write generator

1 files changed, 193 insertions, 0 deletions

diff --git a/build.py b/build.py
new file mode 100644
index 0000000..7d71b9f
--- /dev/null
+++ b/build.py
@@ -0,0 +1,193 @@
+from dataclasses import dataclass
+from datetime import date, datetime
+from pathlib import Path
+import re
+import shutil
+import typing
+
+from jinja2 import Environment, FileSystemLoader, Markup, StrictUndefined, Template, select_autoescape
+from md2gemini import md2gemini
+import mistune
+import mistune_contrib.highlight
+
+POST_DATE = re.compile(r'(\d{4}-\d{2}-\d{2})-(.*)\.md')
+POST_FRONT_MATTER = re.compile(r'---\n(.*?)---\n\n?(.*)', re.DOTALL)
+
+html_template = Template('')
+gmi_template = Template('')
+
+
+@dataclass
+class Page:
+    title: str
+    description: typing.Optional[str]
+    excerpt: str
+    url: str
+    date: typing.Optional[date]
+
+    raw_content: str
+    html_content: str
+    gmi_content: str
+
+    @staticmethod
+    def load(path: Path, *, template_context: typing.Optional[dict] = None):
+        date_match = POST_DATE.fullmatch(path.name)
+        if date_match is None:
+            slug = path.stem
+            dest = Path(slug)
+            page_date = None
+        else:
+            post_date, slug = date_match.groups()
+            page_date = date.fromisoformat(post_date)
+            dest = Path(page_date.strftime('%Y/%m/%d')) / slug
+
+        file_data = path.read_text()
+        front_matter = POST_FRONT_MATTER.fullmatch(file_data)
+        if front_matter is not None:
+            front_matter, content = POST_FRONT_MATTER.fullmatch(file_data).groups()
+            front_matter = parse_front_matter(front_matter)
+        else:
+            front_matter = {'title': slug}
+            content = file_data
+        title = front_matter['title']
+        description = front_matter.get('description', None)
+        excerpt = get_excerpt(content)
+        if 'permalink' in front_matter:
+            permalink = front_matter['permalink'].lstrip('/')
+            if permalink.endswith('/'):
+                permalink = permalink + 'index'
+            dest = Path(permalink)
+
+        if template_context is not None:
+            content = Template(content)
+            content = content.render(**template_context)
+
+        html_content = markdown(content)
+        gmi_content = md2gemini(content, links='copy', md_links=True).replace('\r\n', '\n')
+
+        return Page(title, description, excerpt, str(dest).replace('\\', '/'), page_date,
+                    content, html_content, gmi_content)
+
+    def describe(self):
+        if self.description is not None:
+            return self.description
+        return self.excerpt
+
+
+class HighlightRenderer(mistune_contrib.highlight.HighlightMixin, mistune.HTMLRenderer):
+    options = {'inlinestyles': False, 'linenos': False}
+
+    def link(self, link, text=None, title=None):
+        if link.startswith('/') and link.endswith('.md'):
+            link = re.sub('.md$', '.html', link)
+        return super(HighlightRenderer, self).link(link, text, title)
+
+
+markdown = mistune.create_markdown(renderer=HighlightRenderer(), plugins=['strikethrough'])
+
+
+def render(site_dir, page: Page):
+    html_dest = (site_dir / 'html' / page.url).with_suffix('.html')
+    html_dest.parent.mkdir(parents=True, exist_ok=True)
+    html_dest.write_text(html_template.render(content=Markup(page.html_content), page=page))
+
+    gmi_dest = (site_dir / 'gmi' / page.url).with_suffix('.gmi')
+    gmi_dest.parent.mkdir(parents=True, exist_ok=True)
+    gmi_dest.write_text(gmi_template.render(content=page.gmi_content, page=page))
+
+
+def parse_front_matter(front_matter: str):
+    lines = front_matter.split('\n')
+    fields = [line.split(': ', 1) for line in lines if len(line.strip()) > 0]
+    return dict((key, value.strip('"')) for key, value in fields)
+
+
+def get_excerpt(body: str):
+    def flatten_ast(node: dict):
+        if not isinstance(node, dict):
+            return str(node)
+        elif node['type'] == 'text':
+            return node['text']
+        else:
+            return ''.join(flatten_ast(child) for child in node['children'])
+
+    ast = mistune.markdown(body, renderer='ast')
+    paragraphs = [node for node in ast if node['type'] == 'paragraph'] + [{'type': 'text', 'text': ''}]
+    first_paragraph = paragraphs[0]
+    excerpt = flatten_ast(first_paragraph)
+    excerpt = re.sub(r'\s+', ' ', excerpt)
+    return excerpt
+
+
+def copy_assets(src: Path, dest: Path):
+    dest_assets = dest / 'assets'
+    if dest_assets.exists():
+        for asset in dest_assets.iterdir():
+            asset.unlink()
+        dest_assets.rmdir()
+    shutil.copytree(src / 'assets', dest_assets)
+
+
+def main():
+    print('building...')
+    env = Environment(
+        loader=FileSystemLoader('_layouts'),
+        autoescape=select_autoescape(['html']),
+        undefined=StrictUndefined,
+    )
+    env.filters['absolute_url'] = lambda x: f"https://www.boringcactus.com/{x}"
+    global html_template
+    html_template = env.get_template('default.html')
+    global gmi_template
+    gmi_template = env.get_template('default.gmi')
+
+    source_dir = Path('.')
+    site_dir = Path('_site')
+    html_site_dir = site_dir / 'html'
+    gmi_site_dir = site_dir / 'gmi'
+
+    html_site_dir.mkdir(parents=True, exist_ok=True)
+    gmi_site_dir.mkdir(parents=True, exist_ok=True)
+
+    copy_assets(source_dir, html_site_dir)
+    copy_assets(source_dir, gmi_site_dir)
+
+    posts = []
+    for post_filename in (source_dir / '_posts').glob('*.md'):
+        print('  -', post_filename.name)
+        page = Page.load(post_filename)
+        render(site_dir, page)
+        posts.append(page)
+
+    posts.sort(key=lambda x: x.date, reverse=True)
+
+    for page_filename in source_dir.glob('*.md'):
+        print('  -', page_filename.name)
+        page = Page.load(page_filename, template_context=dict(posts=posts))
+        render(site_dir, page)
+
+    print('  - feed.xml')
+
+    feed = (source_dir / 'feed.xml').read_text()
+    feed = Template(feed, autoescape=True)
+
+    (html_site_dir / 'feed.xml').write_text(feed.render(
+        root='https://www.boringcactus.com',
+        mime='text/html',
+        ext='html',
+        now=datetime.utcnow(),
+        posts=posts,
+        content_attr='html_content',
+    ))
+    (gmi_site_dir / 'feed.xml').write_text(feed.render(
+        root='gemini://boringcactus.com',
+        mime='text/gemini',
+        ext='gmi',
+        now=datetime.utcnow(),
+        posts=posts,
+        content_attr='gmi_content',
+    ))
+
+
+if __name__ == '__main__':
+    main()