import re
import warnings
from pathlib import Path

import pytest

from logwatcher.classifier import N2_PATTERNS, classify_log_entries
from logwatcher.config import FIXTURE_PATH
from logwatcher.parser import parse_file, parse_lines


def _pattern_name_to_filename(name: str) -> str:
    """Return the fixture corresponding to the error code name."""
    return f"CR_{name.lower()}.txt"


def _read_fixture(path: Path) -> str:
    """Read a file with fallback Windows-1252."""
    try:
        content = path.read_text(encoding="utf-8")
    except UnicodeDecodeError:
        content = path.read_text(encoding="windows-1252")
    if "\ufffd" not in content:
        return content
    return path.read_text(encoding="windows-1252")  # TODO: improve coverage score


@pytest.mark.parametrize("name, pattern", N2_PATTERNS.items())
def test_pattern_matches_fixture(valid_log_dir, name, pattern):
    """Assert each N2 pattern must match at least one line in its fixture."""
    fixture = valid_log_dir / _pattern_name_to_filename(name)
    if not fixture.exists():
        msg = f"No fixture named '{fixture}'"
        warnings.warn(msg)
        pytest.skip(msg)

    lignes = [l for l in _read_fixture(fixture).splitlines() if l.strip()]

    assert lignes, f"empty fixture file : {fixture}"

    match_count = 0

    for ligne in lignes:
        match = re.search(pattern, ligne)
        if match:
            match_count += 1

    assert match_count > 0, (
        f"{name} : no match in file {fixture.name}\npattern: {pattern}"
    )


def test_patterns_do_not_overlap():
    """Assert that a log line can match only one N2 pattern."""
    fixtures_dir = FIXTURE_PATH
    for fixture in fixtures_dir.glob("CR_*.txt"):
        line_list = [
            line for line in _read_fixture(fixture).splitlines() if line.strip()
        ]  # TODO: improve coverage score
        for line in line_list:  # TODO: improve coverage score
            matching = [  # TODO: improve coverage score
                name
                for name, raw in N2_PATTERNS.items()
                if re.compile(raw).search(line)
            ]
            assert len(matching) <= 1, (  # TODO: improve coverage score
                f"Chevauchement détecté sur {fixture.name} : "
                f"{matching} match line : {line[:80]}"
            )


def test_no_orphan_fixtures():
    """Assert that a fixture exist only if its associated pattern exists."""
    expected = {f"CR_{name.lower()}.txt" for name in N2_PATTERNS}
    actual = {f.name for f in FIXTURE_PATH.glob("CR_*.txt")}
    orphan = actual - expected
    assert not orphan, f"Fixtures orphelines (sans pattern) : {sorted(orphan)}"


def test_classify_log_entries_all_relevant(valid_log_dir: Path):
    """Must return an empty irrelevant log entry list."""
    log_file = valid_log_dir / "only_relevant_logs.txt"
    log_entries = parse_file(log_file)

    # all entries are relevant
    relevant, irrelevant = classify_log_entries(log_entries)

    assert relevant and not irrelevant


def test_classify_log_entries_none_relevant(valid_log_dir: Path):
    """Must return an empty relevant log entry list."""
    log_file = valid_log_dir / "no_relevant_logs.txt"
    log_entries = parse_file(log_file)

    # all entries are relevant
    relevant, irrelevant = classify_log_entries(log_entries)

    assert not relevant and irrelevant


def test_classify_log_entries_mixed(valid_log_dir: Path):
    """Test classification on relevant and irrelevant log entry list."""
    log_file = valid_log_dir / "mixed_logs.txt"
    log_entries = parse_file(log_file)

    relevant, irrelevant = classify_log_entries(log_entries)

    assert len(relevant) and len(irrelevant)
    assert len(relevant) + len(irrelevant) == len(log_entries)


def test_classify_log_entries_empty_log_entries(invalid_log_dir: Path):
    """Test classification on an empty log entry list."""
    empty_file = invalid_log_dir / "empty_file.txt"
    log_entries = parse_file(empty_file)

    relevant, irrelevant = classify_log_entries(log_entries)

    assert not relevant and not irrelevant


N2_ERROR_MESSAGE = "Erreur : GetTvaId -> TVA non trouvée : 20,00"
OTHER_ERROR_MESSAGE = "Erreur FTP SDUpdatePrixISF : Requested action not taken"


def _build_log_line(
    error_time: str,
    error_message: str,
    store_name: str = "GABRIEL",
    start_time: str = "22/09/2026 09:28:58",
    mdc_server_name: str = "MDC_220",
    server_ip: str = "192.168.13.22",
) -> str:
    """Build a raw log line as written by a MDC scan."""
    return (
        f"\\\\{server_ip}\\e\\{mdc_server_name}\\Logs\\26\\09\\22\\20260922092847.txt "
        f"[{start_time}] DOSSIER EN COURS : {store_name} "
        f"[{error_time}] {error_message}"
    )


def test_classify_log_entries_merges_consecutive_duplicates():
    """Consecutive identical errors become a single log entry."""
    log_entries = parse_lines(
        [
            _build_log_line("22/09/2026 09:32:19", N2_ERROR_MESSAGE),
            _build_log_line("22/09/2026 09:32:20", N2_ERROR_MESSAGE),
            _build_log_line("22/09/2026 09:32:25", N2_ERROR_MESSAGE),
        ]
    )

    relevant, irrelevant = classify_log_entries(log_entries)

    assert len(relevant) == 1
    assert relevant[0].get_error_time() == "22/09/2026 09:32:19"
    assert not irrelevant


def test_classify_log_entries_merges_same_error_name():
    """Same error code with a different amount is still a single log entry."""
    log_entries = parse_lines(
        [
            _build_log_line(
                "22/09/2026 09:32:19", "Erreur : GetTvaId -> TVA non trouvée : 8,50"
            ),
            _build_log_line(
                "22/09/2026 09:32:20", "Erreur : GetTvaId -> TVA non trouvée : 20,00"
            ),
        ]
    )

    relevant, _ = classify_log_entries(log_entries)

    assert len(relevant) == 1


def test_classify_log_entries_keeps_duplicates_from_other_store_or_scan():
    """Same error from another scan or another store is not a duplicate."""
    log_entries = parse_lines(
        [
            _build_log_line("22/09/2026 09:32:19", N2_ERROR_MESSAGE),
            _build_log_line(
                "22/09/2026 10:29:10",
                N2_ERROR_MESSAGE,
                start_time="22/09/2026 10:28:58",
            ),
            _build_log_line(
                "22/09/2026 09:32:19", N2_ERROR_MESSAGE, store_name="DUTOUR"
            ),
        ]
    )

    relevant, _ = classify_log_entries(log_entries)

    assert len(relevant) == 3


def test_classify_log_entries_merges_consecutive_irrelevant_duplicates():
    """Deduplication also applies to errors that do not require N2 support."""
    log_entries = parse_lines(
        [
            _build_log_line("22/09/2026 09:32:19", OTHER_ERROR_MESSAGE),
            _build_log_line("22/09/2026 09:32:24", OTHER_ERROR_MESSAGE),
        ]
    )

    relevant, irrelevant = classify_log_entries(log_entries)

    assert not relevant
    assert len(irrelevant) == 1


def test_classify_log_entries_keeps_same_error_separated_by_another_error():
    """An interleaved error breaks the run, so both errors are kept."""
    log_entries = parse_lines(
        [
            _build_log_line("22/09/2026 09:32:19", N2_ERROR_MESSAGE),
            _build_log_line("22/09/2026 09:32:20", OTHER_ERROR_MESSAGE),
            _build_log_line("22/09/2026 09:32:21", N2_ERROR_MESSAGE),
        ]
    )

    relevant, irrelevant = classify_log_entries(log_entries)

    assert len(relevant) == 2
    assert len(irrelevant) == 1