Support Markdown-rendered entries

blog/admin.py

@@ -3,7 +3,6 @@
 from django.db.models.functions import Length
 from django.db.models import F
 from django import forms
-from xml.etree import ElementTree
 from .models import (
     Entry,
     Tag,
@@ -44,10 +43,13 @@ class MyEntryForm(forms.ModelForm):
     def clean_body(self):
         # Ensure this is valid XML
         body = self.cleaned_data["body"]
-        try:
-            ElementTree.fromstring("<entry>%s</entry>" % body)
-        except Exception as e:
-            raise forms.ValidationError(str(e))
+        if not self.cleaned_data.get("use_markdown"):
+            from xml.etree import ElementTree
+
+            try:
+                ElementTree.fromstring("<entry>%s</entry>" % body)
+            except Exception as e:
+                raise forms.ValidationError(str(e))
         return body
 
 

blog/feeds.py

@@ -1,6 +1,7 @@
 from django.contrib.syndication.views import Feed
 from django.utils.dateformat import format as date_format
 from django.utils.feedgenerator import Atom1Feed
+from django.utils.safestring import mark_safe
 from django.http import HttpResponse
 from blog.models import Entry, Blogmark, Quotation, Note
 
@@ -63,7 +64,7 @@ def item_description(self, item):
             'Subscribe to <a href="https://simonwillison.net/atom/everything/">/atom/everything/</a> '
             'to get all of my posts, or take a look at my <a href="https://simonwillison.net/about/#subscribe">other subscription options</a>.</em></p>'
         )
-        return item.body + note
+        return mark_safe(str(item.rendered) + note)
 
 
 class Blogmarks(Base):

blog/management/commands/validate_entries_xml.py

@@ -1,5 +1,4 @@
 from django.core.management.base import BaseCommand
-from xml.etree import ElementTree
 from blog.models import Entry
 
 
@@ -8,7 +7,11 @@ class Command(BaseCommand):
 
     def handle(self, *args, **kwargs):
         for entry in Entry.objects.all():
+            if entry.use_markdown:
+                continue
             try:
+                from xml.etree import ElementTree
+
                 ElementTree.fromstring("<entry>%s</entry>" % entry.body.encode("utf8"))
             except Exception as e:
                 print(e)

blog/migrations/0030_entry_use_markdown.py

@@ -0,0 +1,21 @@
+# Generated by Django 5.1.4 on 2025-10-04 19:07
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("blog", "0029_blogmark_title"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="entry",
+            name="use_markdown",
+            field=models.BooleanField(
+                default=False,
+                help_text='Images can use the img element - set width="..." for a specific width and use class="blogmark-image" to center and add a drop shadow.',
+            ),
+        ),
+    ]

blog/models.py

@@ -14,7 +14,7 @@
 import arrow
 import datetime
 from markdown import markdown
-from xml.etree import ElementTree
+from bs4 import BeautifulSoup
 
 tag_re = re.compile("^[a-z0-9]+$")
 
@@ -224,6 +224,10 @@ class Meta:
 class Entry(BaseModel):
     title = models.CharField(max_length=255)
     body = models.TextField()
+    use_markdown = models.BooleanField(
+        default=False,
+        help_text='Images can use the img element - set width="..." for a specific width and use class="blogmark-image" to center and add a drop shadow.',
+    )
     tweet_html = models.TextField(
         blank=True,
         null=True,
@@ -251,13 +255,13 @@ def previous_by_created(self):
 
     def images(self):
         """Extracts images from entry.body"""
-        et = ElementTree.fromstring("<entry>%s</entry>" % self.body)
-        return [i.attrib for i in et.findall(".//img")]
+        soup = BeautifulSoup(str(self.rendered), "html.parser")
+        return [dict(img.attrs) for img in soup.find_all("img")]
 
     def index_components(self):
         return {
             "A": self.title,
-            "C": strip_tags(self.body),
+            "C": strip_tags(str(self.rendered)),
             "B": " ".join(self.tags.values_list("tag", flat=True)),
         }
 
@@ -286,14 +290,22 @@ def series_info(self):
         }
 
     def multi_paragraph(self):
-        return self.body.count("<p") > 1
+        return str(self.rendered).count("<p") > 1
 
     def __str__(self):
         return self.title
 
     class Meta(BaseModel.Meta):
         verbose_name_plural = "Entries"
 
+    @property
+    def rendered(self):
+        if self.use_markdown:
+            rendered = markdown(self.body or "")
+        else:
+            rendered = self.body or ""
+        return mark_safe(rendered)
+
 
 class LiveUpdate(models.Model):
     created = models.DateTimeField(auto_now_add=True)

blog/templatetags/entry_tags.py

@@ -1,14 +1,11 @@
 from django import template
 from django.utils.html import conditional_escape
 from django.utils.safestring import mark_safe
-from xml.etree import ElementTree
+from bs4 import BeautifulSoup, NavigableString, Comment
 import re
 import datetime
 
 register = template.Library()
-entry_stripper = re.compile("^<entry>(.*?)</entry>$", re.DOTALL)
-
-
 @register.filter
 def xhtml(xhtml):
     return XhtmlString(xhtml, contains_markup=True)
@@ -17,32 +14,38 @@ def xhtml(xhtml):
 class XhtmlString(object):
     def __init__(self, value, contains_markup=False):
         if isinstance(value, XhtmlString):
-            self.et = value.et
+            self.soup = value.soup
         else:
+            if value is None:
+                value = ""
             if not contains_markup:
                 # Handle strings like "this & that"
                 value = conditional_escape(value)
-            self.et = ElementTree.fromstring("<entry>%s</entry>" % value)
+            self.soup = BeautifulSoup(f"<entry>{value}</entry>", "html.parser")
+        self.et = self.soup.find("entry")
 
     def __str__(self):
-        m = entry_stripper.match(ElementTree.tostring(self.et, "unicode"))
-        if m:
-            return mark_safe(m.group(1))
-        else:
-            return ""  # If we end up with <entry />
+        if not self.et:
+            return ""
+        return mark_safe("".join(str(content) for content in self.et.contents))
 
 
 @register.filter
 def resize_images_to_fit_width(value, arg):
     max_width = int(arg)
     x = XhtmlString(value)
-    for img in x.et.findall(".//img"):
-        width = int(img.get("width", 0))
-        height = int(img.get("height", 0))
-        if width > max_width:
+    if not x.et:
+        return x
+    for img in x.et.find_all("img"):
+        try:
+            width = int(img.get("width", 0))
+            height = int(img.get("height", 0))
+        except (TypeError, ValueError):
+            continue
+        if width > max_width and height:
             # Scale down
-            img.set("width", str(max_width))
-            img.set("height", str(int(float(max_width) / width * height)))
+            img["width"] = str(max_width)
+            img["height"] = str(int(float(max_width) / width * height))
     return x
 
 
@@ -64,23 +67,27 @@ def split_cutoff(xhtml):
 @register.filter
 def remove_context_paragraph(xhtml):
     x = XhtmlString(xhtml)
+    if not x.et:
+        return x
     p = x.et.find("p")
     if p is None:
         return x
-    xhtml = ElementTree.tostring(p, "unicode")
+    xhtml = str(p)
     if xhtml.startswith("<p><em>My answer to") or xhtml.startswith(
         '<p class="context">'
     ):
-        x.et.remove(p)
+        p.decompose()
     return x
 
 
 @register.filter
 def first_paragraph(xhtml):
     x = XhtmlString(xhtml)
+    if not x.et:
+        return mark_safe("<p>%s</p>" % xhtml)
     p = x.et.find("p")
     if p is not None:
-        return mark_safe(ElementTree.tostring(p, "unicode"))
+        return mark_safe(str(p))
     else:
         return mark_safe("<p>%s</p>" % xhtml)
 
@@ -104,9 +111,11 @@ def ends_with_punctuation(value):
 @register.filter
 def strip_p_ids(xhtml):
     x = XhtmlString(xhtml)
-    for p in x.et.findall(".//p"):
-        if "id" in p.attrib:
-            del p.attrib["id"]
+    if not x.et:
+        return x
+    for p in x.et.find_all("p"):
+        if "id" in p.attrs:
+            del p.attrs["id"]
     return x
 
 
@@ -120,18 +129,26 @@ def break_up_long_words(xhtml, length):
 
 
 def do_break_long_words(et, length):
-    """Pass an ElementTree instance; breaks up long words in it"""
-    if et.text:
-        et.text = do_break_long_words_string(et.text, length)
-    for child in et:
-        do_break_long_words(child, length)
-    if et.tail:
-        et.tail = do_break_long_words_string(et.tail, length)
+    """Pass a BeautifulSoup Tag instance; breaks up long words in it"""
+    if et is None:
+        return
+    for node in list(_iter_text_nodes(et)):
+        new_text = do_break_long_words_string(str(node), length)
+        if new_text != str(node):
+            node.replace_with(new_text)
 
 
 whitespace_re = re.compile(r"(\s+)")
 
 
+def _iter_text_nodes(tag):
+    if tag is None:
+        return []
+    for node in tag.descendants:
+        if isinstance(node, NavigableString) and not isinstance(node, Comment):
+            yield node
+
+
 def do_break_long_words_string(s, length):
     bits = whitespace_re.split(s)
     for i, bit in enumerate(bits):
@@ -167,15 +184,15 @@ def strip_wrapping_p(xhtml):
 
 
 def do_typography(et):
-    # Designed to be called recursively on ElementTree objects
-    if et.tag not in ("pre", "code"):
-        # Don't do et.text or children for those tags; just do et.tail
-        if et.text:
-            et.text = do_typography_string(et.text)
-        for child in et:
-            do_typography(child)
-    if et.tail:
-        et.tail = do_typography_string(et.tail)
+    if et is None:
+        return
+    for node in list(_iter_text_nodes(et)):
+        parent = node.parent
+        if parent and parent.name in ("pre", "code"):
+            continue
+        new_text = do_typography_string(str(node))
+        if new_text != str(node):
+            node.replace_with(new_text)
 
 
 LEFT_DOUBLE_QUOTATION_MARK = "\u201c"

blog/tests.py

@@ -113,6 +113,49 @@ def test_markup(self):
         """.strip(),
         )
 
+    def test_markdown_entry_renders_everywhere(self):
+        tag = Tag.objects.create(tag="markdown-test")
+        entry = EntryFactory(
+            title="Markdown Entry",
+            body="This is **markdown** text.",
+            use_markdown=True,
+        )
+        entry.tags.add(tag)
+
+        year = entry.created.year
+        month = entry.created.strftime("%b")
+        day = entry.created.day
+
+        paths = {
+            "/": True,
+            f"/{year}/": False,
+            f"/{year}/{month}/": True,
+            f"/{year}/{month}/{day}/": True,
+            entry.get_absolute_url(): True,
+            f"/tags/{tag.tag}/": True,
+            f"/search/?tag={tag.tag}": True,
+        }
+
+        for path, should_have_markup in paths.items():
+            response = self.client.get(path)
+            self.assertEqual(response.status_code, 200, msg=path)
+            if should_have_markup and "<strong>markdown</strong>" not in response.content.decode():
+                self.fail(f"{path} missing rendered markdown")
+
+        entries_feed = self.client.get("/atom/entries/")
+        self.assertEqual(entries_feed.status_code, 200)
+        self.assertIn(
+            "&lt;strong&gt;markdown&lt;/strong&gt;",
+            entries_feed.content.decode(),
+        )
+
+        everything_feed = self.client.get("/atom/everything/")
+        self.assertEqual(everything_feed.status_code, 200)
+        self.assertIn(
+            "&lt;strong&gt;markdown&lt;/strong&gt;",
+            everything_feed.content.decode(),
+        )
+
     def test_update_blogmark_runs_commit_hooks(self):
         # This was throwing errors on upgrade Django 2.2 to 2.2.1
         blogmark = BlogmarkFactory()

templates/entry.html

@@ -9,7 +9,7 @@
 {% endblock %}
 
 {% block card_title %}{{ entry.title|typography }}{% endblock %}
-{% block card_description %}{{ entry.body|xhtml|remove_context_paragraph|typography|xhtml2html|striptags|truncatewords:30|force_escape }}{% endblock %}
+{% block card_description %}{{ entry.rendered|xhtml|remove_context_paragraph|typography|xhtml2html|striptags|truncatewords:30|force_escape }}{% endblock %}
 
 {% block item_content %}
 <div data-permalink-context="{{ entry.get_absolute_url }}">
@@ -18,7 +18,7 @@ <h2>{{ entry.title|typography }}</h2>
 
 {% include "_draft_warning.html" %}
 
-{{ entry.body|xhtml|resize_images_to_fit_width:"450"|typography|xhtml2html }}
+{{ entry.rendered|xhtml|resize_images_to_fit_width:"450"|typography|xhtml2html }}
 
 {% if updates %}
 <div id="live-updates">

templates/feeds/everything.html

@@ -1,5 +1,5 @@
 {% if obj.is_entry %}
-    {{ obj.body|safe }}
+    {{ obj.rendered|safe }}
     {% if obj.tags.count %}
         <p>Tags: {% for tag in obj.tags.all %}<a href="https://simonwillison.net/tags/{{ tag.tag }}">{{ tag.tag }}</a>{% if not forloop.last %}, {% endif %}{% endfor %}</p>
     {% endif %}