kotlin-website.py

import copy
import datetime
import json
import os
import sys
import threading
from hashlib import md5
from os import path
from urllib.parse import urlparse, urljoin, ParseResult

import yaml
from bs4 import BeautifulSoup
from flask import Flask, render_template, Response, send_from_directory, request
from flask.helpers import url_for, send_file, make_response
from flask.views import View
from flask_frozen import Freezer, walk_directory
from yaml import FullLoader

from src.Feature import Feature
from src.api import get_api_page
from src.encoder import DateAwareEncoder
from src.externals import process_nav_includes
from src.github import assert_valid_git_hub_url
from src.grammar import get_grammar
from src.ktl_components import KTLComponentExtension
from src.markdown.makrdown import jinja_aware_markdown
from src.navigation import process_nav, get_current_url
from src.pages.MyFlatPages import MyFlatPages
from src.pdf import generate_pdf
from src.processors.processors import process_code_blocks
from src.processors.processors import set_replace_simple_code

app = Flask(__name__, static_folder='_assets')
app.config.from_pyfile('mysettings.py')
app.jinja_env.trim_blocks = True
app.jinja_env.lstrip_blocks = True
pages = MyFlatPages(app)
freezer = Freezer(app)
ignore_stdlib = False
build_mode = False
build_contenteditable = False
build_check_links = True
build_errors = []
url_adapter = app.create_url_adapter(None)

root_folder = path.join(os.path.dirname(__file__))
data_folder = path.join(os.path.dirname(__file__), "data")

_nav_cache = None
_nav_lock = threading.RLock()

_cached_asset_version = {}


def get_asset_version(filename):
    if filename in _cached_asset_version:
        return _cached_asset_version[filename]

    filepath = (root_folder if root_folder else ".") + filename
    if filename and path.exists(filepath):
        with open(filepath, 'rb') as file:
            digest = md5(file.read()).hexdigest()
            _cached_asset_version[filename] = digest
            return digest
    return None


def redirect_to_map(redirects_list):
    result = {}

    for item in redirects_list:
        key = item["from"]
        result.update({
            f"{key}": item["to"]
        })

    return result


def get_site_data():
    data = {}
    for data_file in os.listdir(data_folder):
        if data_file.startswith('_'):
            continue
        if not data_file.endswith(".yml"):
            continue
        data_file_path = path.join(data_folder, data_file)
        with open(data_file_path, encoding="UTF-8") as stream:
            try:
                file_name_without_extension = data_file[:-4] if data_file.endswith(".yml") else data_file
                data[file_name_without_extension] = yaml.load(stream, Loader=FullLoader)
            except yaml.YAMLError as exc:
                sys.stderr.write('Cant parse data file ' + data_file + ': ')
                sys.stderr.write(str(exc))
                sys.exit(-1)
            except IOError as exc:
                sys.stderr.write('Cant read data file ' + data_file + ': ')
                sys.stderr.write(str(exc))
                sys.exit(-1)
    data["core"] = redirect_to_map(
        yaml.load(open("redirects/stdlib-redirects.yml", encoding="UTF-8"), Loader=FullLoader))
    return data


site_data = get_site_data()


def get_nav():
    global _nav_cache
    global _nav_lock

    with _nav_lock:
        if _nav_cache is not None:
            nav = _nav_cache
        else:
            nav = get_nav_impl()

        nav = copy.deepcopy(nav)

        if build_mode:
            _nav_cache = copy.deepcopy(nav)

    # NOTE. This call depends on `request.path`, cannot cache
    process_nav(request.path, nav)
    return nav


def get_countries_size():
    def match_string(entry):
        location = entry.get("location", "")
        # Extract the last part as the country code
        return location.split(",")[-1].strip()

    # Extract unique countries, ignoring any None results
    matches = set(filter(None, map(match_string, site_data['universities'])))
    return len(matches)


def get_education_courses():
    return [{attr: x[attr] for attr in ["title", "location", "courses"]}
            for x in site_data['universities']]


def get_nav_impl():
    with open(path.join(data_folder, "_nav.yml")) as stream:
        nav = yaml.load(stream, Loader=FullLoader)
        nav = process_nav_includes(build_mode, nav)
        return nav


def get_kotlin_features():
    features_dir = path.join(os.path.dirname(__file__), "kotlin-features")
    features = []
    for feature_meta in yaml.load(open(path.join(features_dir, "kotlin-features.yml")), Loader=FullLoader):
        file_path = path.join(features_dir, feature_meta['content_file'])
        with open(file_path, encoding='utf-8') as f:
            content = f.read()
            content = content.replace("\r\n", "\n")
            if file_path.endswith(".md"):
                html_content = BeautifulSoup(jinja_aware_markdown(content, pages), 'html.parser')
                content = process_code_blocks(html_content)
            features.append(Feature(content, feature_meta))
    return features


@app.context_processor
def add_year_to_context():
    return {
        'year': datetime.datetime.now().year
    }


app.jinja_env.add_extension(KTLComponentExtension)


@app.context_processor
def add_data_to_context():
    nav = get_nav()
    return {
        'nav': nav,
        'data': site_data,
        'site': {
            'pdf_url': app.config['PDF_URL'],
            'forum_url': app.config['FORUM_URL'],
            'site_github_url': app.config['SITE_GITHUB_URL'],
            'data': site_data,
            'text_using_gradle': app.config['TEXT_USING_GRADLE'],
            'code_baseurl': app.config['CODE_URL'],
            'contenteditable': build_contenteditable
        },
        'headerCurrentUrl': get_current_url(nav['subnav']['content']),
    }


@app.template_filter('get_domain')
def get_domain(url):
    return urlparse(url).netloc


app.jinja_env.globals['get_domain'] = get_domain


@app.template_filter('split_chunk')
def split_chunk(list, size):
    return [list[i:i + size] for i in range(len(list))[::size]]


app.jinja_env.globals['split_chunk'] = split_chunk


@app.template_filter('autoversion')
def autoversion_filter(filename):
    asset_version = get_asset_version(filename)
    if asset_version is None: return filename
    original = urlparse(filename)._asdict()
    original.update(query=original.get('query') + '&v=' + asset_version)
    return ParseResult(**original).geturl()


@app.route('/data/cities.json')
def get_cities():
    return Response(json.dumps(site_data['cities'], cls=DateAwareEncoder), mimetype='application/json')


@app.route('/data/kotlinconf.json')
def get_kotlinconf():
    return Response(json.dumps(site_data['kotlinconf'], cls=DateAwareEncoder), mimetype='application/json')


@app.route('/data/universities.json')
def get_universities():
    return Response(json.dumps(site_data['universities'], cls=DateAwareEncoder), mimetype='application/json')


@app.route('/docs/reference/grammar.html')
def grammar():
    grammar = get_grammar(build_mode)
    if grammar is None:
        return "Grammar file not found", 404
    return render_template('pages/grammar.html', kotlinGrammar=grammar)


@app.route('/docs/kotlin-reference.pdf')
def kotlin_reference_pdf():
    return send_file(path.join(root_folder, "assets", "kotlin-reference.pdf"))


@app.route('/docs/kotlin-docs.pdf')
def kotlin_docs_pdf():
    return send_file(path.join(root_folder, "assets", "kotlin-reference.pdf"))


@app.route('/docs/<path:path>')
def docs(path):
    return send_from_directory('dist/docs/', path)


@app.route('/_next/<path:path>')
def static_file(path):
    return send_from_directory('out/_next/', path)


@app.route('/community/')
def community_page():
    return send_file(path.join(root_folder, 'out', 'community/index.html'))


@app.route('/community/events/')
def community_events_page():
    return send_file(path.join(root_folder, 'out', 'community/events/index.html'))


@app.route('/community/user-groups/')
def community_user_groups_page():
    return send_file(path.join(root_folder, 'out', 'community/user-groups/index.html'))


@app.route('/education/')
def education_page():
    return render_template(
        'pages/education/index.html',
        universities_count=len(site_data['universities']),
        countries_count=get_countries_size()
    )


@app.route('/education/why-teach-kotlin.html')
def why_teach_page():
    return render_template('pages/education/why-teach-kotlin.html')


@app.route('/education/courses.html')
def education_courses():
    return render_template('pages/education/courses.html', universities_data=get_education_courses())


@app.route('/')
def next_index_page():
    return send_file(path.join(root_folder, 'out', 'index.html'))


def process_page(page_path):
    # get_nav() has side effect to copy and patch files from the `external` folder
    # under site folder. We need it for dev mode to make sure file is up-to-date
    # TODO: extract get_nav and implement the explicit way to avoid side-effects
    get_nav()

    page = pages.get_or_404(page_path)
    if 'redirect_path' in page.meta and page.meta['redirect_path'] is not None:
        page_path = page.meta['redirect_path']
        if page_path.startswith('https://') or page_path.startswith('http://'):
            return render_template('redirect.html', url=page_path)
        else:
            return render_template('redirect.html', url=url_for('page', page_path=page_path))

    if 'date' in page.meta and page['date'] is not None:
        page.meta['formatted_date'] = page.meta['date'].strftime('%d %B %Y')
        if page.meta['formatted_date'].startswith('0'):
            page.meta['formatted_date'] = page.meta['formatted_date'][1:]

    if 'github_edit_url' in page.meta:
        edit_on_github_url = page.meta['github_edit_url']
    else:
        edit_on_github_url = app.config['EDIT_ON_GITHUB_URL'] + app.config['FLATPAGES_ROOT'] + "/" + page_path + \
                             app.config['FLATPAGES_EXTENSION']

    assert_valid_git_hub_url(edit_on_github_url, page_path)

    template = page.meta["layout"] if 'layout' in page.meta else 'default.html'
    if not template.endswith(".html"):
        template += ".html"

    if build_check_links:
        validate_links_weak(page, page_path)

    return render_template(
        template,
        page=page,
        baseurl="",
        edit_on_github_url=edit_on_github_url,

    )


def validate_links_weak(page, page_path):
    for link in page.parsed_html.select('a'):
        if 'href' not in link.attrs:
            continue

        href = urlparse(urljoin('/' + page_path, link['href']))
        if href.scheme != '':
            continue

        endpoint, params = url_adapter.match(href.path, 'GET', query_args={})
        if endpoint != 'page' and endpoint != 'get_index_page':
            response = app.test_client().get(href.path)
            if response.status_code == 404:
                build_errors.append("Broken link: " + str(href.path) + " on page " + page_path)
            continue

        referenced_page = pages.get(params['page_path'])
        if referenced_page is None:
            build_errors.append("Broken link: " + str(href.path) + " on page " + page_path)
            continue

        if href.fragment == '':
            continue

        ids = []
        for x in referenced_page.parsed_html.select('h1,h2,h3,h4'):
            try:
                ids.append(x['id'])
            except KeyError:
                pass

        for x in referenced_page.parsed_html.select('a'):
            try:
                ids.append(x['name'])
            except KeyError:
                pass

        if href.fragment not in ids:
            build_errors.append("Bad anchor: " + str(href.fragment) + " on page " + page_path)

    if not build_mode and len(build_errors) > 0:
        errors_copy = []

        for item in build_errors:
            errors_copy.append(item)

        build_errors.clear()
        raise Exception("Validation errors " + str(len(errors_copy)) + ":\n\n" +
                        "\n".join(str(item) for item in errors_copy))


@freezer.register_generator
def page():
    for page in pages:
        yield {'page_path': page.path}


@app.route('/<path:page_path>.html')
def page(page_path):
    return process_page(page_path)


@app.route('/404.html')
def page_404():
    return send_file(path.join(root_folder, 'out', '404.html'))


@freezer.register_generator
def api_page():
    api_folder = path.join(root_folder, 'api')
    for root, dirs, files in os.walk(api_folder):
        for file in files:
            yield {'page_path': path.join(path.relpath(root, api_folder), file).replace(os.sep, '/')}


class RedirectTemplateView(View):
    def __init__(self, url):
        self.redirect_url = url

    def dispatch_request(self):
        return render_template('redirect.html', url=self.redirect_url)


def generate_redirect_pages():
    redirects_folder = path.join(root_folder, 'redirects')
    for root, dirs, files in os.walk(redirects_folder):
        for file in files:
            if not file.endswith(".yml"):
                continue

            redirects_file_path = path.join(redirects_folder, file)

            with open(redirects_file_path, encoding="UTF-8") as stream:
                try:
                    redirects = yaml.load(stream, Loader=FullLoader)

                    for entry in redirects:
                        url_to = entry["to"]
                        url_from = entry["from"]
                        url_list = url_from if isinstance(url_from, list) else [url_from]

                        for url in url_list:
                            if file == 'api.yml' and path.isfile(path.join(root_folder, url[1:])):
                                print("The file " + url + " is already exist.")
                            else:
                                app.add_url_rule(url, view_func=RedirectTemplateView.as_view(url, url=url_to))

                except yaml.YAMLError as exc:
                    sys.stderr.write('Cant parse data file ' + file + ': ')
                    sys.stderr.write(str(exc))
                    sys.exit(-1)
                except IOError as exc:
                    sys.stderr.write('Cant read data file ' + file + ': ')
                    sys.stderr.write(str(exc))
                    sys.exit(-1)


@app.errorhandler(404)
def page_not_found(e):
    return send_file(path.join(root_folder, 'out', '404.html')), 404


app.register_error_handler(404, page_not_found)


@app.route('/api/<path:page_path>')
def api_page(page_path):
    path_other, ext = path.splitext(page_path)
    if ext == '.html':
        return process_api_page(page_path[:-5])
    elif path.basename(page_path) == "package-list" or ext:
        return respond_with_package_list(page_path)
    elif not page_path.endswith('/'):
        page_path += '/'
    return process_api_page(page_path + 'index')


def process_api_page(page_path):
    return render_template(
        'api.html',
        page=get_api_page(build_mode, page_path)
    )


def respond_with_package_list(page_path):
    file_path = path.join(root_folder, 'api', page_path)
    if not path.exists(file_path):
        return make_response(path.basename(page_path) + " not found", 404)
    return send_file(file_path, mimetype="text/plain")


@app.route('/assets/<path:path>')
def asset(path):
    return send_from_directory('assets', path)


@app.route('/assets/images/tutorials/<path:filename>')
def tutorial_img(filename):
    return send_from_directory(path.join('assets', 'images', 'tutorials'), filename)


@freezer.register_generator
def asset():
    for filename in walk_directory(path.join(root_folder, "assets")):
        yield {'path': filename}


@app.route('/<path:page_path>')
def get_index_page(page_path):
    """
    Handle requests which urls don't end with '.html' (for example, '/doc/')

    We don't need any generator here, because such urls are equivalent to the same urls
    with 'index.html' at the end.

    :param page_path: str
    :return: str
    """
    if not page_path.endswith('/'):
        page_path += '/'
    return process_page(page_path + 'index')


generate_redirect_pages()


@app.after_request
def add_header(request):
    request.headers["Cache-Control"] = "no-cache, no-store, must-revalidate"
    request.headers["Pragma"] = "no-cache"
    request.headers["Expires"] = "0"
    request.headers['Cache-Control'] = 'public, max-age=0'
    return request


if __name__ == '__main__':
    print("\n\n\nRunning new KotlinWebSite generator/dev-mode:\n")

    argv_copy = []
    for arg in sys.argv:
        print("arg: " + arg)

        if arg == "--ignore-stdlib":
            ignore_stdlib = True
        elif arg == "--no-check-links":
            build_check_links = False
        elif arg == "--editable":
            build_contenteditable = True
        else:
            argv_copy.append(arg)

    print("\n\n")
    print("ignore_stdlib: " + str(ignore_stdlib))
    print("build_check_links: " + str(build_check_links))
    print("build_contenteditable: " + str(build_contenteditable))
    print("\n\n")

    set_replace_simple_code(build_contenteditable)

    with (open(path.join(root_folder, "_nav-mapped.yml"), 'w')) as output:
        yaml.dump(get_nav_impl(), output)

    if len(argv_copy) > 1:
        if argv_copy[1] == "build":
            build_mode = True

            urls = freezer.freeze()
            if len(build_errors) > 0:
                for error in build_errors:
                    sys.stderr.write(error + '\n')
                sys.exit(-1)

        elif argv_copy[1] == "reference-pdf":
            generate_pdf("kotlin-docs.pdf", site_data)
        else:
            print("Unknown argument: " + argv_copy[1])
            sys.exit(1)
    else:
        app.run(host="0.0.0.0", port=8080, debug=True, threaded=True, use_debugger=False, use_reloader=False)