Skip to main content

Core Concepts

This guide explains the fundamental concepts you need to understand when working with PDFDancer, including both PDF-specific concepts and PDFDancer's content model.

PDF Fundamentals​

PDF Coordinate System​

PDF uses a Cartesian coordinate system with the origin at the bottom-left corner of the page:

  • X-axis: Increases from left (0) to right
  • Y-axis: Increases from bottom (0) to top
  • Units: PDF points (1 point = 1/72 inch)

PDF Coordinate System - Bottom-left origin with X increasing right, Y increasing up

Common Page Sizes:

  • Letter (US): 612 × 792 points (8.5" × 11")
  • A4: 595 × 842 points (210mm × 297mm)
  • Legal: 612 × 1008 points (8.5" × 14")

PDF Points​

All measurements in PDF use points as the base unit:

  • 1 point = 1/72 inch
  • 72 points = 1 inch
  • 1 inch margin = 72 points

Page Sizes​

PDFDancer provides constants for standard page sizes. All dimensions are in points.

ISO A Series:

from pdfdancer import PageSize

# ISO A Series
PageSize.A4     # 595 × 842 points (210mm × 297mm)
PageSize.A3     # 842 × 1191 points (297mm × 420mm)
PageSize.A5     # 420 × 595 points (148mm × 210mm)

US/North American Sizes:

# US Sizes
PageSize.LETTER    # 612 × 792 points (8.5" × 11")
PageSize.LEGAL     # 612 × 1008 points (8.5" × 14")
PageSize.TABLOID   # 792 × 1224 points (11" × 17")

Custom Page Sizes:

# Create custom page size (width, height in points)
custom_size = PageSize(name=None, width=500.0, height=700.0)

Bounding Rectangles​

Every PDF element has a bounding rectangle that defines its position and size:

{
    "x": 100,      # Left edge (from page left)
    "y": 500,      # Bottom edge (from page bottom)
    "width": 200,  # Width in points
    "height": 50   # Height in points
}

PDFDancer Content Model​

PDFDancer provides a structured way to interact with PDF content through several key object types.

Pages​

Pages are the fundamental containers in a PDF document. Each page has:

  • A page number (page 1 is the first page)
  • Dimensions (width and height in points)
  • A bounding rectangle defining its size
  • Content (paragraphs, images, paths, form fields)

Pages are accessed using pdf.page(number):

PDFDancer uses standard page numbering — page 1 is the first page.

# Get first page
first_page = pdf.page(1)

# Get all pages
all_pages = pdf.pages()

Paragraphs​

Paragraphs are PDFDancer's high-level text abstraction. A paragraph represents a logical block of text that may span multiple lines.

Key Properties:

  • text: The complete text content
  • position: Bounding rectangle and page location
  • internal_id: Unique identifier within the PDF

When to use Paragraphs:

  • Finding text blocks by content (e.g., "Invoice #12345")
  • Editing multi-line text blocks
  • Replacing entire sections of text
  • Adding formatted text content
# Select all paragraphs
paragraphs = pdf.select_paragraphs()

# Select by text prefix
headers = pdf.select_paragraphs_starting_with("Chapter")

# Access paragraph properties
for para in paragraphs:
    print(f"Text: {para.text}")
    print(f"Position: {para.position.bounding_rect}")

TextLines​

TextLines represent individual lines of text within a paragraph. They provide finer-grained control than paragraphs.

Key Properties:

  • text: The text content of the line
  • position: Bounding rectangle of the line
  • internal_id: Unique identifier

When to use TextLines:

  • Precise line-by-line text manipulation
  • Finding single-line text elements
  • Working with tabular data or structured text
# Select all text lines
lines = pdf.page(1).select_lines()

# Select lines by prefix
date_lines = pdf.select_text_lines_starting_with("Date:")

for line in lines:
    print(f"Line text: {line.text}")

Paragraph vs TextLine:

  • Use Paragraphs for semantic text blocks (headings, body text, captions)
  • Use TextLines for precise line-level control or single-line elements

Images​

Images represent raster graphics (PNG, JPEG, etc.) embedded in the PDF.

Key Properties:

  • internal_id: Unique identifier
  • position: Bounding rectangle and location
  • Image data (for export/manipulation)

Common Operations:

  • Selecting images by position
  • Adding new images at specific coordinates
  • Deleting existing images
  • Replacing images
# Select all images on a page
images = pdf.page(1).select_images()

# Select images at coordinates
images_at_point = pdf.page(1).select_images_at(x=100, y=500)

# Add a new image
pdf.new_image() \
    .from_file("logo.png") \
    .at(page=1, x=50, y=700) \
    .add()

Paths (Vector Graphics)​

Paths are vector graphics elements that can represent lines, curves, shapes, and complex drawings.

What Paths Represent:

  • Lines and curves (straight lines, Bézier curves)
  • Shapes (rectangles, circles, polygons)
  • Borders and decorative elements
  • Technical drawings and diagrams

Key Concepts:

  • Bézier curves: Mathematical curves defined by control points
  • Stroke: The outline of a path (color, width)
  • Fill: The interior color of closed paths
# Select all paths on a page
paths = pdf.page(1).select_paths()

# Select paths at specific coordinates
paths_at_point = pdf.page(1).select_paths_at(x=150, y=320)

for path in paths:
    print(f"Path ID: {path.internal_id}")

Form Fields (AcroForms)​

Form Fields are interactive elements in PDF forms (AcroForms) that can be filled programmatically.

Common Field Types:

  • Text fields: Single-line or multi-line text input
  • Checkboxes: Boolean on/off values
  • Radio buttons: Single choice from multiple options
  • Dropdowns: Selection from a list

Key Properties:

  • name: Field identifier (e.g., "firstName", "email")
  • object_type: Type of field
  • position: Location on the page
# Select all form fields
fields = pdf.select_form_fields()

# Select by name
name_fields = pdf.select_form_fields_by_name("firstName")

# Fill a field
if name_fields:
    name_fields[0].edit().value("John Doe").apply()

FormXObjects​

FormXObjects (also called XObjects) are reusable content streams that can be referenced multiple times throughout a document.

Use Cases:

  • Company logos appearing on every page
  • Page headers and footers
  • Watermarks
  • Template overlays

Benefits:

  • Efficiency: Content is stored once, referenced many times
  • Consistency: Ensures identical appearance across pages
  • Smaller file size: No content duplication

FormXObjects can be transformed (scaled, rotated, positioned) each time they're used without modifying the original content.

Working with FormXObjects:

from pdfdancer import PDFDancer

with PDFDancer.open("document.pdf") as pdf:
    # Select all FormXObjects on a page
    formxobjects = pdf.page(1).select_formxobjects()

    # Select FormXObjects at specific coordinates
    formxobjects_at_point = pdf.page(1).select_formxobjects_at(x=100, y=500)

    for fxo in formxobjects:
        print(f"FormXObject ID: {fxo.internal_id}")
        print(f"Position: {fxo.position.bounding_rect}")

Fonts​

PDF supports both standard and custom fonts.

Standard PDF Fonts​

The 14 standard PDF fonts are guaranteed to be available in all PDF readers and do not need to be embedded in the PDF document.

Serif Fonts (Times family):

from pdfdancer import StandardFonts

StandardFonts.TIMES_ROMAN          # Times-Roman
StandardFonts.TIMES_BOLD           # Times-Bold
StandardFonts.TIMES_ITALIC         # Times-Italic
StandardFonts.TIMES_BOLD_ITALIC    # Times-BoldItalic

Sans-serif Fonts (Helvetica family):

StandardFonts.HELVETICA              # Helvetica
StandardFonts.HELVETICA_BOLD         # Helvetica-Bold
StandardFonts.HELVETICA_OBLIQUE      # Helvetica-Oblique
StandardFonts.HELVETICA_BOLD_OBLIQUE # Helvetica-BoldOblique

Monospace Fonts (Courier family):

StandardFonts.COURIER              # Courier
StandardFonts.COURIER_BOLD         # Courier-Bold
StandardFonts.COURIER_OBLIQUE      # Courier-Oblique
StandardFonts.COURIER_BOLD_OBLIQUE # Courier-BoldOblique

Symbol and Decorative Fonts:

StandardFonts.SYMBOL          # Symbol (mathematical and special characters)
StandardFonts.ZAPF_DINGBATS   # ZapfDingbats (decorative symbols)

Using Standard Fonts:

# Use standard font constant
pdf.new_paragraph() \
    .text("Hello World") \
    .font(StandardFonts.HELVETICA.value, 12) \
    .add()

# Or use font name string directly
pdf.new_paragraph() \
    .text("Hello World") \
    .font("Helvetica", 12) \
    .add()

Custom Fonts​

PDFDancer supports embedding custom TrueType fonts (.ttf) for precise typography.

# Use standard font
pdf.new_paragraph() \
    .text("Hello World") \
    .font("Helvetica", 12) \
    .add()

# Use custom font
pdf.new_paragraph() \
    .text("Custom Typography") \
    .font_file("custom-font.ttf", 14) \
    .add()

Position Objects​

Position objects encapsulate coordinate information for precise element placement and selection.

Creating Positions​

from pdfdancer import Position, PositionMode

# Create position at point
position = Position.at_page_coordinates(page=0, x=100, y=200)

# Create position with bounding rect
position = Position(
    page_number=1,
    bounding_rect={"x": 100, "y": 200, "width": 50, "height": 30},
    mode=PositionMode.INTERSECT
)

# Use for selection
paragraphs = pdf.select_paragraphs_at(position)

Position Modes​

  • INTERSECT: Select elements that overlap with the position area
  • CONTAIN: Select elements fully contained within the position area
  • EXACT: Select elements at exact coordinates

Color​

PDFDancer uses RGB color values for text and graphics.

from pdfdancer import Color

# Create colors
black = Color(0, 0, 0)
red = Color(255, 0, 0)
gray = Color(128, 128, 128)
custom = Color(70, 130, 180)  # Steel blue

# Apply to text
pdf.new_paragraph() \
    .text("Colored text") \
    .color(red) \
    .add()

Selection vs Creation​

PDFDancer provides two primary workflows:

Selection (Read/Modify)​

Use select_* methods to find existing content:

# Find existing content
paragraphs = pdf.select_paragraphs()
images = pdf.page(1).select_images()
fields = pdf.select_form_fields_by_name("email")

# Modify it
paragraphs[0].edit().replace("New text").apply()

Creation (Add)​

Use new_* methods to add new content:

# Add new content
pdf.new_paragraph() \
    .text("New content") \
    .at(page_number=1, x=100, y=500) \
    .add()

pdf.new_image() \
    .from_file("logo.png") \
    .at(page=1, x=50, y=700) \
    .add()

Fluent Builders​

PDFDancer uses fluent builder patterns for creating and editing content. Builders allow you to chain method calls for readable, declarative code:

# Paragraph builder
pdf.new_paragraph() \
    .text("Hello World") \
    .font("Helvetica", 12) \
    .color(Color(0, 0, 0)) \
    .line_spacing(1.5) \
    .at(page_number=1, x=100, y=500) \
    .add()

# Edit builder
paragraph.edit() \
    .replace("New text") \
    .font("Helvetica-Bold", 14) \
    .color(Color(255, 0, 0)) \
    .apply()

Thread Safety​

Important

PDFDancer sessions are not thread-safe and must not be used concurrently.

Each session instance should only be accessed from a single thread at a time. Do not share session objects across threads or use them in concurrent operations.

Why This Matters:

When you call PDFDancer.open(), you create a session that maintains state on both the client and server. Concurrent access from multiple threads can lead to:

  • Race conditions and unpredictable behavior
  • Corrupted PDF state
  • API errors and failed operations

Safe Patterns:

from pdfdancer import PDFDancer
from concurrent.futures import ThreadPoolExecutor

# ✓ SAFE: Each thread creates its own session
def process_pdf(file_path: str) -> None:
    with PDFDancer.open(file_path) as pdf:
        # Operations on this session
        paragraphs = pdf.select_paragraphs()
        pdf.save(f"output_{file_path}")

# Process multiple PDFs in parallel - each gets its own session
with ThreadPoolExecutor() as executor:
    executor.map(process_pdf, ["doc1.pdf", "doc2.pdf", "doc3.pdf"])


# ✗ UNSAFE: Sharing a session across threads
pdf = PDFDancer.open("document.pdf")
def unsafe_operation():
    # DON'T DO THIS - multiple threads using the same session
    pdf.select_paragraphs()  # Not thread-safe!

with ThreadPoolExecutor() as executor:
    executor.submit(unsafe_operation)
    executor.submit(unsafe_operation)

Best Practice: Always create a new session instance for each thread or concurrent operation. Sessions are lightweight to create and are designed for single-threaded access.

Next Steps​

Now that you understand the core concepts, explore how to use them: