Merge pull request #516 from Kozea/pdfrw

Use pdfrw instead of editing PDF files directly

docs/features.rst

@@ -272,14 +272,13 @@ paged media "describing how:
 - content such as page counters are placed in the headers and footers; and
 - orphans and widows can be controlled."
 
-The ``bleed`` and ``marks`` properties from this document are **not**
-implemented. All the other features are available, including:
+All the features of this draft are available, including:
 
 - the ``@page`` rule and the ``:left``, ``:right``, ``:first`` and ``:blank``
   selectors;
 - the page margin boxes;
 - the page-based counters (with known bugs `#91`_, `#93`_, `#289`_);
-- the page ``size`` property;
+- the page ``size``, ``bleed`` and ``marks`` properties;
 - the named pages.
 
 .. _CSS Paged Media Module Level 3: http://dev.w3.org/csswg/css3-page/

docs/install.rst

@@ -13,6 +13,7 @@ WeasyPrint |version| depends on:
 * cssselect2_ ≥ 0.1
 * CairoSVG_ ≥ 1.0.20
 * Pyphen_ ≥ 0.8
+* pdfrw_ ≥ 0.4
 * Optional: GDK-PixBuf_ [#]_
 
 .. _CPython: http://www.python.org/
@@ -25,6 +26,7 @@ WeasyPrint |version| depends on:
 .. _cssselect2: https://cssselect2.readthedocs.io/
 .. _CairoSVG: http://cairosvg.org/
 .. _Pyphen: http://pyphen.org/
+.. _pdfrw: https://github.com/pmaupin/pdfrw/
 .. _GDK-PixBuf: https://live.gnome.org/GdkPixbuf
 
 

setup.py

@@ -34,7 +34,8 @@
     'cssselect2>=0.1',
     'cffi>=0.6',
     'cairocffi>=0.5',
-    'Pyphen>=0.8'
+    'Pyphen>=0.8',
+    'pdfrw>=0.4',
     # C dependencies: Gdk-Pixbuf (optional), Pango, cairo.
 ]
 

weasyprint/css/__init__.py

@@ -43,7 +43,7 @@
 # A test function that returns True if the given property name has an
 # initial value that is not always the same when computed.
 RE_INITIAL_NOT_COMPUTED = re.compile(
-    '^(display|column_gap|'
+    '^(display|column_gap|(bleed_(left|right|top|bottom))|'
     '(border_[a-z]+|outline|column_rule)_(width|color))$').match
 
 

weasyprint/css/computed_values.py

@@ -135,7 +135,7 @@ def _computing_order():
     """Some computed values are required by others, so order matters."""
     first = [
         'font_stretch', 'font_weight', 'font_family', 'font_variant',
-        'font_style', 'font_size', 'line_height']
+        'font_style', 'font_size', 'line_height', 'marks']
     order = sorted(INITIAL_VALUES)
     for name in first:
         order.remove(name)
@@ -329,6 +329,20 @@ def length(computer, name, value, font_size=None, pixels_only=False):
     return result if pixels_only else Dimension(result, 'px')
 
 
+@register_computer('bleed-left')
+@register_computer('bleed-right')
+@register_computer('bleed-top')
+@register_computer('bleed-bottom')
+def bleed(computer, name, value):
+    if value == 'auto':
+        if 'crop' in computer.computed['marks']:
+            return Dimension(8, 'px')  # 6pt
+        else:
+            return Dimension(0, 'px')
+    else:
+        return length(computer, name, value)
+
+
 @register_computer('letter-spacing')
 def pixel_length(computer, name, value):
     if value == 'normal':

weasyprint/css/properties.py

@@ -146,6 +146,11 @@
     # Paged Media 3 (WD): https://www.w3.org/TR/css3-page/
     'size': None,  # set to A4 in computed_values
     'page': 'auto',
+    'bleed_left': 'auto',
+    'bleed_right': 'auto',
+    'bleed_top': 'auto',
+    'bleed_bottom': 'auto',
+    'marks': 'none',
 
     # Text 3/4 (WD/WD): https://www.w3.org/TR/css-text-4/
     'hyphenate_character': '‐',  # computed value chosen by the user agent

weasyprint/css/tests_ua.css

@@ -1,6 +1,7 @@
 /*
     Simplified user-agent stylesheet for HTML5 in tests.
 */
+@page { bleed: 0 }
 html, body, div, h1, h2, h3, h4, ol, p, ul, hr, pre, section, article
                 { display: block; }
 li              { display: list-item }

weasyprint/css/validation.py

@@ -617,6 +617,35 @@ def page(token):
         return 'auto' if token.lower_value == 'auto' else token.value
 
 
+@validator("bleed-left")
+@validator("bleed-right")
+@validator("bleed-top")
+@validator("bleed-bottom")
+@single_token
+def bleed(token):
+    """``bleed`` property validation."""
+    keyword = get_keyword(token)
+    if keyword == 'auto':
+        return 'auto'
+    else:
+        return get_length(token)
+
+
+@validator()
+def marks(tokens):
+    """``marks`` property validation."""
+    if len(tokens) == 2:
+        keywords = [get_keyword(token) for token in tokens]
+        if 'crop' in keywords and 'cross' in keywords:
+            return keywords
+    elif len(tokens) == 1:
+        keyword = get_keyword(tokens[0])
+        if keyword in ('crop', 'cross'):
+            return [keyword]
+        elif keyword == 'none':
+            return 'none'
+
+
 @validator('outline-style')
 @single_keyword
 def outline_style(keyword):
@@ -1632,6 +1661,7 @@ def expander_decorator(function):
 @expander('border-width')
 @expander('margin')
 @expander('padding')
+@expander('bleed')
 def expand_four_sides(base_url, name, tokens):
     """Expand properties setting a token for the four sides of a box."""
     # Make sure we have 4 tokens

weasyprint/document.py

@@ -156,6 +156,11 @@ def __init__(self, page_box, enable_hinting=False):
         #: The page height, including margins, in CSS pixels.
         self.height = page_box.margin_height()
 
+        #: The page bleed width, in CSS pixels.
+        self.bleed = {
+            side: page_box.style['bleed_%s' % side].value
+            for side in ('top', 'right', 'bottom', 'left')}
+
         #: A list of ``(bookmark_level, bookmark_label, target)`` tuples.
         #: :obj:`bookmark_level` and :obj:`bookmark_label` are respectively
         #: an integer and a Unicode string, based on the CSS properties
@@ -481,10 +486,15 @@ def write_pdf(self, target=None, zoom=1, attachments=None):
         LOGGER.info('Step 6 - Drawing')
         for page in self.pages:
             surface.set_size(
-                math.floor(page.width * scale),
-                math.floor(page.height * scale))
-            page.paint(context, scale=scale)
-            surface.show_page()
+                math.floor(scale * (
+                    page.width + page.bleed['left'] + page.bleed['right'])),
+                math.floor(scale * (
+                    page.height + page.bleed['top'] + page.bleed['bottom'])))
+            with stacked(context):
+                context.translate(
+                    page.bleed['left'] * scale, page.bleed['top'] * scale)
+                page.paint(context, scale=scale)
+                surface.show_page()
         surface.finish()
 
         LOGGER.info('Step 7 - Adding PDF metadata')

weasyprint/draw.py

@@ -20,10 +20,67 @@
 
 from .compat import xrange
 from .formatting_structure import boxes
+from .images import SVGImage
+from .layout.backgrounds import BackgroundLayer
 from .stacking import StackingContext
 from .text import show_first_line
 
 SIDES = ('top', 'right', 'bottom', 'left')
+CROP = '''
+  <!-- horizontal top left -->
+  <path d="M0,{bleed_top} h{half_bleed_left}" />
+  <!-- horizontal top right -->
+  <path d="M0,{bleed_top} h{half_bleed_right}"
+        transform="translate({width},0) scale(-1,1)" />
+  <!-- horizontal bottom right -->
+  <path d="M0,{bleed_bottom} h{half_bleed_right}"
+        transform="translate({width},{height}) scale(-1,-1)" />
+  <!-- horizontal bottom left -->
+  <path d="M0,{bleed_bottom} h{half_bleed_left}"
+        transform="translate(0,{height}) scale(1,-1)" />
+  <!-- vertical top left -->
+  <path d="M{bleed_left},0 v{half_bleed_top}" />
+  <!-- vertical bottom right -->
+  <path d="M{bleed_right},0 v{half_bleed_bottom}"
+        transform="translate({width},{height}) scale(-1,-1)" />
+  <!-- vertical bottom left -->
+  <path d="M{bleed_left},0 v{half_bleed_bottom}"
+        transform="translate(0,{height}) scale(1,-1)" />
+  <!-- vertical top right -->
+  <path d="M{bleed_right},0 v{half_bleed_top}"
+        transform="translate({width},0) scale(-1,1)" />
+'''
+CROSS = '''
+  <!-- top -->
+  <circle r="{half_bleed_top}"
+          transform="scale(0.5)
+                     translate({width},{half_bleed_top}) scale(0.5)" />
+  <path d="M-{half_bleed_top},{half_bleed_top} h{bleed_top}
+           M0,0 v{bleed_top}"
+        transform="scale(0.5) translate({width},0)" />
+  <!-- bottom -->
+  <circle r="{half_bleed_bottom}"
+          transform="translate(0,{height}) scale(0.5)
+                     translate({width},-{half_bleed_bottom}) scale(0.5)" />
+  <path d="M-{half_bleed_bottom},-{half_bleed_bottom} h{bleed_bottom}
+           M0,0 v-{bleed_bottom}"
+        transform="translate(0,{height}) scale(0.5) translate({width},0)" />
+  <!-- left -->
+  <circle r="{half_bleed_left}"
+          transform="scale(0.5)
+                     translate({half_bleed_left},{height}) scale(0.5)" />
+  <path d="M{half_bleed_left},-{half_bleed_left} v{bleed_left}
+           M0,0 h{bleed_left}"
+        transform="scale(0.5) translate(0,{height})" />
+  <!-- right -->
+  <circle r="{half_bleed_right}"
+          transform="translate({width},0) scale(0.5)
+                     translate(-{half_bleed_right},{height}) scale(0.5)" />
+  <path d="M-{half_bleed_right},-{half_bleed_right} v{bleed_right}
+           M0,0 h-{bleed_right}"
+        transform="translate({width},0)
+                   scale(0.5) translate(0,{height})" />
+'''
 
 
 @contextlib.contextmanager
@@ -91,10 +148,14 @@ def lighten(color):
 
 def draw_page(page, context, enable_hinting):
     """Draw the given PageBox."""
+    bleed = {
+        side: page.style['bleed_%s' % side].value
+        for side in ('top', 'right', 'bottom', 'left')}
+    marks = page.style['marks']
     stacking_context = StackingContext.from_page(page)
     draw_background(
         context, stacking_context.box.background, enable_hinting,
-        clip_box=False)
+        clip_box=False, bleed=bleed, marks=marks)
     draw_background(
         context, page.canvas_background, enable_hinting, clip_box=False)
     draw_border(context, page, enable_hinting)
@@ -261,7 +322,8 @@ def rounded_box_path(context, radii):
         context.restore()
 
 
-def draw_background(context, bg, enable_hinting, clip_box=True):
+def draw_background(context, bg, enable_hinting, clip_box=True, bleed=None,
+                    marks=()):
     """Draw the background color and image to a ``cairo.Context``.
 
     If ``clip_box`` is set to ``False``, the background is not clipped to the
@@ -286,11 +348,57 @@ def draw_background(context, bg, enable_hinting, clip_box=True):
             with stacked(context):
                 painting_area = bg.layers[-1].painting_area
                 if painting_area:
+                    if bleed:
+                        # Painting area is the PDF BleedBox
+                        x, y, width, height = painting_area
+                        painting_area = (
+                            x - bleed['left'], y - bleed['top'],
+                            width + bleed['left'] + bleed['right'],
+                            height + bleed['top'] + bleed['bottom'])
                     context.rectangle(*painting_area)
                     context.clip()
                 context.set_source_rgba(*bg.color)
                 context.paint()
 
+        if bleed and marks:
+            x, y, width, height = bg.layers[-1].painting_area
+            x -= bleed['left']
+            y -= bleed['top']
+            width += bleed['left'] + bleed['right']
+            height += bleed['top'] + bleed['bottom']
+            svg = '''
+              <svg height="{height}" width="{width}"
+                   fill="transparent" stroke="black" stroke-width="1"
+                   xmlns="http://www.w3.org/2000/svg"
+                   xmlns:xlink="http://www.w3.org/1999/xlink">
+            '''
+            if 'crop' in marks:
+                svg += CROP
+            if 'cross' in marks:
+                svg += CROSS
+            svg += '</svg>'
+            half_bleed = {key: value * 0.5 for key, value in bleed.items()}
+            image = SVGImage(svg.format(
+                height=height, width=width,
+                bleed_left=bleed['left'], bleed_right=bleed['right'],
+                bleed_top=bleed['top'], bleed_bottom=bleed['bottom'],
+                half_bleed_left=half_bleed['left'],
+                half_bleed_right=half_bleed['right'],
+                half_bleed_top=half_bleed['top'],
+                half_bleed_bottom=half_bleed['bottom'],
+            ), '', None)
+            # Painting area is the PDF media box
+            size = (width, height)
+            position = (x, y)
+            repeat = ('no-repeat', 'no-repeat')
+            unbounded = True
+            painting_area = position + size
+            positioning_area = (0, 0, width, height)
+            clipped_boxes = []
+            layer = BackgroundLayer(
+                image, size, position, repeat, unbounded, painting_area,
+                positioning_area, clipped_boxes)
+            bg.layers.insert(0, layer)
         # Paint in reversed order: first layer is "closest" to the viewer.
         for layer in reversed(bg.layers):
             draw_background_image(context, layer, bg.image_rendering)

weasyprint/layout/backgrounds.py

@@ -60,14 +60,16 @@ def layout_box_backgrounds(page, box, get_image_from_uri):
     style = box.style
     if style.visibility == 'hidden':
         box.background = None
-        return
+        if page != box:  # Pages need a background for bleed box
+            return
 
     images = [get_image_from_uri(value) if type_ == 'url' else value
               for type_, value in style.background_image]
     color = style.get_color('background_color')
     if color.alpha == 0 and not any(images):
         box.background = None
-        return
+        if page != box:  # Pages need a background for bleed box
+            return
 
     box.background = Background(
         color=color, image_rendering=style.image_rendering, layers=[