#!/usr/bin/env python3
"""Translation helpers for the generated content pages.

  python3 -E scripts/i18n.py extract <out_dir>
      Write one flat {path: text} JSON per English page with only the strings to translate.

  python3 -E scripts/i18n.py apply <locale> <dir>
      Read translated flat files from <dir> and write src/content/pages/<locale>/<slug>.json,
      keeping the English structure. Fails if a translation is missing or extra.

  python3 -E scripts/i18n.py index
      Regenerate src/content/pages/index.ts. A locale without a page falls back to English.
"""

from __future__ import annotations

import copy
import json
import pathlib
import re
import sys

ROOT = pathlib.Path(__file__).resolve().parent.parent
PAGES = ROOT / "src" / "content" / "pages"
LOCALES = ["en", "es", "ar"]
SKIP_KEYS = {"slug", "path", "form", "level", "kind", "num"}


def keep_english(value: str, section: dict | None, key: str) -> bool:
    """Office addresses and stat figures stay as they are in every language."""
    if "Phone:" in value:
        return True
    if section and section.get("kind") == "stats" and key == "title" and re.match(r"^[\d.,]+", value):
        return True
    return False


def walk(node, path=(), section=None):
    if isinstance(node, dict):
        if "kind" in node:
            section = node
        for k, v in node.items():
            if k in SKIP_KEYS:
                continue
            yield from walk(v, path + (k,), section)
    elif isinstance(node, list):
        for i, v in enumerate(node):
            yield from walk(v, path + (str(i),), section)
    elif isinstance(node, str) and node.strip():
        if not keep_english(node, section, path[-1]):
            yield ".".join(path), node


def set_path(node, path: str, value: str):
    parts = path.split(".")
    for p in parts[:-1]:
        node = node[int(p)] if isinstance(node, list) else node[p]
    last = parts[-1]
    if isinstance(node, list):
        node[int(last)] = value
    else:
        node[last] = value


def english_pages():
    return sorted((PAGES / "en").glob("*.json"))


def extract(out_dir: str):
    out = pathlib.Path(out_dir)
    out.mkdir(parents=True, exist_ok=True)
    for f in english_pages():
        flat = dict(walk(json.loads(f.read_text())))
        (out / f.name).write_text(json.dumps(flat, ensure_ascii=False, indent=1) + "\n")
        print(f"{f.name}: {len(flat)} strings")


def apply(locale: str, src_dir: str):
    src = pathlib.Path(src_dir)
    dest = PAGES / locale
    dest.mkdir(parents=True, exist_ok=True)
    problems = 0
    for f in english_pages():
        tf = src / f.name
        if not tf.exists():
            print(f"skip {f.name}: no translation yet")
            continue
        page = json.loads(f.read_text())
        wanted = dict(walk(page))
        got = json.loads(tf.read_text())
        missing = [k for k in wanted if k not in got]
        extra = [k for k in got if k not in wanted]
        if missing or extra:
            problems += 1
            print(f"{f.name}: missing {missing[:5]} extra {extra[:5]}")
            continue
        out = copy.deepcopy(page)
        for k, v in got.items():
            set_path(out, k, v)
        (dest / f.name).write_text(json.dumps(out, ensure_ascii=False, indent=2) + "\n")
        print(f"{locale}/{f.name}: ok")
    if problems:
        sys.exit(1)


def ident(locale: str, name: str) -> str:
    return locale + "_" + re.sub(r"\W", "_", name)


def index():
    names = [f.stem for f in english_pages()]
    lines = ["// Generated by scripts/i18n.py index — do not edit by hand.", ""]
    table = []
    for locale in LOCALES:
        entries = []
        for n in names:
            if (PAGES / locale / f"{n}.json").exists():
                lines.append(f'import {ident(locale, n)} from "./{locale}/{n}.json";')
                entries.append(f'    "{n}": {ident(locale, n)},')
            else:
                entries.append(f'    "{n}": {ident("en", n)},')
        table.append(f"  {locale}: {{\n" + "\n".join(entries) + "\n  },")
    lines += ["", "export const content = {", *table, "};", ""]
    (PAGES / "index.ts").write_text("\n".join(lines))
    print(f"index.ts: {len(names)} pages × {len(LOCALES)} locales")


if __name__ == "__main__":
    cmd, *args = sys.argv[1:]
    {"extract": extract, "apply": apply, "index": index}[cmd](*args)