defmodule VavPipeline.Tablet39ParashahBreaksTest do
  @moduledoc """
  Tablet 39 (2026-10-04) — L1 had no paragraphs.

  The WLC marks every petuchah (פ) and setumah (ס). `scripts/build_l1.py`
  read past them, so the scroll in the aron was 5,853 verses with no
  division between them but chapter numbers, which are not Torah's.
  They are now engraved per verse as `parashah_breaks`, and the reader
  carries them.
  """

  use ExUnit.Case, async: true

  alias VavPipeline.L1Reader

  defp breaks(book, chapter, verse) do
    {:ok, v} = L1Reader.read_verse(book, chapter, verse)
    {v.parashah_breaks, length(v.words)}
  end

  test "the six days of creation each close with a petuchah" do
    for verse <- [5, 8, 13, 19, 23, 31] do
      {brks, word_count} = breaks("bereishit", 1, verse)
      assert brks == [%{after_word: word_count, type: :petuchah}], "Bereishit 1:#{verse}"
    end
  end

  test "a verse with no break after it carries an empty list, not a missing field" do
    assert {[], _} = breaks("bereishit", 1, 1)
    assert {[], _} = breaks("bereishit", 1, 6)
  end

  test "the Shema opens after a petuchah and its first paragraph closes with a setumah" do
    # Devarim 6:3 ends the preceding section; 6:4–9 is one parashah.
    {before, n} = breaks("devarim", 6, 3)
    assert before == [%{after_word: n, type: :petuchah}]

    for verse <- 4..8, do: assert({[], _} = breaks("devarim", 6, verse))

    {closing, m} = breaks("devarim", 6, 9)
    assert closing == [%{after_word: m, type: :setumah}]
  end

  test "pisqa b'emtza pasuq — a break inside the verse is placed by the word it follows" do
    # Bereishit 35:22: "…vayishma Yisrael" [פ] "vayihyu v'nei Ya'akov…"
    {[brk], word_count} = breaks("bereishit", 35, 22)
    assert brk == %{after_word: 14, type: :petuchah}
    assert brk.after_word < word_count

    {:ok, v} = L1Reader.read_verse("bereishit", 35, 22)
    assert Enum.at(v.words, 13).consonants == "ישראל"
    assert Enum.at(v.words, 14).consonants == "ויהיו"
  end

  test "Devarim 5:21 has two setumot — one within the verse, one after it" do
    {brks, word_count} = breaks("devarim", 5, 21)

    assert brks == [
             %{after_word: 4, type: :setumah},
             %{after_word: word_count, type: :setumah}
           ]
  end

  test "the text of a verse is unchanged — the break is structure, not letters" do
    {:ok, v} = L1Reader.read_verse("bereishit", 1, 5)
    assert String.ends_with?(v.text_hebrew, "׃")
    refute String.ends_with?(String.trim_trailing(v.text_hebrew, "׃") |> String.trim(), "פ")
  end

  test "the whole scroll: 293 petuchot and 394 setumot, in 686 verses" do
    index = L1Reader.l1_root() |> Path.join("torah/index.json") |> File.read!() |> Jason.decode!()

    counts =
      for book <- index["books"],
          {chapter, verses} <- book["chapter_verse_counts"],
          verse <- 1..verses,
          reduce: %{petuchah: 0, setumah: 0, verses: 0} do
        acc ->
          {:ok, v} = L1Reader.read_verse(book["book"], String.to_integer(chapter), verse)

          Enum.reduce(v.parashah_breaks, acc, fn b, acc -> Map.update!(acc, b.type, &(&1 + 1)) end)
          |> Map.update!(:verses, &(&1 + if(v.parashah_breaks == [], do: 0, else: 1)))
      end

    assert counts == %{petuchah: 293, setumah: 394, verses: 686}
  end
end