#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""validate_draft.py — would this draft pass the catalog's copy gate? Checked on a THROWAWAY copy, never the repo.

    python3 validate_draft.py <catalog-root> <draft.json>

Copies `templates/<app>/`, `scripts/` and the freeze into a temp directory, writes the guide there with
apply_guide.py, admits the app to the temp freeze, and runs check-copy-i18n.py --root <temp> <app>. Also checks the
night's format: 3-5 sections, each with a title, a one-sentence intro and 2-5 steps; 4-6 FAQ entries; no URL, no
markdown, no bare DOMAIN outside `<label>.DOMAIN`. Prints PASS or the findings; exit 0 / 1. The repo is only read.
"""
import io
import json
import os
import re
import shutil
import subprocess
import sys
import tempfile

HERE = os.path.dirname(os.path.abspath(__file__))


def fmt_problems(d):
    p = []
    for lang in ("hu", "en"):
        g = d.get(lang) or {}
        secs, faq = g.get("sections") or [], g.get("faq") or []
        if not 3 <= len(secs) <= 5:
            p.append("%s: %d sections (format: 3-5)" % (lang, len(secs)))
        if not 4 <= len(faq) <= 6:
            p.append("%s: %d FAQ entries (format: 4-6)" % (lang, len(faq)))
        for i, se in enumerate(secs):
            if not se.get("intro"):
                p.append("%s: sections[%d] has no intro (format: one sentence each)" % (lang, i))
            if not 2 <= len(se.get("steps") or []) <= 5:
                p.append("%s: sections[%d] has %d steps (format: 2-5)" % (lang, i, len(se.get("steps") or [])))
        strings = []
        for se in secs:
            strings += [se.get("title", ""), se.get("intro", "")] + list(se.get("steps") or [])
        for qa in faq:
            strings += [qa.get("q", ""), qa.get("a", "")]
        for s in strings:
            if re.search(r"https?://|www\.", s):
                p.append("%s: a URL in guide text (not allowed): %r" % (lang, s))
            if re.search(r"\*\*|__|`|<[a-zA-Z/][^>]*>|^#|\[[^\]]+\]\(", s):
                p.append("%s: markdown or HTML in guide text (rendered as plain text): %r" % (lang, s))
            for m in re.finditer(r"\bDOMAIN\b", s):
                if not re.search(r"[a-z0-9-]+\.DOMAIN\b", s[max(0, m.start() - 40):m.end()]):
                    p.append("%s: bare DOMAIN (the box fills it with the household's domain): %r" % (lang, s))
            if "—" in s and lang == "hu":
                p.append("hu: em dash (R-932): %r" % s)
            if re.search(r"[\U0001F300-\U0001FAFF☀-➿]", s):
                p.append("%s: emoji: %r" % (lang, s))
    return p


def main(argv):
    root, draft = os.path.abspath(argv[1]), os.path.abspath(argv[2])
    d = json.load(io.open(draft, encoding="utf-8"))
    app = d["app"]
    probs = fmt_problems(d)
    tmp = tempfile.mkdtemp(prefix="guide-validate-")
    try:
        shutil.copytree(os.path.join(root, "scripts"), os.path.join(tmp, "scripts"),
                        ignore=shutil.ignore_patterns("__pycache__"))
        os.makedirs(os.path.join(tmp, "templates"))
        for a in os.listdir(os.path.join(root, "templates")):
            src = os.path.join(root, "templates", a)
            if os.path.isdir(src):
                shutil.copytree(src, os.path.join(tmp, "templates", a))
        r = subprocess.run([sys.executable, os.path.join(HERE, "apply_guide.py"), tmp, draft], capture_output=True,
                           text=True)
        if r.returncode != 0:
            probs.append("apply: " + (r.stdout + r.stderr).strip())
        else:
            subprocess.run([sys.executable, os.path.join(tmp, "scripts", "check-copy-i18n.py"), "--root", tmp,
                            "--add-app", app, "--reason", "validate_draft"], capture_output=True, text=True)
            # The ratchet is about the whole catalog; measure it on the temp copy and hand it back.
            probe = subprocess.run([sys.executable, os.path.join(tmp, "scripts", "check-copy-i18n.py"), "--root", tmp,
                                    "--expect-missing", "-1"], capture_output=True, text=True)
            m = re.search(r"(\d+) strings have no English", probe.stdout)
            miss = m.group(1) if m else "0"
            r = subprocess.run([sys.executable, os.path.join(tmp, "scripts", "check-copy-i18n.py"), "--root", tmp,
                                "--expect-missing", miss, app], capture_output=True, text=True)
            if r.returncode != 0:
                probs += ["copy gate: " + l.strip() for l in r.stdout.splitlines() if l.strip().startswith("- ")
                          and app in l]
                if not any(app in l for l in r.stdout.splitlines() if l.strip().startswith("- ")):
                    probs.append("copy gate failed:\n" + r.stdout[-1500:])
            if miss != "0":
                probs.append("English coverage: %s copy string(s) of the catalog would have no English" % miss)
    finally:
        shutil.rmtree(tmp, ignore_errors=True)
    if probs:
        print("FAIL %s" % app)
        for x in probs:
            print("  - " + x)
        return 1
    print("PASS %s — format and copy gate clean" % app)
    return 0


if __name__ == "__main__":
    sys.exit(main(sys.argv))