run.font.name = name

run._element.get_or_add_rPr().rFonts.set(qn("w:ascii"), name)

run._element.get_or_add_rPr().rFonts.set(qn("w:hAnsi"), name)

run.font.size = Pt(size)

run.font.color.rgb = RGBColor.from_string(color)

document = Document()

section = document.sections[0]

section.page_width = Inches(8.5)

section.page_height = Inches(11)

section.top_margin = Inches(1)

section.right_margin = Inches(1)

section.bottom_margin = Inches(1)

section.left_margin = Inches(1)

section.header_distance = Inches(0.492)

section.footer_distance = Inches(0.492)

normal = document.styles["Normal"]

normal.font.name = "Calibri"

normal._element.rPr.rFonts.set(qn("w:ascii"), "Calibri")

normal._element.rPr.rFonts.set(qn("w:hAnsi"), "Calibri")

normal.font.size = Pt(11)

normal.paragraph_format.space_before = Pt(0)

normal.paragraph_format.space_after = Pt(6)

normal.paragraph_format.line_spacing = 1.25

heading_tokens = {

"Heading 1": (16, "2E74B5", 18, 10),

"Heading 2": (13, "2E74B5", 14, 7),

"Heading 3": (12, "1F4D78", 10, 5),

}

for style_name, (size, color, before, after) in heading_tokens.items():

style = document.styles[style_name]

style.font.name = "Calibri"

style._element.rPr.rFonts.set(qn("w:ascii"), "Calibri")

style._element.rPr.rFonts.set(qn("w:hAnsi"), "Calibri")

style.font.size = Pt(size)

style.font.color.rgb = RGBColor.from_string(color)

style.paragraph_format.space_before = Pt(before)

style.paragraph_format.space_after = Pt(after)

header = section.header.paragraphs[0]

header.alignment = WD_ALIGN_PARAGRAPH.LEFT

header.paragraph_format.space_after = Pt(0)

set_run_font(header.add_run("FLUENTREAD · DOCUMENT TRANSLATION EXAMPLE"), "Calibri", 8.5, "7A8294", True)

footer = section.footer.paragraphs[0]

footer.alignment = WD_ALIGN_PARAGRAPH.RIGHT

footer.paragraph_format.space_before = Pt(0)

set_run_font(footer.add_run("Local browser fixture"), "Calibri", 8.5, "7A8294")

kicker = document.add_paragraph()

kicker.paragraph_format.space_before = Pt(8)

kicker.paragraph_format.space_after = Pt(6)

set_run_font(kicker.add_run("REFERENCE GUIDE"), "Calibri", 10, "E83B6B", True)

title = document.add_paragraph()

title.paragraph_format.space_before = Pt(0)

title.paragraph_format.space_after = Pt(6)

set_run_font(title.add_run("Document Translation Example"), "Calibri", 26, "202533", True)

subtitle = document.add_paragraph()

subtitle.paragraph_format.space_before = Pt(0)

subtitle.paragraph_format.space_after = Pt(18)

set_run_font(subtitle.add_run("A realistic DOCX fixture for bilingual reading regression"), "Calibri", 13, "5F687C")

document.add_heading("Why local document translation matters", level=1)

document.add_paragraph(

"A useful document translator should preserve the source file while presenting each translated paragraph directly beneath its original text."

)

paragraph = document.add_paragraph()

set_run_font(paragraph.add_run("Privacy first. "), "Calibri", 11, "202533", True)

set_run_font(

paragraph.add_run("The file is parsed inside the browser, and only extracted text is sent to the translation service selected by the reader."),

"Calibri",

11,

)

document.add_heading("What the regression should protect", level=1)

for label, detail in (

("Structure", "Headings, paragraph order, headers, and footers remain valid after export."),

("Reading flow", "The original paragraph appears first and the editable translation follows naturally."),

("Format", "The downloaded result remains a standards-compliant DOCX that opens in Word and LibreOffice."),

):

paragraph = document.add_paragraph()

set_run_font(paragraph.add_run(f"{label}: "), "Calibri", 11, "2E74B5", True)

set_run_font(paragraph.add_run(detail), "Calibri", 11)

output.parent.mkdir(parents=True, exist_ok=True)

output.parent.mkdir(parents=True, exist_ok=True)

page_width, page_height = letter

pdf = canvas.Canvas(str(output), pagesize=letter)

pdf.setTitle("Document Translation Example")

pdf.setAuthor("FluentRead")

def wrapped_lines(text: str, font: str, size: float, width: float) -> list[str]:

words = text.split()

lines: list[str] = []

current = ""

for word in words:

candidate = f"{current} {word}".strip()

if current and stringWidth(candidate, font, size) > width:

lines.append(current)

current = word

else:

current = candidate

if current:

lines.append(current)

return lines

def draw_wrapped(text: str, x: float, y: float, width: float, *, font: str = "Times-Roman", size: float = 9.2, leading: float = 12.3) -> float:

pdf.setFont(font, size)

pdf.setFillColor(HexColor("#202533"))

for line in wrapped_lines(text, font, size, width):

pdf.drawString(x, y, line)

y -= leading

return y

def draw_heading(text: str, x: float, y: float) -> float:

pdf.setFont("Helvetica-Bold", 11)

pdf.setFillColor(HexColor("#D93463"))

pdf.drawString(x, y, text)

return y - 17

def draw_footer(page_number: int) -> None:

pdf.setStrokeColor(HexColor("#DDE2EC"))

pdf.line(48, 39, page_width - 48, 39)

pdf.setFont("Helvetica", 7.5)

pdf.setFillColor(HexColor("#7A8294"))

pdf.drawString(48, 25, "FLUENTREAD DOCUMENT LAYOUT REGRESSION")

pdf.drawRightString(page_width - 48, 25, str(page_number))

# 第 1 页刻意组合通栏标题、双栏正文和图示，用于捕获 PDF 翻译被压平成

# 提取字符串列表、而没有写回原页面文本框的回归。

pdf.setFont("Helvetica-Bold", 22)

pdf.setFillColor(HexColor("#202533"))

pdf.drawCentredString(page_width / 2, 742, "Document Translation Example")

pdf.setFont("Helvetica", 9)

pdf.setFillColor(HexColor("#647086"))

pdf.drawCentredString(page_width / 2, 724, "Layout-preserving bilingual PDF regression · FluentRead Research")

pdf.setFillColor(HexColor("#F7F8FB"))

pdf.roundRect(48, 652, page_width - 96, 51, 7, fill=1, stroke=0)

pdf.setFillColor(HexColor("#202533"))

pdf.setFont("Helvetica-Bold", 8.5)

pdf.drawString(60, 688, "ABSTRACT")

draw_wrapped(

"A document translator should preserve columns, figures, captions, and reading order while replacing text inside the corresponding page regions.",

60, 674, page_width - 120, font="Helvetica", size=8.4, leading=10.4,

)

left_x, right_x, column_width = 48, 318, 246

left_y = draw_heading("1 PAGE-AWARE TRANSLATION", left_x, 626)

left_y = draw_wrapped(

"FluentRead parses the PDF text layer into layout blocks with page coordinates. Each translated block is fitted back into its original region instead of being shown as a detached transcript.",

left_x, left_y, column_width,

) - 9

left_y = draw_heading("2 READING EXPERIENCE", left_x, left_y)

left_y = draw_wrapped(

"The reader presents the untouched source page and the translated page side by side. Pagination, multi-column flow, graphics, and captions remain visible for visual comparison.",

left_x, left_y, column_width,

) - 9

left_y = draw_heading("3 EXPORT CONTRACT", left_x, left_y)

draw_wrapped(

"A bilingual download contains one wide page for every source page: the original is placed on the left and the translated layout is placed on the right.",

left_x, left_y, column_width,

)

right_y = draw_heading("FIGURE 1 DOCUMENT PIPELINE", right_x, 626)

box_y = right_y - 56

box_width = 66

for index, (label, color) in enumerate((("Parse", "#EAF0FF"), ("Translate", "#FFF0F4"), ("Render", "#E7F8F5"))):

box_x = right_x + index * 84

pdf.setFillColor(HexColor(color))

pdf.setStrokeColor(HexColor("#C9D1E2"))

pdf.roundRect(box_x, box_y, box_width, 38, 6, fill=1, stroke=1)

pdf.setFillColor(HexColor("#202533"))

pdf.setFont("Helvetica-Bold", 8)

pdf.drawCentredString(box_x + box_width / 2, box_y + 15, label)

if index < 2:

pdf.setStrokeColor(HexColor("#9BA6BB"))

pdf.line(box_x + box_width + 5, box_y + 19, box_x + box_width + 15, box_y + 19)

pdf.line(box_x + box_width + 11, box_y + 22, box_x + box_width + 15, box_y + 19)

pdf.line(box_x + box_width + 11, box_y + 16, box_x + box_width + 15, box_y + 19)

right_y = box_y - 19

pdf.setFont("Times-Italic", 8.2)

pdf.setFillColor(HexColor("#5F687C"))

pdf.drawString(right_x, right_y, "Text extraction is an internal step, not the final reading interface.")

right_y -= 29

right_y = draw_heading("4 VISUAL INVARIANTS", right_x, right_y)

draw_wrapped(

"Regression evidence must prove that the translated page retains the figure, column boundaries, title hierarchy, and page dimensions. A generic list of OCR-like text fragments does not satisfy this contract.",

right_x, right_y, column_width,

)

draw_footer(1)

pdf.showPage()

# 第 2 页覆盖表格、图表、通栏说明和独立分栏，确保解析与导出定位都受到验证。

pdf.setFont("Helvetica-Bold", 19)

pdf.setFillColor(HexColor("#202533"))

pdf.drawString(48, 742, "Regression Coverage")

pdf.setFont("Helvetica", 9)

pdf.setFillColor(HexColor("#647086"))

pdf.drawString(48, 724, "Expected invariants for preview and downloadable PDF output")

table_x, table_y = 48, 634

widths = [155, 105, 248]

row_height = 25

headers = ["Capability", "Expected", "Evidence"]

rows = [

("Page dimensions", "Preserved", "Source and translated pages share the same height"),

("Two-column flow", "Preserved", "Blocks remain in their original left or right column"),

("Figures and charts", "Preserved", "Non-text page graphics remain in place"),

]

pdf.setFillColor(HexColor("#202533"))

pdf.rect(table_x, table_y + row_height, sum(widths), row_height, fill=1, stroke=0)

cursor_x = table_x

pdf.setFont("Helvetica-Bold", 8)

pdf.setFillColor(HexColor("#FFFFFF"))

for header, width in zip(headers, widths):

pdf.drawString(cursor_x + 8, table_y + 34, header)

cursor_x += width

for row_index, row in enumerate(rows):

row_y = table_y - row_index * row_height

pdf.setFillColor(HexColor("#F7F8FB") if row_index % 2 == 0 else HexColor("#FFFFFF"))

pdf.setStrokeColor(HexColor("#DDE2EC"))

pdf.rect(table_x, row_y, sum(widths), row_height, fill=1, stroke=1)

cursor_x = table_x

pdf.setFillColor(HexColor("#202533"))

pdf.setFont("Helvetica", 7.8)

for value, width in zip(row, widths):

pdf.drawString(cursor_x + 8, row_y + 9, value)

cursor_x += width

left_y = draw_heading("5 MULTI-COLUMN CONTENT", left_x, 527)

left_y = draw_wrapped(

"Page boundaries and paragraph order survive parsing. The translated page uses the same canvas as the source page before text regions are replaced, which leaves lines, colored shapes, and charts untouched.",

left_x, left_y, column_width,

) - 8

left_y = draw_heading("6 FAILURE BOUNDARY", left_x, left_y)

draw_wrapped(

"Image-only scans without a usable text layer are reported clearly. This beta does not pretend that extracted OCR text alone is a translated PDF document.",

left_x, left_y, column_width,

)

right_y = draw_heading("FIGURE 2 BLOCK RETENTION", right_x, 527)

chart_x, chart_y, chart_w, chart_h = right_x + 12, 382, 215, 92

pdf.setStrokeColor(HexColor("#AAB3C5"))

pdf.line(chart_x, chart_y, chart_x, chart_y + chart_h)

pdf.line(chart_x, chart_y, chart_x + chart_w, chart_y)

for index, (label, value, color) in enumerate((("Text", .86, "#ED4775"), ("Layout", .96, "#5A7BEF"), ("Graphics", 1.0, "#2DB483"))):

bar_x = chart_x + 25 + index * 62

pdf.setFillColor(HexColor(color))

pdf.rect(bar_x, chart_y, 29, chart_h * value, fill=1, stroke=0)

pdf.setFillColor(HexColor("#5F687C"))

pdf.setFont("Helvetica", 7)

pdf.drawCentredString(bar_x + 14.5, chart_y - 11, label)

pdf.setFont("Times-Italic", 8.1)

pdf.setFillColor(HexColor("#5F687C"))

pdf.drawString(right_x, 350, "Figure 2. Layout and non-text graphics remain anchored to the page.")

draw_footer(2)

mimetype = "application/epub+zip"

container = """<?xml version="1.0" encoding="UTF-8"?>

content_opf = """<?xml version="1.0" encoding="UTF-8"?>

nav = """<?xml version="1.0" encoding="UTF-8"?>

chapter_1 = """<?xml version="1.0" encoding="UTF-8"?>

chapter_2 = """<?xml version="1.0" encoding="UTF-8"?>

output.parent.mkdir(parents=True, exist_ok=True)

with zipfile.ZipFile(output, "w") as archive:

archive.writestr("mimetype", mimetype, compress_type=zipfile.ZIP_STORED)

archive.writestr("META-INF/container.xml", container, compress_type=zipfile.ZIP_DEFLATED)

archive.writestr("OEBPS/content.opf", content_opf, compress_type=zipfile.ZIP_DEFLATED)

archive.writestr("OEBPS/nav.xhtml", nav, compress_type=zipfile.ZIP_DEFLATED)

archive.writestr("OEBPS/chapter-1.xhtml", chapter_1, compress_type=zipfile.ZIP_DEFLATED)

parser = argparse.ArgumentParser()

parser.add_argument("--output-dir", type=Path, required=True)

args = parser.parse_args()

generate_pdf(args.output_dir / "sample.pdf")

generate_epub(args.output_dir / "sample.epub")