2015-11-25 10:38:01 +03:00
|
|
|
|
# coding: utf-8
|
2012-10-02 20:59:02 +04:00
|
|
|
|
"""
|
|
|
|
|
weasyprint.document
|
|
|
|
|
-------------------
|
|
|
|
|
|
2014-01-10 18:27:02 +04:00
|
|
|
|
:copyright: Copyright 2011-2014 Simon Sapin and contributors, see AUTHORS.
|
2012-10-02 20:59:02 +04:00
|
|
|
|
:license: BSD, see LICENSE for details.
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
|
|
|
|
|
from __future__ import division, unicode_literals
|
|
|
|
|
|
2017-03-25 02:33:36 +03:00
|
|
|
|
import functools
|
2012-10-02 20:59:02 +04:00
|
|
|
|
import io
|
|
|
|
|
import math
|
|
|
|
|
import shutil
|
|
|
|
|
|
2012-12-29 04:00:30 +04:00
|
|
|
|
import cairocffi as cairo
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2017-03-25 02:33:36 +03:00
|
|
|
|
from . import CSS, images
|
|
|
|
|
from .compat import FILESYSTEM_ENCODING, iteritems, izip
|
2012-10-02 20:59:02 +04:00
|
|
|
|
from .css import get_all_computed_styles
|
2017-03-25 02:33:36 +03:00
|
|
|
|
from .draw import draw_page, stacked
|
2016-10-27 18:36:24 +03:00
|
|
|
|
from .fonts import FontConfiguration
|
2012-10-04 13:35:25 +04:00
|
|
|
|
from .formatting_structure import boxes
|
2012-10-02 20:59:02 +04:00
|
|
|
|
from .formatting_structure.build import build_formatting_structure
|
|
|
|
|
from .layout import layout_document
|
2013-04-03 19:46:14 +04:00
|
|
|
|
from .layout.backgrounds import percentage
|
2017-03-25 02:33:36 +03:00
|
|
|
|
from .logger import LOGGER
|
2012-10-02 20:59:02 +04:00
|
|
|
|
from .pdf import write_pdf_metadata
|
|
|
|
|
|
|
|
|
|
|
2012-10-06 13:26:55 +04:00
|
|
|
|
def _get_matrix(box):
|
|
|
|
|
"""Return the matrix for the CSS transforms on this box.
|
|
|
|
|
|
2013-03-28 19:41:51 +04:00
|
|
|
|
:returns: a :class:`cairocffi.Matrix` object or :obj:`None`.
|
2012-10-06 13:26:55 +04:00
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
# "Transforms apply to block-level and atomic inline-level elements,
|
|
|
|
|
# but do not apply to elements which may be split into
|
|
|
|
|
# multiple inline-level boxes."
|
|
|
|
|
# http://www.w3.org/TR/css3-2d-transforms/#introduction
|
|
|
|
|
if box.style.transform and not isinstance(box, boxes.InlineBox):
|
|
|
|
|
border_width = box.border_width()
|
|
|
|
|
border_height = box.border_height()
|
|
|
|
|
origin_x, origin_y = box.style.transform_origin
|
|
|
|
|
origin_x = box.border_box_x() + percentage(origin_x, border_width)
|
|
|
|
|
origin_y = box.border_box_y() + percentage(origin_y, border_height)
|
|
|
|
|
|
|
|
|
|
matrix = cairo.Matrix()
|
|
|
|
|
matrix.translate(origin_x, origin_y)
|
|
|
|
|
for name, args in box.style.transform:
|
|
|
|
|
if name == 'scale':
|
|
|
|
|
matrix.scale(*args)
|
|
|
|
|
elif name == 'rotate':
|
|
|
|
|
matrix.rotate(args)
|
|
|
|
|
elif name == 'translate':
|
|
|
|
|
translate_x, translate_y = args
|
|
|
|
|
matrix.translate(
|
|
|
|
|
percentage(translate_x, border_width),
|
|
|
|
|
percentage(translate_y, border_height),
|
|
|
|
|
)
|
|
|
|
|
else:
|
|
|
|
|
if name == 'skewx':
|
|
|
|
|
args = (1, 0, math.tan(args), 1, 0, 0)
|
|
|
|
|
elif name == 'skewy':
|
|
|
|
|
args = (1, math.tan(args), 0, 1, 0, 0)
|
|
|
|
|
else:
|
|
|
|
|
assert name == 'matrix'
|
|
|
|
|
matrix = cairo.Matrix(*args) * matrix
|
|
|
|
|
matrix.translate(-origin_x, -origin_y)
|
|
|
|
|
box.transformation_matrix = matrix
|
|
|
|
|
return matrix
|
|
|
|
|
|
|
|
|
|
|
2012-10-07 00:09:17 +04:00
|
|
|
|
def rectangle_aabb(matrix, pos_x, pos_y, width, height):
|
|
|
|
|
"""Apply a transformation matrix to an axis-aligned rectangle
|
|
|
|
|
and return its axis-aligned bounding box as ``(x, y, width, height)``
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
transform_point = matrix.transform_point
|
|
|
|
|
x1, y1 = transform_point(pos_x, pos_y)
|
|
|
|
|
x2, y2 = transform_point(pos_x + width, pos_y)
|
|
|
|
|
x3, y3 = transform_point(pos_x, pos_y + height)
|
|
|
|
|
x4, y4 = transform_point(pos_x + width, pos_y + height)
|
|
|
|
|
box_x1 = min(x1, x2, x3, x4)
|
|
|
|
|
box_y1 = min(y1, y2, y3, y4)
|
|
|
|
|
box_x2 = max(x1, x2, x3, x4)
|
|
|
|
|
box_y2 = max(y1, y2, y3, y4)
|
|
|
|
|
return box_x1, box_y1, box_x2 - box_x1, box_y2 - box_y1
|
|
|
|
|
|
|
|
|
|
|
2012-10-04 13:35:25 +04:00
|
|
|
|
class _TaggedTuple(tuple):
|
|
|
|
|
"""A tuple with a :attr:`sourceline` attribute,
|
|
|
|
|
The line number in the HTML source for whatever the tuple represents.
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
|
2012-12-20 23:10:48 +04:00
|
|
|
|
|
2013-07-14 15:08:02 +04:00
|
|
|
|
def _gather_links_and_bookmarks(box, bookmarks, links, anchors, matrix):
|
|
|
|
|
transform = _get_matrix(box)
|
|
|
|
|
if transform:
|
|
|
|
|
matrix = transform * matrix if matrix else transform
|
|
|
|
|
|
2015-05-01 22:14:10 +03:00
|
|
|
|
bookmark_label = box.style.bookmark_label
|
|
|
|
|
if box.style.bookmark_level == 'none':
|
|
|
|
|
bookmark_level = None
|
|
|
|
|
else:
|
|
|
|
|
bookmark_level = box.style.bookmark_level
|
2012-10-04 13:35:25 +04:00
|
|
|
|
link = box.style.link
|
|
|
|
|
anchor_name = box.style.anchor
|
|
|
|
|
has_bookmark = bookmark_label and bookmark_level
|
|
|
|
|
# 'link' is inherited but redundant on text boxes
|
|
|
|
|
has_link = link and not isinstance(box, boxes.TextBox)
|
|
|
|
|
# In case of duplicate IDs, only the first is an anchor.
|
|
|
|
|
has_anchor = anchor_name and anchor_name not in anchors
|
2014-04-23 19:10:31 +04:00
|
|
|
|
is_attachment = hasattr(box, 'is_attachment') and box.is_attachment
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
|
|
|
|
if has_bookmark or has_link or has_anchor:
|
|
|
|
|
pos_x, pos_y, width, height = box.hit_area()
|
|
|
|
|
if has_link:
|
|
|
|
|
link_type, target = link
|
2016-08-17 19:30:53 +03:00
|
|
|
|
if isinstance(target, bytes):
|
|
|
|
|
# Links are filesystem_encoding/utf-8 bytestrings in Python 2
|
|
|
|
|
# and ASCII unicode in Python 3. See ``iri_to_uri`` and
|
|
|
|
|
# standard library's ``quote`` source.
|
|
|
|
|
target = target.decode(
|
|
|
|
|
FILESYSTEM_ENCODING if target.startswith('file:')
|
|
|
|
|
else 'utf-8')
|
2014-04-04 20:46:00 +04:00
|
|
|
|
if link_type == 'external' and is_attachment:
|
|
|
|
|
link_type = 'attachment'
|
2012-10-07 00:09:17 +04:00
|
|
|
|
if matrix:
|
|
|
|
|
link = _TaggedTuple(
|
|
|
|
|
(link_type, target, rectangle_aabb(
|
|
|
|
|
matrix, pos_x, pos_y, width, height)))
|
|
|
|
|
else:
|
|
|
|
|
link = _TaggedTuple(
|
|
|
|
|
(link_type, target, (pos_x, pos_y, width, height)))
|
2012-10-04 13:35:25 +04:00
|
|
|
|
link.sourceline = box.sourceline
|
|
|
|
|
links.append(link)
|
2012-10-07 00:09:17 +04:00
|
|
|
|
if matrix and (has_bookmark or has_anchor):
|
|
|
|
|
pos_x, pos_y = matrix.transform_point(pos_x, pos_y)
|
|
|
|
|
if has_bookmark:
|
|
|
|
|
bookmarks.append((bookmark_level, bookmark_label, (pos_x, pos_y)))
|
2012-10-04 13:35:25 +04:00
|
|
|
|
if has_anchor:
|
|
|
|
|
anchors[anchor_name] = pos_x, pos_y
|
|
|
|
|
|
|
|
|
|
for child in box.all_children():
|
2013-07-14 15:08:02 +04:00
|
|
|
|
_gather_links_and_bookmarks(child, bookmarks, links, anchors, matrix)
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
|
|
|
|
|
2012-10-02 20:59:02 +04:00
|
|
|
|
class Page(object):
|
2012-10-04 13:35:25 +04:00
|
|
|
|
"""Represents a single rendered page.
|
|
|
|
|
|
2012-10-08 21:51:18 +04:00
|
|
|
|
.. versionadded:: 0.15
|
|
|
|
|
|
2012-10-04 13:35:25 +04:00
|
|
|
|
Should be obtained from :attr:`Document.pages` but not
|
|
|
|
|
instantiated directly.
|
|
|
|
|
|
|
|
|
|
"""
|
2012-10-05 22:12:05 +04:00
|
|
|
|
def __init__(self, page_box, enable_hinting=False):
|
2012-10-08 21:51:18 +04:00
|
|
|
|
#: The page width, including margins, in CSS pixels.
|
2012-10-05 22:12:05 +04:00
|
|
|
|
self.width = page_box.margin_width()
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
2012-10-08 21:51:18 +04:00
|
|
|
|
#: The page height, including margins, in CSS pixels.
|
2012-10-05 22:12:05 +04:00
|
|
|
|
self.height = page_box.margin_height()
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
2012-10-08 21:51:18 +04:00
|
|
|
|
#: A list of ``(bookmark_level, bookmark_label, target)`` tuples.
|
|
|
|
|
#: :obj:`bookmark_level` and :obj:`bookmark_label` are respectively
|
|
|
|
|
#: an integer and an Unicode string, based on the CSS properties
|
|
|
|
|
#: of the same names. :obj:`target` is a ``(x, y)`` point
|
|
|
|
|
#: in CSS pixels from the top-left of the page.
|
2012-10-06 13:26:55 +04:00
|
|
|
|
self.bookmarks = bookmarks = []
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
2012-10-05 20:50:40 +04:00
|
|
|
|
#: A list of ``(link_type, target, rectangle)`` tuples.
|
2012-10-08 21:51:18 +04:00
|
|
|
|
#: A rectangle is ``(x, y, width, height)``, in CSS pixels from
|
|
|
|
|
#: the top-left of the page. :obj:`link_type` is one of two strings:
|
2012-10-04 13:35:25 +04:00
|
|
|
|
#:
|
|
|
|
|
#: * ``'external'``: :obj:`target` is an absolute URL
|
|
|
|
|
#: * ``'internal'``: :obj:`target` is an anchor name (see
|
2013-07-14 15:08:02 +04:00
|
|
|
|
#: :attr:`Page.anchors`).
|
|
|
|
|
# The anchor might be defined in another page,
|
|
|
|
|
# in multiple pages (in which case the first occurence is used),
|
|
|
|
|
# or not at all.
|
2014-04-04 20:46:00 +04:00
|
|
|
|
#: * ``'attachment'``: :obj:`target` is an absolute URL and points
|
|
|
|
|
#: to a resource to attach to the document.
|
2012-10-06 13:26:55 +04:00
|
|
|
|
self.links = links = []
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
2012-10-08 21:51:18 +04:00
|
|
|
|
#: A dict mapping anchor names to their target, ``(x, y)`` points
|
|
|
|
|
#: in CSS pixels form the top-left of the page.)
|
2012-10-06 13:26:55 +04:00
|
|
|
|
self.anchors = anchors = {}
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
2013-07-14 15:08:02 +04:00
|
|
|
|
_gather_links_and_bookmarks(
|
|
|
|
|
page_box, bookmarks, links, anchors, matrix=None)
|
2012-10-05 22:12:05 +04:00
|
|
|
|
self._page_box = page_box
|
2012-10-04 13:35:25 +04:00
|
|
|
|
self._enable_hinting = enable_hinting
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2012-10-05 22:12:05 +04:00
|
|
|
|
def paint(self, cairo_context, left_x=0, top_y=0, scale=1, clip=False):
|
2013-03-28 15:44:28 +04:00
|
|
|
|
"""Paint the page in cairo, on any type of surface.
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2013-03-28 19:41:51 +04:00
|
|
|
|
:param cairo_context:
|
|
|
|
|
Any :class:`cairocffi.Context` object.
|
|
|
|
|
|
|
|
|
|
.. note::
|
|
|
|
|
|
|
|
|
|
In case you get a :class:`cairo.Context` object
|
|
|
|
|
(eg. form PyGTK),
|
|
|
|
|
it is possible to :ref:`convert it to cairocffi
|
|
|
|
|
<converting_pycairo>`.
|
2012-10-02 20:59:02 +04:00
|
|
|
|
:param left_x:
|
2012-10-05 22:12:05 +04:00
|
|
|
|
X coordinate of the left of the page, in cairo user units.
|
2012-10-02 20:59:02 +04:00
|
|
|
|
:param top_y:
|
2012-10-05 22:12:05 +04:00
|
|
|
|
Y coordinate of the top of the page, in cairo user units.
|
|
|
|
|
:param scale:
|
2012-10-08 21:51:18 +04:00
|
|
|
|
Zoom scale in cairo user units per CSS pixel.
|
2012-10-02 20:59:02 +04:00
|
|
|
|
:param clip:
|
|
|
|
|
Whether to clip/cut content outside the page. If false or
|
|
|
|
|
not provided, content can overflow.
|
2013-03-28 19:41:51 +04:00
|
|
|
|
:type left_x: float
|
|
|
|
|
:type top_y: float
|
|
|
|
|
:type scale: float
|
|
|
|
|
:type clip: bool
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
with stacked(cairo_context):
|
|
|
|
|
if self._enable_hinting:
|
|
|
|
|
left_x, top_y = cairo_context.user_to_device(left_x, top_y)
|
|
|
|
|
# Hint in device space
|
|
|
|
|
left_x = int(left_x)
|
|
|
|
|
top_y = int(top_y)
|
|
|
|
|
left_x, top_y = cairo_context.device_to_user(left_x, top_y)
|
2012-10-05 22:12:05 +04:00
|
|
|
|
# Make (0, 0) the top-left corner:
|
2012-10-02 20:59:02 +04:00
|
|
|
|
cairo_context.translate(left_x, top_y)
|
2012-10-05 22:12:05 +04:00
|
|
|
|
# Make user units CSS pixels:
|
|
|
|
|
cairo_context.scale(scale, scale)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
if clip:
|
2012-10-05 22:12:05 +04:00
|
|
|
|
width = self.width
|
|
|
|
|
height = self.height
|
|
|
|
|
if self._enable_hinting:
|
|
|
|
|
width, height = (
|
|
|
|
|
cairo_context.user_to_device_distance(width, height))
|
|
|
|
|
# Hint in device space
|
|
|
|
|
width = int(math.ceil(width))
|
|
|
|
|
height = int(math.ceil(height))
|
|
|
|
|
width, height = (
|
|
|
|
|
cairo_context.device_to_user_distance(width, height))
|
2012-10-02 20:59:02 +04:00
|
|
|
|
cairo_context.rectangle(0, 0, width, height)
|
|
|
|
|
cairo_context.clip()
|
|
|
|
|
draw_page(self._page_box, cairo_context, self._enable_hinting)
|
|
|
|
|
|
|
|
|
|
|
2013-07-14 15:08:02 +04:00
|
|
|
|
class DocumentMetadata(object):
|
|
|
|
|
"""Contains meta-information about a :class:`Document`
|
|
|
|
|
that do not belong to specific pages but to the whole document.
|
|
|
|
|
|
|
|
|
|
New attributes may be added in future versions of WeasyPrint.
|
|
|
|
|
|
|
|
|
|
.. _W3C’s profile of ISO 8601: http://www.w3.org/TR/NOTE-datetime
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
def __init__(self, title=None, authors=None, description=None,
|
2014-04-04 14:32:21 +04:00
|
|
|
|
keywords=None, generator=None, created=None, modified=None,
|
|
|
|
|
attachments=None):
|
2013-07-14 15:08:02 +04:00
|
|
|
|
#: The title of the document, as a string or :obj:`None`.
|
|
|
|
|
#: Extracted from the ``<title>`` element in HTML
|
|
|
|
|
#: and written to the ``/Title`` info field in PDF.
|
|
|
|
|
self.title = title
|
|
|
|
|
#: The authors of the document as a list of strings.
|
|
|
|
|
#: Extracted from the ``<meta name=author>`` elements in HTML
|
|
|
|
|
#: and written to the ``/Author`` info field in PDF.
|
|
|
|
|
self.authors = authors or []
|
|
|
|
|
#: The description of the document, as a string or :obj:`None`.
|
|
|
|
|
#: Extracted from the ``<meta name=description>`` element in HTML
|
|
|
|
|
#: and written to the ``/Subject`` info field in PDF.
|
|
|
|
|
self.description = description
|
|
|
|
|
#: Keywords associated with the document, as a list of strings.
|
|
|
|
|
#: (Defaults to the empty list.)
|
|
|
|
|
#: Extracted from ``<meta name=keywords>`` elements in HTML
|
|
|
|
|
#: and written to the ``/Keywords`` info field in PDF.
|
|
|
|
|
self.keywords = keywords or []
|
|
|
|
|
#: The name of one of the software packages
|
|
|
|
|
#: used to generate the document, as a string or :obj:`None`.
|
|
|
|
|
#: Extracted from the ``<meta name=generator>`` element in HTML
|
|
|
|
|
#: and written to the ``/Creator`` info field in PDF.
|
|
|
|
|
self.generator = generator
|
|
|
|
|
#: The creation date of the document, as a string or :obj:`None`.
|
|
|
|
|
#: Dates are in one of the six formats specified in
|
|
|
|
|
#: `W3C’s profile of ISO 8601`_.
|
|
|
|
|
#: Extracted from the ``<meta name=dcterms.created>`` element in HTML
|
|
|
|
|
#: and written to the ``/CreationDate`` info field in PDF.
|
|
|
|
|
self.created = created
|
|
|
|
|
#: The modification date of the document, as a string or :obj:`None`.
|
|
|
|
|
#: Dates are in one of the six formats specified in
|
|
|
|
|
#: `W3C’s profile of ISO 8601`_.
|
|
|
|
|
#: Extracted from the ``<meta name=dcterms.modified>`` element in HTML
|
|
|
|
|
#: and written to the ``/ModDate`` info field in PDF.
|
|
|
|
|
self.modified = modified
|
2014-04-04 14:32:21 +04:00
|
|
|
|
#: File attachments as a list of tuples of URL and a description or
|
|
|
|
|
#: :obj:`None`.
|
|
|
|
|
#: Extracted from the ``<link rel=attachment>`` elements in HTML
|
|
|
|
|
#: and written to the ``/EmbeddedFiles`` dictionary in PDF.
|
|
|
|
|
self.attachments = attachments or []
|
2013-07-14 15:08:02 +04:00
|
|
|
|
|
|
|
|
|
|
2012-10-02 20:59:02 +04:00
|
|
|
|
class Document(object):
|
2012-10-04 13:35:25 +04:00
|
|
|
|
"""A rendered document, with access to individual pages
|
|
|
|
|
ready to be painted on any cairo surfaces.
|
|
|
|
|
|
2013-07-14 15:08:02 +04:00
|
|
|
|
Typically obtained from :meth:`HTML.render() <weasyprint.HTML.render>`,
|
|
|
|
|
but can also be instantiated directly
|
2014-04-18 17:11:45 +04:00
|
|
|
|
with a list of :class:`pages <Page>`,
|
|
|
|
|
a set of :class:`metadata <DocumentMetadata>` and a ``url_fetcher``.
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
|
|
|
|
"""
|
2012-10-02 20:59:02 +04:00
|
|
|
|
@classmethod
|
2016-08-30 19:15:30 +03:00
|
|
|
|
def _render(cls, html, stylesheets, enable_hinting,
|
|
|
|
|
presentational_hints=False):
|
2016-10-27 18:36:24 +03:00
|
|
|
|
font_config = FontConfiguration()
|
2016-08-30 19:15:30 +03:00
|
|
|
|
style_for = get_all_computed_styles(
|
|
|
|
|
html, presentational_hints=presentational_hints, user_stylesheets=[
|
|
|
|
|
css if hasattr(css, 'rules')
|
|
|
|
|
else CSS(guess=css, media_type=html.media_type)
|
2016-10-27 12:41:34 +03:00
|
|
|
|
for css in stylesheets or []],
|
|
|
|
|
font_config=font_config)
|
2012-10-05 20:50:40 +04:00
|
|
|
|
get_image_from_uri = functools.partial(
|
2012-10-02 20:59:02 +04:00
|
|
|
|
images.get_image_from_uri, {}, html.url_fetcher)
|
|
|
|
|
page_boxes = layout_document(
|
|
|
|
|
enable_hinting, style_for, get_image_from_uri,
|
|
|
|
|
build_formatting_structure(
|
2016-10-27 18:36:24 +03:00
|
|
|
|
html.root_element, style_for, get_image_from_uri),
|
|
|
|
|
font_config)
|
2016-10-27 12:41:34 +03:00
|
|
|
|
rendering = cls(
|
|
|
|
|
[Page(p, enable_hinting) for p in page_boxes],
|
|
|
|
|
DocumentMetadata(**html._get_metadata()), html.url_fetcher)
|
2016-10-27 18:36:24 +03:00
|
|
|
|
font_config.clean()
|
2016-10-27 12:41:34 +03:00
|
|
|
|
return rendering
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2014-04-18 17:11:45 +04:00
|
|
|
|
def __init__(self, pages, metadata, url_fetcher):
|
2012-10-02 20:59:02 +04:00
|
|
|
|
#: A list of :class:`Page` objects.
|
|
|
|
|
self.pages = pages
|
2013-07-14 15:08:02 +04:00
|
|
|
|
#: A :class:`DocumentMetadata` object.
|
|
|
|
|
#: Contains information that does not belong to a specific page
|
|
|
|
|
#: but to the whole document.
|
|
|
|
|
self.metadata = metadata
|
2014-04-18 17:11:45 +04:00
|
|
|
|
#: A ``url_fetcher`` for resources that have to be read when writing
|
|
|
|
|
#: the output.
|
|
|
|
|
self.url_fetcher = url_fetcher
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
|
|
|
|
def copy(self, pages='all'):
|
2012-10-05 20:50:40 +04:00
|
|
|
|
"""Take a subset of the pages.
|
|
|
|
|
|
|
|
|
|
:param pages:
|
|
|
|
|
An iterable of :class:`Page` objects from :attr:`pages`.
|
|
|
|
|
:return:
|
|
|
|
|
A new :class:`Document` object.
|
|
|
|
|
|
2013-07-14 12:17:40 +04:00
|
|
|
|
Examples:
|
2012-10-05 20:50:40 +04:00
|
|
|
|
|
2013-07-14 12:17:40 +04:00
|
|
|
|
Write two PDF files for odd-numbered and even-numbered pages::
|
|
|
|
|
|
|
|
|
|
# Python lists count from 0 but pages are numbered from 1.
|
2012-10-05 20:50:40 +04:00
|
|
|
|
# [::2] is a slice of even list indexes but odd-numbered pages.
|
|
|
|
|
document.copy(document.pages[::2]).write_pdf('odd_pages.pdf')
|
|
|
|
|
document.copy(document.pages[1::2]).write_pdf('even_pages.pdf')
|
|
|
|
|
|
2013-07-14 12:17:40 +04:00
|
|
|
|
Write each page to a numbred PNG file::
|
|
|
|
|
|
2012-10-05 20:50:40 +04:00
|
|
|
|
for i, page in enumerate(document.pages):
|
|
|
|
|
document.copy(page).write_png('page_%s.png' % i)
|
|
|
|
|
|
2013-07-14 12:17:40 +04:00
|
|
|
|
Combine multiple documents into one PDF file,
|
|
|
|
|
using metadata from the first::
|
|
|
|
|
|
|
|
|
|
all_pages = [p for p in doc.pages for doc in documents]
|
|
|
|
|
documents[0].copy(all_pages).write_pdf('combined.pdf')
|
|
|
|
|
|
2012-10-05 20:50:40 +04:00
|
|
|
|
"""
|
2012-10-02 20:59:02 +04:00
|
|
|
|
if pages == 'all':
|
|
|
|
|
pages = self.pages
|
2012-10-05 20:50:40 +04:00
|
|
|
|
elif not isinstance(pages, list):
|
|
|
|
|
pages = list(pages)
|
2014-04-18 17:11:45 +04:00
|
|
|
|
return type(self)(pages, self.metadata, self.url_fetcher)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2012-10-04 13:35:25 +04:00
|
|
|
|
def resolve_links(self):
|
|
|
|
|
"""Resolve internal hyperlinks.
|
|
|
|
|
|
|
|
|
|
Links to a missing anchor are removed with a warning.
|
|
|
|
|
If multiple anchors have the same name, the first is used.
|
|
|
|
|
|
|
|
|
|
:returns:
|
|
|
|
|
A generator yielding lists (one per page) like :attr:`Page.links`,
|
|
|
|
|
except that :obj:`target` for internal hyperlinks is
|
|
|
|
|
``(page_number, x, y)`` instead of an anchor name.
|
|
|
|
|
The page number is an index (0-based) in the :attr:`pages` list,
|
2012-10-08 21:51:18 +04:00
|
|
|
|
``x, y`` are in CSS pixels from the top-left of the page.
|
2012-10-04 13:35:25 +04:00
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
anchors = {}
|
|
|
|
|
for i, page in enumerate(self.pages):
|
|
|
|
|
for anchor_name, (point_x, point_y) in iteritems(page.anchors):
|
|
|
|
|
anchors.setdefault(anchor_name, (i, point_x, point_y))
|
|
|
|
|
for page in self.pages:
|
|
|
|
|
page_links = []
|
|
|
|
|
for link in page.links:
|
|
|
|
|
link_type, anchor_name, rectangle = link
|
|
|
|
|
if link_type == 'internal':
|
|
|
|
|
target = anchors.get(anchor_name)
|
|
|
|
|
if target is None:
|
2013-08-19 16:38:09 +04:00
|
|
|
|
LOGGER.warning(
|
2012-10-04 13:35:25 +04:00
|
|
|
|
'No anchor #%s for internal URI reference '
|
2017-03-28 10:32:50 +03:00
|
|
|
|
'at line %s', anchor_name, link.sourceline)
|
2012-10-04 13:35:25 +04:00
|
|
|
|
else:
|
|
|
|
|
page_links.append((link_type, target, rectangle))
|
|
|
|
|
else:
|
|
|
|
|
# External link
|
|
|
|
|
page_links.append(link)
|
|
|
|
|
yield page_links
|
|
|
|
|
|
|
|
|
|
def make_bookmark_tree(self):
|
|
|
|
|
"""Make a tree of all bookmarks in the document.
|
|
|
|
|
|
|
|
|
|
:return: a list of bookmark subtrees.
|
|
|
|
|
A subtree is ``(label, target, children)``. :obj:`label` is
|
|
|
|
|
a string, :obj:`target` is ``(page_number, x, y)`` like in
|
|
|
|
|
:meth:`resolve_links`, and :obj:`children` is itself a (recursive)
|
|
|
|
|
list of subtrees.
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
root = []
|
|
|
|
|
# At one point in the document, for each "output" depth, how much
|
|
|
|
|
# to add to get the source level (CSS values of bookmark-level).
|
|
|
|
|
# Eg. with <h1> then <h3>, level_shifts == [0, 1]
|
|
|
|
|
# 1 means that <h3> has depth 3 - 1 = 2 in the output.
|
|
|
|
|
skipped_levels = []
|
|
|
|
|
last_by_depth = [root]
|
|
|
|
|
previous_level = 0
|
|
|
|
|
for page_number, page in enumerate(self.pages):
|
|
|
|
|
for level, label, (point_x, point_y) in page.bookmarks:
|
|
|
|
|
if level > previous_level:
|
|
|
|
|
# Example: if the previous bookmark is a <h2>, the next
|
|
|
|
|
# depth "should" be for <h3>. If now we get a <h6> we’re
|
|
|
|
|
# skipping two levels: append 6 - 3 - 1 = 2
|
|
|
|
|
skipped_levels.append(level - previous_level - 1)
|
|
|
|
|
else:
|
|
|
|
|
temp = level
|
|
|
|
|
while temp < previous_level:
|
|
|
|
|
temp += 1 + skipped_levels.pop()
|
|
|
|
|
if temp > previous_level:
|
|
|
|
|
# We remove too many "skips", add some back:
|
|
|
|
|
skipped_levels.append(temp - previous_level - 1)
|
|
|
|
|
|
|
|
|
|
previous_level = level
|
|
|
|
|
depth = level - sum(skipped_levels)
|
|
|
|
|
assert depth == len(skipped_levels)
|
|
|
|
|
assert depth >= 1
|
|
|
|
|
|
|
|
|
|
children = []
|
|
|
|
|
subtree = label, (page_number, point_x, point_y), children
|
|
|
|
|
last_by_depth[depth - 1].append(subtree)
|
|
|
|
|
del last_by_depth[depth:]
|
|
|
|
|
last_by_depth.append(children)
|
|
|
|
|
return root
|
|
|
|
|
|
2014-04-22 22:40:46 +04:00
|
|
|
|
def write_pdf(self, target=None, zoom=1, attachments=None):
|
2012-10-05 20:50:40 +04:00
|
|
|
|
"""Paint the pages in a PDF file, with meta-data.
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2012-10-05 20:50:40 +04:00
|
|
|
|
PDF files written directly by cairo do not have meta-data such as
|
|
|
|
|
bookmarks/outlines and hyperlinks.
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2012-10-05 20:50:40 +04:00
|
|
|
|
:param target:
|
2012-10-08 21:51:18 +04:00
|
|
|
|
A filename, file-like object, or :obj:`None`.
|
2012-11-23 01:27:34 +04:00
|
|
|
|
:type zoom: float
|
|
|
|
|
:param zoom:
|
|
|
|
|
The zoom factor in PDF units per CSS units.
|
|
|
|
|
**Warning**: All CSS units (even physical, like ``cm``)
|
|
|
|
|
are affected.
|
|
|
|
|
For values other than 1, physical CSS units will thus be “wrong”.
|
|
|
|
|
Page size declarations are affected too, even with keyword values
|
|
|
|
|
like ``@page { size: A3 landscape; }``
|
2014-04-22 22:40:46 +04:00
|
|
|
|
:param attachments: A list of additional file attachments for the
|
2014-04-26 01:35:43 +04:00
|
|
|
|
generated PDF document or :obj:`None`. The list's elements are
|
|
|
|
|
:class:`Attachment` objects, filenames, URLs or file-like objects.
|
2012-10-05 20:50:40 +04:00
|
|
|
|
:returns:
|
2012-10-08 21:51:18 +04:00
|
|
|
|
The PDF as byte string if :obj:`target` is :obj:`None`, otherwise
|
|
|
|
|
:obj:`None` (the PDF is written to :obj:`target`.)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
|
|
|
|
"""
|
2012-11-23 01:27:34 +04:00
|
|
|
|
# 0.75 = 72 PDF point (cairo units) per inch / 96 CSS pixel per inch
|
|
|
|
|
scale = zoom * 0.75
|
2012-10-02 20:59:02 +04:00
|
|
|
|
# Use an in-memory buffer. We will need to seek for metadata
|
|
|
|
|
# TODO: avoid this if target can seek? Benchmark first.
|
|
|
|
|
file_obj = io.BytesIO()
|
|
|
|
|
# (1, 1) is overridden by .set_size() below.
|
|
|
|
|
surface = cairo.PDFSurface(file_obj, 1, 1)
|
|
|
|
|
context = cairo.Context(surface)
|
|
|
|
|
for page in self.pages:
|
2015-04-29 10:32:33 +03:00
|
|
|
|
surface.set_size(
|
|
|
|
|
math.floor(page.width * scale),
|
|
|
|
|
math.floor(page.height * scale))
|
2012-11-22 21:47:20 +04:00
|
|
|
|
page.paint(context, scale=scale)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
surface.show_page()
|
|
|
|
|
surface.finish()
|
|
|
|
|
|
2014-04-22 22:40:46 +04:00
|
|
|
|
write_pdf_metadata(self, file_obj, scale, self.metadata, attachments,
|
2014-04-27 21:16:14 +04:00
|
|
|
|
self.url_fetcher)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
|
|
|
|
if target is None:
|
|
|
|
|
return file_obj.getvalue()
|
|
|
|
|
else:
|
|
|
|
|
file_obj.seek(0)
|
|
|
|
|
if hasattr(target, 'write'):
|
|
|
|
|
shutil.copyfileobj(file_obj, target)
|
|
|
|
|
else:
|
|
|
|
|
with open(target, 'wb') as fd:
|
|
|
|
|
shutil.copyfileobj(file_obj, fd)
|
|
|
|
|
|
2012-12-29 04:00:30 +04:00
|
|
|
|
def write_image_surface(self, resolution=96):
|
2012-10-05 22:12:05 +04:00
|
|
|
|
dppx = resolution / 96
|
|
|
|
|
|
2012-10-02 20:59:02 +04:00
|
|
|
|
# This duplicates the hinting logic in Page.paint. There is a
|
|
|
|
|
# dependency cycle otherwise:
|
|
|
|
|
# this → hinting logic → context → surface → this
|
|
|
|
|
# But since we do no transform here, cairo_context.user_to_device and
|
|
|
|
|
# friends are identity functions.
|
2012-10-05 22:12:05 +04:00
|
|
|
|
widths = [int(math.ceil(p.width * dppx)) for p in self.pages]
|
|
|
|
|
heights = [int(math.ceil(p.height * dppx)) for p in self.pages]
|
|
|
|
|
|
2012-10-02 20:59:02 +04:00
|
|
|
|
max_width = max(widths)
|
|
|
|
|
sum_heights = sum(heights)
|
2012-10-05 22:12:05 +04:00
|
|
|
|
surface = cairo.ImageSurface(
|
|
|
|
|
cairo.FORMAT_ARGB32, max_width, sum_heights)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
context = cairo.Context(surface)
|
|
|
|
|
pos_y = 0
|
|
|
|
|
for page, width, height in izip(self.pages, widths, heights):
|
|
|
|
|
pos_x = (max_width - width) / 2
|
2012-10-05 22:12:05 +04:00
|
|
|
|
page.paint(context, pos_x, pos_y, scale=dppx, clip=True)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
pos_y += height
|
2012-12-29 04:00:30 +04:00
|
|
|
|
return surface, max_width, sum_heights
|
2012-10-02 20:59:02 +04:00
|
|
|
|
|
2012-12-29 04:00:30 +04:00
|
|
|
|
def write_png(self, target=None, resolution=96):
|
|
|
|
|
"""Paint the pages vertically to a single PNG image.
|
|
|
|
|
|
|
|
|
|
There is no decoration around pages other than those specified in CSS
|
|
|
|
|
with ``@page`` rules. The final image is as wide as the widest page.
|
|
|
|
|
Each page is below the previous one, centered horizontally.
|
|
|
|
|
|
|
|
|
|
:param target:
|
|
|
|
|
A filename, file-like object, or :obj:`None`.
|
|
|
|
|
:type resolution: float
|
|
|
|
|
:param resolution:
|
|
|
|
|
The output resolution in PNG pixels per CSS inch. At 96 dpi
|
|
|
|
|
(the default), PNG pixels match the CSS ``px`` unit.
|
|
|
|
|
:returns:
|
|
|
|
|
A ``(png_bytes, png_width, png_height)`` tuple. :obj:`png_bytes`
|
|
|
|
|
is a byte string if :obj:`target` is :obj:`None`, otherwise
|
|
|
|
|
:obj:`None` (the image is written to :obj:`target`.)
|
|
|
|
|
:obj:`png_width` and :obj:`png_height` are the size of the
|
|
|
|
|
final image, in PNG pixels.
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
surface, max_width, sum_heights = self.write_image_surface(resolution)
|
2012-10-02 20:59:02 +04:00
|
|
|
|
if target is None:
|
|
|
|
|
target = io.BytesIO()
|
|
|
|
|
surface.write_to_png(target)
|
|
|
|
|
png_bytes = target.getvalue()
|
|
|
|
|
else:
|
|
|
|
|
surface.write_to_png(target)
|
|
|
|
|
png_bytes = None
|
2012-10-05 20:19:17 +04:00
|
|
|
|
return png_bytes, max_width, sum_heights
|