif ord(ch) < 0x20: out.append('\\x%02x' % ord(ch))
        else: out.append(ch)
    return ''.join(out)


def read_lang(path):
    """key -> value, with the file's own \\n/\\t/\\\\ escapes decoded."""
    rows = {}
    for line in open(path, encoding='utf-8'):
        line = line.rstrip('\n').rstrip('\r')
        if '\t' not in line:          # header/comment lines carry no tab
            continue
        k, v = line.split('\t', 1)
        def un(t):
            o, i = [], 0
            while i < len(t):
                if t[i] == '\\' and i + 1 < len(t):
                    n = t[i + 1]
                    o.append('\n' if n == 'n' else '\t' if n == 't' else n)
                    i += 2
                else:
                    o.append(t[i]); i += 1
            return ''.join(o)
        k, v = un(k), un(v)
        if k and v:
            rows[k] = v
    return rows


def main():
    langs = {}
    for code in CODES[1:]:                      # English is the key itself
        p = os.path.join(LANGDIR, code + '.lang')
        if os.path.exists(p):
            langs[code] = read_lang(p)

    if not langs:
        sys.exit('no .lang files found in ' + LANGDIR)

    w = ['// Generated by scripts/build/gen-lang-builtin.py — DO NOT EDIT.',
         '// Source: deploy/apps/lang/*.lang (the same files the Lua Store serves).',
         '// Compiled-in fallback for boards that cannot rely on downloading a language.',
         '#pragma once', '#include "i18n.h"', '']

    for code in CODES[1:]:
        rows = langs.get(code, {})
        sym = 'kBuiltin_' + code.replace('-', '_')
        w.append('static const I18nPair %s[] = {' % sym)
        # A row whose translation IS the key does nothing: TR() already returns the
        # key when it finds no match. The .lang files carry those deliberately, as
        # placeholders telling translators the string exists, but baking them costs
        # flash on exactly the board that needs the flash (the V4 sits at ~90%).
        for k in sorted(rows):                  # sorted: TR() binary-searches
            if rows[k] == k:
                continue
            w.append('  { "%s", "%s" },' % (c_escape(k), c_escape(rows[k])))
        w.append('};')

    w.append('')
    w.append('static const I18nPair* const kBuiltinLang[LANG_COUNT] = {')
    w.append('  nullptr,')                       # LANG_EN: the key is the English
    for code in CODES[1:]:
        sym = 'kBuiltin_' + code.replace('-', '_')
        w.append('  %s,' % (sym if langs.get(code) else 'nullptr'))
    w.append('};')
    w.append('static const int kBuiltinLangCount[LANG_COUNT] = {')
    w.append('  0,')
    for code in CODES[1:]:
        sym = 'kBuiltin_' + code.replace('-', '_')
        w.append('  %s,' % (('(int)(sizeof(%s)/sizeof(%s[0]))' % (sym, sym))
                            if langs.get(code) else '0'))
    w.append('};')
    w.append('')

    open(OUT, 'w', encoding='utf-8').write('\n'.join(w))
    total = sum(sum(1 for k, v in r.items() if v != k) for r in langs.values())
    print('%s: %d languages, %d translated rows, %d bytes' %
          (os.path.relpath(OUT, ROOT), len(langs), total, os.path.getsize(OUT)))


if __name__ == '__main__':
    main()