import re
from pathlib import Path

import openpyxl
import pytest

from synthgen import dictionary, inn
from synthgen.dictionary import build

REAL_ECP = Path("/Users/andrey.epifancev/projects/awp-ca/fixtures/directories/ecp-export-russkiy-biskvit.xlsx")
REAL_OSV_DIR = Path("/Users/andrey.epifancev/projects/awp-ca/fixtures/osv/rb/2024-Q1")
_FORMS = re.compile(r"\b(ООО|АО|ПАО|ЗАО|ОАО|ИП)\b", re.IGNORECASE)


def _norm(text: str) -> str:
    text = _FORMS.sub(" ", text)
    return re.sub(r"[^0-9a-zа-яё]+", " ", text.lower().replace("ё", "е")).strip()


def test_dictionary_size_and_kinds():
    orgs = dictionary.load()
    assert len(orgs) == 400
    assert sum(o.kind == "банк" for o in orgs) == 30
    assert 30 <= sum(o.form == "ИП" for o in orgs) <= 50
    names = [o.name.lower() for o in orgs]
    assert len(set(names)) == len(names)


def test_dictionary_inn():
    for o in dictionary.load():
        assert inn.is_valid(o.inn)
        if o.form == "ИП":
            assert len(o.inn) == 12 and o.kpp == ""
            number = int(o.inn[2:-2])
        else:
            assert len(o.inn) == 10
            assert re.fullmatch(r"\d{9}", o.kpp)
            number = int(o.inn[2:-1])
        assert 100001 <= number <= 109999


def _real_names() -> set[str]:
    names: set[str] = set()
    wb = openpyxl.load_workbook(REAL_ECP, read_only=True)
    for ws in wb.worksheets:
        rows = ws.iter_rows(values_only=True)
        header = next(rows, None)
        if not header or "Организация" not in header:
            continue
        col = list(header).index("Организация")
        names.update(_norm(str(r[col])) for r in rows if r[col])
    for path in sorted(REAL_OSV_DIR.glob("*.xlsx")):
        wb = openpyxl.load_workbook(path, read_only=True)
        ws = wb.worksheets[0]
        names.update(_norm(str(r[0])) for r in ws.iter_rows(max_col=1, values_only=True) if r and r[0])
    names.discard("")
    return names


def test_dictionary_not_real():
    if not REAL_ECP.exists() or not REAL_OSV_DIR.is_dir():
        pytest.skip("нет реальных фикстур")
    real = _real_names()
    assert len(real) > 50
    clash = [o.name for o in dictionary.load() if _norm(o.name) in real]
    assert clash == []


def test_build_is_reproducible(tmp_path):
    out = tmp_path / "companies.csv"
    build.write_csv(out)
    assert out.read_bytes() == dictionary.CSV_PATH.read_bytes()