view piecrust/importing/jekyll.py @ 1188:a7c43131d871

bake: Fix file write flushing problem with Python 3.8+ Writing the cache files fails in Python 3.8 because it looks like flushing behaviour has changed. We need to explicitly flush. And even then, in very rare occurrences, it looks like it can still run into racing conditions, so we do a very hacky and ugly "retry" loop when fetching cached data :(
author Ludovic Chabant <ludovic@chabant.com>
date Tue, 15 Jun 2021 22:36:23 -0700
parents 727110ea112a
children
line wrap: on
line source

import os
import os.path
import re
import shutil
import yaml
import logging
from piecrust.configuration import parse_config_header
from piecrust.importing.base import FileWalkingImporter
from piecrust.uriutil import multi_replace


logger = logging.getLogger(__name__)


class JekyllImporter(FileWalkingImporter):
    name = 'jekyll'
    description = "Imports content from a Jekyll or Octopress blog."

    def setupParser(self, parser, app):
        super(JekyllImporter, self).setupParser(parser, app)
        parser.add_argument('root_dir',
                help="The root directory of the Jekyll or Octopress website.")

    def importWebsite(self, app, args):
        logger.debug("Importing Jekyll site from: %s" % args.root_dir)
        self._startWalk(args.root_dir, args.exclude, app)
        logger.info("The Jekyll website was successfully imported.")

    def _importFile(self, full_fn, rel_fn, app):
        logger.debug("- %s" % rel_fn)
        if rel_fn == '_config.yml':
            self.convertConfig(app, full_fn)
        elif rel_fn.startswith('_layouts'):
            self.convertLayout(app, full_fn, rel_fn[len('_layouts/'):])
        elif rel_fn.startswith('_includes'):
            self.convertInclude(app, full_fn, rel_fn[len('_includes/'):])
        elif rel_fn.startswith('_posts'):
            self.convertPost(app, full_fn, rel_fn[len('_posts/'):])
        else:
            with open(full_fn, 'rb') as fp:
                firstline = fp.read(3)
            if firstline == '---':
                self.convertPage(app, full_fn, rel_fn)
            else:
                self.convertStatic(app, full_fn, rel_fn)

    def convertConfig(self, app, src_path):
        logger.debug("  Converting configuration file.")
        with open(src_path, 'r', encoding='utf8') as fp:
            config = yaml.safe_load(fp)

        if 'site' not in config:
            config['site'] = {}
        config['site']['related_posts'] = []
        config['site']['posts_fs'] = 'flat'
        config['site']['templates_dirs'] = ['includes', 'layouts']
        config['site']['tag_url'] = 'tags/%tag%'
        if 'permalink' in config:
            permalink = config['permalink']
            if permalink == 'date':
                permalink = '/:categories/:year/:month/:day/:title.html'
            elif permalink == 'pretty':
                permalink = '/:categories/:year/:month/:day/:title/'
            elif permalink == 'none':
                permalink = '/:categories/:title.html'

            # TODO: handle `:categories` token.
            post_url = multi_replace(
                    permalink,
                    {':year': '%year%', ':month': '%month%', ':day': '%day%',
                        ':title': '%slug%', ':categories': ''})
            post_url = post_url.replace('//', '/').strip('/')
            config['site']['post_url'] = post_url
        if 'exclude' in config:
            if 'baker' not in config:
                config['baker'] = {}
            config['baker']['ignore'] = list(map(
                    lambda i: '^/_%s/' % re.escape(i)))
        if 'jinja' not in config:
            config['jinja'] = {}
        config['jinja']['auto_escape'] = False
        if 'markdown' in config:
            if not isinstance(config['markdown'], dict):
                logger.warning("Discarding markdown setting: %s" %
                        config['markdown'])
                del config['markdown']

        with open(os.path.join(app.root_dir, 'config.yml'), 'w') as fp:
            yaml.dump(config, stream=fp)

    def convertPage(self, app, path, rel_path):
        logger.debug("  Converting page: %s" % rel_path)
        is_index = False
        is_static = False
        _, ext = os.path.splitext(rel_path)
        if re.search(r'^index\.(html?|textile|markdown)$', rel_path):
            out_path = os.path.join(app.root_dir, 'pages', '_index' + ext)
            is_index = True
        else:
            out_path = os.path.join(app.root_dir, 'pages', rel_path)

        if ext not in ['htm', 'html', 'textile', 'markdown']:
            # There could be static files (SCSS or Less files) that look like
            # pages because they have a YAML front matter.
            is_static = True
            out_path = os.path.join(app.root_dir, 'assets', rel_path)

        if is_static:
            logger.debug("  Actually a static file... forwarding converstion.")
            self.convertStatic(app, path, rel_path, True)
            return

        self._doConvertPage(app, path, out_path)
        if is_index:
            shutil.copy2(out_path, os.path.join(app.root_dir, 'pages', '_tag.%s' % ext))

    def convertPost(self, app, path, rel_path):
        logger.debug("  Converting post: %s" % rel_path)
        out_path = re.sub(
                r'(\d{4}\-\d{2}\-\d{2})\-(.*)$',
                r'\1_\2',
                rel_path)
        out_path = os.path.join(app.root_dir, 'posts', out_path)
        self._doConvertPage(app, path, out_path)

    def convertLayout(self, app, path, rel_path):
        logger.debug("  Converting layout: %s" % rel_path)
        out_path = os.path.join(app.root_dir, 'layouts', rel_path)
        self._doConvertPage(app, path, out_path, True)

    def convertInclude(self, app, path, rel_path):
        logger.debug("  Converting include: %s" % rel_path)
        out_path = os.path.join(app.root_dir, 'includes', rel_path)
        self._doConvertPage(app, path, out_path, True)

    def convertStatic(self, app, path, rel_path, strip_header=False):
        logger.debug("  Converting static: %s" % rel_path)
        out_path = os.path.join(app.root_dir, 'assets', rel_path)
        logger.debug("  %s -> %s" % (path, out_path))
        os.makedirs(os.path.dirname(out_path), 0o755, True)

        if strip_header:
            with open(path, 'r', encoding='utf8') as fp:
                content = fp.write()
            config, offset = parse_config_header(content)
            content = content[offset:]
            with open(out_path, 'w', encoding='utf8') as fp:
                fp.write(content)
            return

        shutil.copy2(path, out_path)

    def _doConvertPage(self, app, path, out_path, is_template=False):
        logger.debug("  %s -> %s" % (path, out_path))
        os.makedirs(os.path.dirname(out_path), 0o755, True)

        with open(path, 'r', encoding='utf8') as fp:
            contents = fp.read()

        config, offset = parse_config_header(contents)
        text = contents[offset:]
        text_before = text

        wrap_content_tag = True

        if is_template:
            if 'layout' in config:
                # Liquid doesn't support template inheritance but
                # Jinja does.
                text = ("{%% extends '%s.html' %%}\n\n"
                        "{%% block jekyllcontent %%}\n"
                        "%s\n"
                        "{%% endblock %%}\n" % (config['layout'], text))
                wrap_content_tag = False
        else:
            if 'layout' in config:
                if config['layout'] == 'nil':
                    config['layout'] = 'none'

        # Convert the template stuff we can:
        # - content tag may have to be wrapped in a `jekyllcontent`
        #   because Jekyll uses implicit layout inheritance
        #   placements.
        if wrap_content_tag:
            text = re.sub(
                    r'{{\s*content\s*}}',
                    r'{% block jekyllcontent %}{{ content }}{% endblock %}',
                    text)
        # - list of posts
        text = re.sub(
            '(?<=\{%|\{\{)([^\}]*)site.posts',
            '\\1blog.posts',
            text);
        text = re.sub(
            '(?<=\{%|\{\{)([^\}]*)paginator.posts',
            '\\1pagination.posts',
            text);
        # - list of categories or tags
        text = re.sub(
            '(?<=\{%|\{\{)([^\}]*)site.categories',
            '\\1blog.categories',
            text);
        text = re.sub(
            '(?<=\{%|\{\{)([^\}]*)site.tags',
            '\\1blog.tags',
            text);
        # - list of related posts
        text = re.sub(
            '(?<=\{%|\{\{)(?<!%\})site.related_posts',
            '\\1pagination.related_posts',
            text);
        # - enumeration limits
        text = re.sub(
            '{%\s*for\s+([^}]+)\s+limit\:\s*(\d+)',
            '{% for \\1[:\\2]',
            text);
        text = re.sub(
            '{%\s*for\s+([^}]+)\s+offset\:\s*(\d+)',
            '{% for \\1[\\2:]',
            text);
        # - code highlighting
        text = re.sub(
            '{%\s*highlight\s+([\w\d]+)\s*%}',
            '{% geshi \'\\1\' %}',
            text);
        text = re.sub(
            '{%\s*endhighlight\s*%}',
            '{% endgeshi %}',
            text);
        # - unless tag
        text = re.sub(
            '{%\s*unless\s+([^}]+)\s*%}',
            '{% if not \\1 %}',
            text);
        text = re.sub(
            '{%\s*endunless\s*%}',
            '{% endif %}',
            text);
        # - variable assignment
        text = re.sub(
            '\{%\s*assign\s+',
            '{% set ',
            text);
        # - include tag
        text = re.sub(
            '\{%\s*include\s+([\w\d\.\-_]+)\s*%}',
            '{% include "\\1" %}',
            text);
        # - truncate filter
        text = re.sub(
            '\|\s*truncate\:\s*(\d+)',
            '|truncate(\\1)',
            text);
        # - date filter
        text = re.sub(
            '\|\s*date\:\s*"([^"]+)"',
            '|date("\\1")',
            text);
        # - some filters we don't need
        text = re.sub(
            '\|\s*date_to_string',
            '',
            text);

        if text != text_before:
            # We changed the text, so create a backup.
            shutil.copy2(path, '%s.orig' % out_path)

        with open(out_path, 'w', encoding='utf8') as fp:
            if not is_template:
                fp.write("---\n")
                fp.write(yaml.dump(config))
                fp.write("---\n")
            fp.write(text)