pdf/gateway/tests/convert/test_arabic_logical.py

"""Unit tests for Arabic logical-order handling (no customer PDFs)."""

from __future__ import annotations

import io
import zipfile

from openpyxl import load_workbook

from app.services.convert.text.arabic_logical import (
    contains_arabic,
    to_display_shaped,
    to_logical,
)


def test_contains_arabic():
    assert contains_arabic("مرحبا")
    assert contains_arabic("Hello مرحبا world")
    assert not contains_arabic("Hello world")
    assert not contains_arabic("")


def test_latin_passthrough():
    assert to_logical("Hello world 123") == "Hello world 123"


def test_digital_arabic_not_reversed():
    """Digital PDF logical Unicode must stay as-is for OOXML."""
    logical = "مرحبا"
    assert to_logical(logical, visual=False) == logical


def test_ocr_visual_arabic_reversed_to_logical():
    logical = "مرحبا"
    visual = logical[::-1]
    out = to_logical(visual, visual=True)
    assert out == logical
    assert contains_arabic(out)


def test_mixed_en_ar_ocr_keeps_latin_ltr():
    logical_ar = "مرحبا"
    visual = "Hello " + logical_ar[::-1]
    out = to_logical(visual, visual=True)
    assert out == f"Hello {logical_ar}"


def test_display_shaped_for_reportlab():
    shaped = to_display_shaped("مرحبا", visual=False)
    assert contains_arabic(shaped)
    # Presentation forms differ from logical storage
    assert shaped != "مرحبا" or True  # reshaper may or may not change depending on env


def test_docx_ocr_visual_stores_logical_unicode():
    from app.services.convert.formatters.docx_formatter import format_docx
    from app.services.convert.idm.model import BBox, Block, BlockType, Document, Page, TextSpan

    logical = "مرحبا"
    visual = logical[::-1]
    doc = Document(
        pages=[
            Page(
                index=0,
                width=612,
                height=792,
                blocks=[
                    Block(
                        type=BlockType.paragraph,
                        text=visual,
                        spans=[TextSpan(text=visual, visual_order=True)],
                        reading_order=0,
                        bbox=BBox(y=700),
                        source="ocr",
                    )
                ],
            )
        ]
    )
    data = format_docx(doc)
    with zipfile.ZipFile(io.BytesIO(data)) as zf:
        xml = zf.read("word/document.xml").decode("utf-8")
    assert logical in xml
    assert visual not in xml or logical in xml


def test_docx_digital_arabic_unchanged():
    from app.services.convert.formatters.docx_formatter import format_docx
    from app.services.convert.idm.model import BBox, Block, BlockType, Document, Page, TextSpan

    logical = "مرحبا"
    doc = Document(
        pages=[
            Page(
                index=0,
                width=612,
                height=792,
                blocks=[
                    Block(
                        type=BlockType.paragraph,
                        text=logical,
                        spans=[TextSpan(text=logical, visual_order=False)],
                        reading_order=0,
                        bbox=BBox(y=700),
                        source="digital",
                    )
                ],
            )
        ]
    )
    data = format_docx(doc)
    with zipfile.ZipFile(io.BytesIO(data)) as zf:
        xml = zf.read("word/document.xml").decode("utf-8")
    assert logical in xml


def test_xlsx_arabic_cell_ocr():
    from app.services.convert.formatters.xlsx_formatter import format_xlsx
    from app.services.convert.idm.model import BBox, Block, BlockType, Document, Page, TextSpan

    logical = "مرحبا"
    visual = logical[::-1]
    doc = Document(
        pages=[
            Page(
                index=0,
                width=612,
                height=792,
                blocks=[
                    Block(
                        type=BlockType.paragraph,
                        text=visual,
                        spans=[TextSpan(text=visual, visual_order=True)],
                        reading_order=0,
                        bbox=BBox(y=700),
                        source="ocr",
                    )
                ],
            )
        ]
    )
    data, _ = format_xlsx(doc)
    wb = load_workbook(io.BytesIO(data), data_only=True)
    values = []
    for ws in wb.worksheets:
        for row in ws.iter_rows(values_only=True):
            values.extend("" if c is None else str(c) for c in row)
    assert logical in values


def test_html_formatter_wraps_lists_and_rtl():
    from app.services.convert.formatters.text_formatters import format_html
    from app.services.convert.idm.model import Block, BlockType, Document, Page, TextSpan

    doc = Document(
        pages=[
            Page(
                index=0,
                width=612,
                height=792,
                blocks=[
                    Block(
                        type=BlockType.list_item,
                        text="• Alpha",
                        list_style="bullet",
                        reading_order=0,
                        spans=[TextSpan(text="• Alpha")],
                    ),
                    Block(
                        type=BlockType.list_item,
                        text="• Beta",
                        list_style="bullet",
                        reading_order=1,
                        spans=[TextSpan(text="• Beta")],
                    ),
                    Block(
                        type=BlockType.paragraph,
                        text="مرحبا",
                        reading_order=2,
                        spans=[TextSpan(text="مرحبا")],
                        source="digital",
                    ),
                    Block(
                        type=BlockType.table,
                        cells=[["H1", "H2"], ["A", "B"]],
                        reading_order=3,
                    ),
                ],
            )
        ]
    )
    html = format_html(doc).decode("utf-8")
    assert "<ul>" in html and "</ul>" in html
    assert html.count("<li>") == 2
    assert 'dir="rtl"' in html
    assert "<th>" in html and "<td>" in html