nerv-project/sphinx_weasyprint_builder
8
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

Initial release

Browse source
This commit is contained in:
Kujiu 2020-06-19 02:21:35 +02:00
parent 441ec2418e
commit 309307f1ce
Signed by: kujiu
GPG key ID: ABBB2CAC6855599F
11 changed files with 284 additions and 130 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
CHANGES
View file

@ -5,4 +5,4 @@ Changes
0.1.0 (*2020-06-10*)
====================
- Initial release
- Initial release (no theme included...)

44
README.rst
View file

@ -49,8 +49,20 @@ for `html` builder.
text overflows the page.
You can use the sphinx_weasyprint_theme installed
with this plugin to start.
with this plugin to start (coming soon!).
In addition, you can set these three variables:
- weasyprint_main_selector
- weasyprint_footer_selector
- weasyprint_header_selector
This builder analyzes generated HTML to extract with
BeautifulSoup a footer and a header. They are removed
from main HTML and reinjected on each page, conforming
to theme's CSS. By default, no header and no footer.
Use compatible selectors for BeautifulSoup.
Use
---
@ -61,36 +73,6 @@ Just launch the following:
make weasyprint
Signing
-------
.. warning::
You need to install `pykcs11` and `endesive` to
sign PDF.
Following parameters are needed to sign:
.. code:: python
weasyprint_sign_method = 'pkcs11'
weasyprint_sign_certid = '0x..........'
weasyprint_sign_reason = 'My Company'
weasyprint_sign_location = 'Where I am'
weasyprint_sign_contact = "Yeah, it's me"
weasyprint_sign_image = 'path_to_image'
weasyprint_sign_text = __('Signed on {sign_date}\nat {sign_location}\nby {sign_contact}\nfor {sign_reason}') #default
weasyprint_sign_position = (page, x, y) #default: None
.. note::
Signing methods are really limited. Do not
hesitate to contribute !
Available: `none` (no signing), `pkcs11`
To be implemented: `p12`, `gpg`
Why an other PDF builder for Sphinx?
------------------------------------

33
build/lib/sphinx_weasyprint_builder/__init__.py
View file

@ -4,7 +4,6 @@ from .weasyprint_builder import WeasyPrintPDFBuilder
from typing import Dict, Any
from sphinx.application import Sphinx
from sphinx.util.osutil import make_filename
from sphinx.locale import __
version = (0, 1, 0)
@ -65,47 +64,21 @@ def setup(app: Sphinx) -> Dict[str, Any]:
True,
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_method',
'none',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_certid',
'weasyprint_main_selector',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_reason',
'weasyprint_footer_selector',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_contact',
'weasyprint_header_selector',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_location',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_image',
None,
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_position',
None,
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_text',
__('Signed on {sign_date}\nat {sign_location}\nby {sign_contact}\nfor {sign_reason}'),
'weasyprint'
)
return {
'version': version,

15
build/lib/sphinx_weasyprint_builder/pdf_signing.py
View file

@ -1,15 +0,0 @@
#!/usr/bin/env python3
def sign_pdf_pkcs11(filename, certid, reason, contact, location, image, text):
pass
def sign_pdf_p12(filename, certid, reason, contact, location, image, text):
pass
def sign_pdf_gpg(filename, certid, reason, contact, location, image, text):
pass

135
build/lib/sphinx_weasyprint_builder/weasyprint_builder.py
View file

@ -2,19 +2,142 @@
import os
from typing import Dict, Set, Tuple
from copy import deepcopy
from sphinx.builders.singlehtml import SingleFileHTMLBuilder
from sphinx.util import progress_message
from sphinx.util.osutil import os_path
from sphinx.locale import __
from sphinx.util import logging
from bs4 import BeautifulSoup
import weasyprint
from .loghandler import init_wpsphinx_log
logger = logging.getLogger('weasyprint')
init_wpsphinx_log()
def extract(soup, selector: str) -> None:
elements = soup.select(selector)
for element in elements:
element.extract()
class PdfGenerator:
"""
From WeasyPrint 47 documentation Tips & Tricks
"""
def __init__(
self, base_url: str,
main_selector: str,
footer_selector: str,
header_selector: str
) -> None:
self.base_url = base_url
with open(self.base_url) as htmlfile:
self.main_html = htmlfile.read()
self.footer_html = None
self.header_html = None
if not (footer_selector or header_selector):
return
if not main_selector:
logger.error(
__('You must define a selector for content if you set selector for footer or header.'))
return
main_soup = BeautifulSoup(self.main_html, 'html.parser')
if footer_selector:
footer_soup = deepcopy(main_soup)
if header_selector:
extract(footer_soup, header_selector)
extract(footer_soup, main_selector)
if header_selector:
header_soup = deepcopy(main_soup)
if footer_selector:
extract(header_soup, footer_selector)
extract(header_soup, main_selector)
extract(main_soup, header_selector)
if footer_selector:
extract(main_soup, footer_selector)
self.main_html = str(main_soup)
if footer_soup:
self.footer_html = str(footer_soup)
if header_soup:
self.header_html = str(header_soup)
def _compute_overlay_element(self, element: str) -> weasyprint.HTML:
html = weasyprint.HTML(
string=getattr(self, f'{element}_html'),
base_url=self.base_url,
)
element_doc = html.render()
element_page = element_doc.pages[0]
element_body = PdfGenerator.get_element(
element_page._page_box.all_children(), 'body'
)
element_body = element_body.copy_with_children(
element_body.all_children()
)
return element_body
def _apply_overlay_on_main(
self, main_doc, header_body=None, footer_body=None
) -> None:
for page in main_doc.pages:
page_body = PdfGenerator.get_element(
page._page_box.all_children(), 'body'
)
if header_body:
page_body.children += header_body.all_children()
if footer_body:
page_body.children += footer_body.all_children()
def write_pdf(self, target: str) -> None:
if self.header_html:
header_body = self._compute_overlay_element(
'header'
)
else:
header_body = None
if self.footer_html:
footer_body = self._compute_overlay_element(
'footer'
)
else:
footer_body = None
html = weasyprint.HTML(
string=self.main_html,
base_url=self.base_url,
)
main_doc = html.render()
if self.header_html or self.footer_html:
self._apply_overlay_on_main(main_doc, header_body, footer_body)
main_doc.write_pdf(target)
@staticmethod
def get_element(boxes, element):
for box in boxes:
if box.element_tag == element:
return box
return PdfGenerator.get_element(box.all_children(), element)
class WeasyPrintPDFBuilder(SingleFileHTMLBuilder):
name = 'weasyprint'
epilog = __('The PDF file has been saved in %(outdir)s.')
@ -57,7 +180,8 @@ class WeasyPrintPDFBuilder(SingleFileHTMLBuilder):
self.globalcontext['use_opensearch'] = False
self.globalcontext['docstitle'] = self.config.weasyprint_title
self.globalcontext['shorttitle'] = self.config.weasyprint_short_title
self.globalcontext['show_copyright'] = self.config.weasyprint_show_copyright
self.globalcontext['show_copyright'] = \
self.config.weasyprint_show_copyright
self.globalcontext['show_sphinx'] = self.config.weasyprint_show_sphinx
self.globalcontext['style'] = stylename
self.globalcontext['favicon'] = None
@ -73,6 +197,9 @@ class WeasyPrintPDFBuilder(SingleFileHTMLBuilder):
self.outdir,
self.config.weasyprint_basename + '.pdf'
)
weasy_html = weasyprint.HTML(infile)
weasy_html.write_pdf(outfile)
# progress_message('Signing PDF')
generator = PdfGenerator(
infile,
self.config.weasyprint_main_selector,
self.config.weasyprint_footer_selector,
self.config.weasyprint_header_selector)
generator.write_pdf(outfile)

1
requirements.txt
View file

@ -2,3 +2,4 @@ Sphinx>=3.0.0
WeasyPrint
endesive
pykcs11
bs4

1
setup.py
View file

@ -50,6 +50,7 @@ setup(
install_requires=[
"sphinx>=3.0.0",
"WeasyPrint",
"bs4",
],
classifiers=[
"Framework :: Sphinx",

BIN
sphinx_weasyprint_builder/.weasyprint_builder.py.swp
View file

Binary file not shown.

33
sphinx_weasyprint_builder/__init__.py
View file

@ -4,7 +4,6 @@ from .weasyprint_builder import WeasyPrintPDFBuilder
from typing import Dict, Any
from sphinx.application import Sphinx
from sphinx.util.osutil import make_filename
from sphinx.locale import __
version = (0, 1, 0)
@ -65,47 +64,21 @@ def setup(app: Sphinx) -> Dict[str, Any]:
True,
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_method',
'none',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_certid',
'weasyprint_main_selector',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_reason',
'weasyprint_footer_selector',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_contact',
'weasyprint_header_selector',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_location',
'',
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_image',
None,
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_position',
None,
'weasyprint'
)
app.add_config_value(
'weasyprint_sign_text',
__('Signed on {sign_date}\nat {sign_location}\nby {sign_contact}\nfor {sign_reason}'),
'weasyprint'
)
return {
'version': version,

15
sphinx_weasyprint_builder/pdf_signing.py
View file

@ -1,15 +0,0 @@
#!/usr/bin/env python3
def sign_pdf_pkcs11(filename, certid, reason, contact, location, image, text):
pass
def sign_pdf_p12(filename, certid, reason, contact, location, image, text):
pass
def sign_pdf_gpg(filename, certid, reason, contact, location, image, text):
pass

135
sphinx_weasyprint_builder/weasyprint_builder.py
View file

@ -2,19 +2,142 @@
import os
from typing import Dict, Set, Tuple
from copy import deepcopy
from sphinx.builders.singlehtml import SingleFileHTMLBuilder
from sphinx.util import progress_message
from sphinx.util.osutil import os_path
from sphinx.locale import __
from sphinx.util import logging
from bs4 import BeautifulSoup
import weasyprint
from .loghandler import init_wpsphinx_log
logger = logging.getLogger('weasyprint')
init_wpsphinx_log()
def extract(soup, selector: str) -> None:
elements = soup.select(selector)
for element in elements:
element.extract()
class PdfGenerator:
"""
From WeasyPrint 47 documentation Tips & Tricks
"""
def __init__(
self, base_url: str,
main_selector: str,
footer_selector: str,
header_selector: str
) -> None:
self.base_url = base_url
with open(self.base_url) as htmlfile:
self.main_html = htmlfile.read()
self.footer_html = None
self.header_html = None
if not (footer_selector or header_selector):
return
if not main_selector:
logger.error(
__('You must define a selector for content if you set selector for footer or header.'))
return
main_soup = BeautifulSoup(self.main_html, 'html.parser')
if footer_selector:
footer_soup = deepcopy(main_soup)
if header_selector:
extract(footer_soup, header_selector)
extract(footer_soup, main_selector)
if header_selector:
header_soup = deepcopy(main_soup)
if footer_selector:
extract(header_soup, footer_selector)
extract(header_soup, main_selector)
extract(main_soup, header_selector)
if footer_selector:
extract(main_soup, footer_selector)
self.main_html = str(main_soup)
if footer_soup:
self.footer_html = str(footer_soup)
if header_soup:
self.header_html = str(header_soup)
def _compute_overlay_element(self, element: str) -> weasyprint.HTML:
html = weasyprint.HTML(
string=getattr(self, f'{element}_html'),
base_url=self.base_url,
)
element_doc = html.render()
element_page = element_doc.pages[0]
element_body = PdfGenerator.get_element(
element_page._page_box.all_children(), 'body'
)
element_body = element_body.copy_with_children(
element_body.all_children()
)
return element_body
def _apply_overlay_on_main(
self, main_doc, header_body=None, footer_body=None
) -> None:
for page in main_doc.pages:
page_body = PdfGenerator.get_element(
page._page_box.all_children(), 'body'
)
if header_body:
page_body.children += header_body.all_children()
if footer_body:
page_body.children += footer_body.all_children()
def write_pdf(self, target: str) -> None:
if self.header_html:
header_body = self._compute_overlay_element(
'header'
)
else:
header_body = None
if self.footer_html:
footer_body = self._compute_overlay_element(
'footer'
)
else:
footer_body = None
html = weasyprint.HTML(
string=self.main_html,
base_url=self.base_url,
)
main_doc = html.render()
if self.header_html or self.footer_html:
self._apply_overlay_on_main(main_doc, header_body, footer_body)
main_doc.write_pdf(target)
@staticmethod
def get_element(boxes, element):
for box in boxes:
if box.element_tag == element:
return box
return PdfGenerator.get_element(box.all_children(), element)
class WeasyPrintPDFBuilder(SingleFileHTMLBuilder):
name = 'weasyprint'
epilog = __('The PDF file has been saved in %(outdir)s.')
@ -57,7 +180,8 @@ class WeasyPrintPDFBuilder(SingleFileHTMLBuilder):
self.globalcontext['use_opensearch'] = False
self.globalcontext['docstitle'] = self.config.weasyprint_title
self.globalcontext['shorttitle'] = self.config.weasyprint_short_title
self.globalcontext['show_copyright'] = self.config.weasyprint_show_copyright
self.globalcontext['show_copyright'] = \
self.config.weasyprint_show_copyright
self.globalcontext['show_sphinx'] = self.config.weasyprint_show_sphinx
self.globalcontext['style'] = stylename
self.globalcontext['favicon'] = None
@ -73,6 +197,9 @@ class WeasyPrintPDFBuilder(SingleFileHTMLBuilder):
self.outdir,
self.config.weasyprint_basename + '.pdf'
)
weasy_html = weasyprint.HTML(infile)
weasy_html.write_pdf(outfile)
# progress_message('Signing PDF')
generator = PdfGenerator(
infile,
self.config.weasyprint_main_selector,
self.config.weasyprint_footer_selector,
self.config.weasyprint_header_selector)
generator.write_pdf(outfile)