"""Shared vocabulary regressions; no generated engine facts in these unit inputs."""

from pathlib import Path

import pytest

from scripts.reader_appendix_language import clean_reader_appendix_markdown
from scripts import skill_release_package


ROOT = Path(__file__).resolve().parents[1]
RULE_PATH = "scripts/reader_appendix_language.rules.json"


@pytest.mark.parametrize("source,expected", [
    ("PL9 第 42 页", "外部参照资料"),
    ("PL9第42页", "外部参照资料"),
    ("PL9.pdf 第 42 页", "外部参照资料"),
    ("PL9.pdf第42页后文", "外部参照资料后文"),
    ("参见PL9第43–44页中的字段", "参见外部参照资料中的字段"),
    ("PL9 第 40 / 48 页", "外部参照资料"),
    ("PL9 pages 99-100", "外部参照资料"),
])
def test_chinese_page_boundary(source, expected):
    assert clean_reader_appendix_markdown(source) == expected


@pytest.mark.parametrize("edition", ["basic_git", "premium_cloud_drive"])
def test_shared_rules_are_included_by_skill_package_selection(monkeypatch, edition):
    # Simulate the tracked file inventory without mutating this worktree's index.
    # Exercise the real release selector: scripts assets must not become .py-only.
    inventory = ["scripts/reader_appendix_language.py", RULE_PATH]
    # Release manifest references are relative to the repository, not the caller.
    monkeypatch.chdir(ROOT)
    monkeypatch.setattr(skill_release_package, "_git_files", lambda: inventory)
    selected = skill_release_package._edition_files(edition)
    assert all(path in selected for path in inventory)
    assert (ROOT / RULE_PATH).is_file()


@pytest.mark.parametrize("dockerfile", ["railway-api.Dockerfile", "railway-web.Dockerfile"])
def test_shared_rules_live_inside_existing_image_copy_boundary(dockerfile):
    source = (ROOT / "deploy" / dockerfile).read_text(encoding="utf-8")
    assert any(line.startswith("COPY scripts ") for line in source.splitlines())


def test_crlf_chart_fence_is_byte_preserved():
    fence = '  ```jyotish-chart\r\n{"title":"PL9第42页 parameter_sensitive","id":"D1"}\r\n  ```\r\n'
    source = "PL9第42页\r\n" + fence + "| parameter_sensitive |\r\n"
    assert clean_reader_appendix_markdown(source).encode("utf-8") == (
        "外部参照资料\r\n" + fence + "| 参数敏感 |\r\n"
    ).encode("utf-8")