mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-11-16 20:53:27 +00:00
21d8b6288d
* Refactoring in preparation for AMP support * infrastructure for AMP validation * Add Metrika counter on AMP page + adjust layout * more content fixes * improve amp layout * improve navigation * Move converting admonitions to generation time * strict amp test check * Batch AMP validation * Add date published/modified to docs articles and their ld+json meta * few more content fixes * improve ld+json meta * adjust margins * skip published/modified for stable release docs * adapt single page mode * update po * skip published/modified for single page docs * update po * adjust layout * adjust layout
201 lines
8.3 KiB
Python
201 lines
8.3 KiB
Python
import logging
|
|
import os
|
|
import re
|
|
import shutil
|
|
import subprocess
|
|
import yaml
|
|
|
|
import bs4
|
|
import mkdocs.commands.build
|
|
|
|
import test
|
|
import util
|
|
import website
|
|
|
|
|
|
def recursive_values(item):
|
|
if isinstance(item, dict):
|
|
for _, value in item.items():
|
|
yield from recursive_values(value)
|
|
elif isinstance(item, list):
|
|
for value in item:
|
|
yield from recursive_values(value)
|
|
elif isinstance(item, str):
|
|
yield item
|
|
|
|
|
|
def concatenate(lang, docs_path, single_page_file, nav):
|
|
lang_path = os.path.join(docs_path, lang)
|
|
az_re = re.compile(r'[a-z]')
|
|
|
|
proj_config = f'{docs_path}/toc_{lang}.yml'
|
|
if os.path.exists(proj_config):
|
|
with open(proj_config) as cfg_file:
|
|
nav = yaml.full_load(cfg_file.read())['nav']
|
|
files_to_concatenate = list(recursive_values(nav))
|
|
files_count = len(files_to_concatenate)
|
|
logging.info(f'{files_count} files will be concatenated into single md-file for {lang}.')
|
|
logging.debug('Concatenating: ' + ', '.join(files_to_concatenate))
|
|
assert files_count > 0, f'Empty single-page for {lang}'
|
|
|
|
for path in files_to_concatenate:
|
|
if path.endswith('introduction/info.md'):
|
|
continue
|
|
try:
|
|
with open(os.path.join(lang_path, path)) as f:
|
|
anchors = set()
|
|
tmp_path = path.replace('/index.md', '/').replace('.md', '/')
|
|
prefixes = ['', '../', '../../', '../../../']
|
|
parts = tmp_path.split('/')
|
|
anchors.add(parts[-2] + '/')
|
|
anchors.add('/'.join(parts[1:]))
|
|
|
|
for part in parts[0:-2] if len(parts) > 2 else parts:
|
|
for prefix in prefixes:
|
|
anchor = prefix + tmp_path
|
|
if anchor:
|
|
anchors.add(anchor)
|
|
anchors.add('../' + anchor)
|
|
anchors.add('../../' + anchor)
|
|
tmp_path = tmp_path.replace(part, '..')
|
|
|
|
for anchor in anchors:
|
|
if re.search(az_re, anchor):
|
|
single_page_file.write('<a name="%s"></a>' % anchor)
|
|
|
|
single_page_file.write('\n')
|
|
|
|
in_metadata = False
|
|
for l in f:
|
|
if l.startswith('---'):
|
|
in_metadata = not in_metadata
|
|
if l.startswith('#'):
|
|
l = '#' + l
|
|
if not in_metadata:
|
|
single_page_file.write(l)
|
|
except IOError as e:
|
|
logging.warning(str(e))
|
|
|
|
single_page_file.flush()
|
|
|
|
|
|
def build_single_page_version(lang, args, nav, cfg):
|
|
logging.info(f'Building single page version for {lang}')
|
|
os.environ['SINGLE_PAGE'] = '1'
|
|
extra = cfg.data['extra']
|
|
extra['single_page'] = True
|
|
extra['is_amp'] = False
|
|
|
|
with util.autoremoved_file(os.path.join(args.docs_dir, lang, 'single.md')) as single_md:
|
|
concatenate(lang, args.docs_dir, single_md, nav)
|
|
|
|
with util.temp_dir() as site_temp:
|
|
with util.temp_dir() as docs_temp:
|
|
docs_src_lang = os.path.join(args.docs_dir, lang)
|
|
docs_temp_lang = os.path.join(docs_temp, lang)
|
|
shutil.copytree(docs_src_lang, docs_temp_lang)
|
|
for root, _, filenames in os.walk(docs_temp_lang):
|
|
for filename in filenames:
|
|
if filename != 'single.md' and filename.endswith('.md'):
|
|
os.unlink(os.path.join(root, filename))
|
|
|
|
cfg.load_dict({
|
|
'docs_dir': docs_temp_lang,
|
|
'site_dir': site_temp,
|
|
'extra': extra,
|
|
'nav': [
|
|
{cfg.data.get('site_name'): 'single.md'}
|
|
]
|
|
})
|
|
|
|
if not args.test_only:
|
|
mkdocs.commands.build.build(cfg)
|
|
|
|
if args.version_prefix:
|
|
single_page_output_path = os.path.join(args.docs_dir, args.docs_output_dir, args.version_prefix, lang, 'single')
|
|
else:
|
|
single_page_output_path = os.path.join(args.docs_dir, args.docs_output_dir, lang, 'single')
|
|
|
|
if os.path.exists(single_page_output_path):
|
|
shutil.rmtree(single_page_output_path)
|
|
|
|
shutil.copytree(
|
|
os.path.join(site_temp, 'single'),
|
|
single_page_output_path
|
|
)
|
|
|
|
single_page_index_html = os.path.join(single_page_output_path, 'index.html')
|
|
single_page_content_js = os.path.join(single_page_output_path, 'content.js')
|
|
with open(single_page_index_html, 'r') as f:
|
|
sp_prefix, sp_js, sp_suffix = f.read().split('<!-- BREAK -->')
|
|
with open(single_page_index_html, 'w') as f:
|
|
f.write(sp_prefix)
|
|
f.write(sp_suffix)
|
|
with open(single_page_content_js, 'w') as f:
|
|
if args.minify:
|
|
import jsmin
|
|
sp_js = jsmin.jsmin(sp_js)
|
|
f.write(sp_js)
|
|
|
|
logging.info(f'Re-building single page for {lang} pdf/test')
|
|
with util.temp_dir() as test_dir:
|
|
extra['single_page'] = False
|
|
cfg.load_dict({
|
|
'docs_dir': docs_temp_lang,
|
|
'site_dir': test_dir,
|
|
'extra': extra,
|
|
'nav': [
|
|
{cfg.data.get('site_name'): 'single.md'}
|
|
]
|
|
})
|
|
mkdocs.commands.build.build(cfg)
|
|
|
|
css_in = ' '.join(website.get_css_in(args))
|
|
js_in = ' '.join(website.get_js_in(args))
|
|
subprocess.check_call(f'cat {css_in} > {test_dir}/css/base.css', shell=True)
|
|
subprocess.check_call(f'cat {js_in} > {test_dir}/js/base.js', shell=True)
|
|
if args.save_raw_single_page:
|
|
shutil.copytree(test_dir, args.save_raw_single_page)
|
|
|
|
if not args.version_prefix: # maybe enable in future
|
|
logging.info(f'Running tests for {lang}')
|
|
test.test_single_page(
|
|
os.path.join(test_dir, 'single', 'index.html'), lang)
|
|
|
|
if not args.skip_pdf:
|
|
single_page_index_html = os.path.join(test_dir, 'single', 'index.html')
|
|
single_page_pdf = os.path.abspath(
|
|
os.path.join(single_page_output_path, f'clickhouse_{lang}.pdf')
|
|
)
|
|
|
|
with open(single_page_index_html, 'r') as f:
|
|
soup = bs4.BeautifulSoup(
|
|
f.read(),
|
|
features='html.parser'
|
|
)
|
|
soup_prefix = f'file://{test_dir}'
|
|
for img in soup.findAll('img'):
|
|
if img['src'].startswith('/'):
|
|
img['src'] = soup_prefix + img['src']
|
|
for script in soup.findAll('script'):
|
|
script_src = script.get('src')
|
|
if script_src:
|
|
script['src'] = soup_prefix + script_src.split('?', 1)[0]
|
|
for link in soup.findAll('link'):
|
|
link['href'] = soup_prefix + link['href'].split('?', 1)[0]
|
|
|
|
with open(single_page_index_html, 'w') as f:
|
|
f.write(str(soup))
|
|
|
|
create_pdf_command = [
|
|
'wkhtmltopdf',
|
|
'--print-media-type',
|
|
'--log-level', 'warn',
|
|
single_page_index_html, single_page_pdf
|
|
]
|
|
|
|
logging.info(' '.join(create_pdf_command))
|
|
subprocess.check_call(' '.join(create_pdf_command), shell=True)
|
|
|
|
logging.info(f'Finished building single page version for {lang}')
|