docs/build-multilang.py
ErickSkrauch 5cebcf8abe
Completely upgrade project:
- removed all translations and used Russian as the base language
- integrated crowdin
- new multilang site layout
- upgraded sphinx
- fixed search, now it correctly handles articles language
2021-03-20 03:27:22 +01:00

170 lines
5.7 KiB
Python
Executable File

#!/usr/bin/env python
import os
import re
import shutil
import subprocess
import sys
from glob import glob
from pathlib import Path
from lxml import html, etree
base_lang: str = "ru"
build_dir: str = "build"
per_lang_static: list[str] = [
r"documentation_options\.js",
r"language_data\.js",
r"^(?!basic)\w+-stemmer\.js",
r"translations\.js",
]
# Cleanup
shutil.rmtree(build_dir, True)
os.mkdir(build_dir)
# Build languages
def sphinx_build(lang: str) -> None:
subprocess.run([
sys.executable, "-m", "sphinx",
"-b", "html",
"-D", f"language={lang}",
"source",
f"{build_dir}/{lang}",
])
languages: list[str] = [base_lang]
if os.path.isdir("locale"):
languages += os.listdir("locale")
index_page_lang: str = "en" if "en" in languages else base_lang
for lang in languages:
sphinx_build(lang)
# Extract common assets
for i, lang in enumerate(languages):
if i == 0:
os.rename(f"{build_dir}/{lang}/_downloads", f"{build_dir}/_downloads")
os.rename(f"{build_dir}/{lang}/_images", f"{build_dir}/_images")
os.rename(f"{build_dir}/{lang}/CNAME", f"{build_dir}/CNAME")
os.rename(f"{build_dir}/{lang}/.nojekyll", f"{build_dir}/.nojekyll")
else:
shutil.rmtree(f"{build_dir}/{lang}/_downloads")
shutil.rmtree(f"{build_dir}/{lang}/_images")
os.remove(f"{build_dir}/{lang}/CNAME")
os.remove(f"{build_dir}/{lang}/.nojekyll")
for static_file in glob(f"{build_dir}/{lang}/_static/**/*", recursive=True):
if not os.path.isfile(static_file):
continue
relative_path = static_file.removeprefix(f"{build_dir}/{lang}/_static/")
matched: bool = False
for pattern in per_lang_static:
if re.match(pattern, relative_path):
matched = True
break
# I have no idea how to make "continue 2" from the loop above so...
if matched:
continue
dir = os.path.dirname(relative_path)
os.makedirs(f"{build_dir}/_static/{dir}", exist_ok=True)
os.rename(static_file, f"{build_dir}/_static/{relative_path}")
shutil.rmtree(f"{build_dir}/{lang}/_sources")
shutil.rmtree(f"{build_dir}/{lang}/.doctrees")
os.remove(f"{build_dir}/{lang}/.buildinfo")
os.remove(f"{build_dir}/{lang}/objects.inv")
# Convert all relative routes into absolute
def make_links_absolute(file_path: str, link: str) -> str:
# Skip local anchor links
if link.startswith("#"):
return link
# Skip external links (including links without protocol "//ely.by")
if link.startswith("//") or re.match(r"^https?://", link):
return link
if ".html" in link:
working_directory = os.getcwd()
os.chdir(Path(file_path).parent.absolute())
resolved = str(Path(link).resolve())
os.chdir(working_directory)
link = resolved.removeprefix(f"{working_directory}/{build_dir}")
link = link.replace("\\", "/") # fix for the Windows
return link
# Other links are links to some static assets
# There is no need to resolve relative links since _static is placed in the root directory,
# so after removing all ../ parts we can safely append / to make the path absolute
while link.startswith("../"):
link = link.removeprefix("../")
no_static_link = link.removeprefix("_static/")
for pattern in per_lang_static:
if re.match(pattern, no_static_link):
lang = re.match(fr"^{build_dir}/(\w+)/", file_path).group(1)
return f"/{lang}/{link}"
return f"/{link}"
for file in glob(f"{build_dir}/**/*.html", recursive=True):
tree = html.parse(file) # type: etree._ElementTree
root = tree.getroot() # type: html.HtmlElement
root.rewrite_links(lambda link: make_links_absolute(file, link))
tree.write(file, method="html", encoding="UTF-8")
# Create index.html for the site root
shutil.copyfile(f"{build_dir}/{index_page_lang}/index.html", f"{build_dir}/index.html")
index_file_tree = html.parse(f"{build_dir}/index.html") # type: etree._ElementTree
index_file_root = index_file_tree.getroot() # type: html.HtmlElement
index_file_last_toctree = index_file_root.find_class("toctree-wrapper")[0] # type: html.HtmlElement
for lang in languages:
if lang == index_page_lang:
continue
tree = html.parse(f"{build_dir}/{lang}/index.html") # type: etree._ElementTree
root = tree.getroot() # type: html.HtmlElement
toctree = root.find_class("toctree-wrapper")[0] # type: html.HtmlElement
index_file_last_toctree.addnext(toctree)
index_file_last_toctree = index_file_root.find_class("toctree-wrapper")[-1] # type: html.HtmlElement
index_file_tree.write(f"{build_dir}/index.html", method="html", encoding="UTF-8")
# Add cross-lang links to the sidebar
sidebar_menus: dict[str, html.HtmlElement] = {}
for lang in languages:
tree = html.parse(f"{build_dir}/{lang}/index.html") # type: etree._ElementTree
root = tree.getroot() # type: html.HtmlElement
sidebar_menus[lang] = root.find_class("wy-menu")[0]
for file in glob(f"{build_dir}/**/*.html") + [f"{build_dir}/index.html"]:
result = re.match(fr"^{build_dir}/(\w+)/", file)
lang: str = result.group(1) if result is not None else index_page_lang
tree = html.parse(file) # type: etree._ElementTree
root = tree.getroot() # type: html.HtmlElement
sidebar_menu_last = root.find_class("wy-menu")[0] # type: html.HtmlElement
for menuLang in sidebar_menus:
if menuLang == lang:
continue
sidebar_menu_last.addnext(sidebar_menus[menuLang])
sidebar_menu_last = root.find_class("wy-menu")[-1] # type: html.HtmlElement
tree.write(file, method="html", encoding="UTF-8")