Make empty actually empty

CHANGELOG.rst

@@ -5,6 +5,10 @@ Next version
 ~~~~~~~~~~~~
 
 - Allowed installing the package in Python 3.10 environments too.
+- Tweaked the cleaning methods of ``ProseEditorField`` and
+  ``SanitizedProseEditorField`` to produce empty strings when no content is
+  entered. Previously they would produce an empty paragraph (``<p></p>``) since
+  our ProseMirror schema says that there exists always one or more block nodes.
 
 
 0.3 (2024-04-09)

django_prose_editor/fields.py

@@ -1,3 +1,5 @@
+import re
+
 from django import forms
 from django.db import models
 from django.utils.html import strip_tags
@@ -6,13 +8,21 @@
 from django_prose_editor.widgets import ProseEditorWidget
 
 
-def _identity(x):
+def _actually_empty(x):
+    """
+    ProseMirror's schema always adds at least one empty paragraph
+
+    We want empty fields to actually be empty strings so that those field
+    values evaluate as ``False`` in a boolean context.
+    """
+    if re.match(r"^<(?P<tag>\w+)></(?P=tag)>$", x):
+        return ""
     return x
 
 
 class ProseEditorField(models.TextField):
     def __init__(self, *args, **kwargs):
-        self.sanitize = kwargs.pop("sanitize", _identity)
+        self.sanitize = kwargs.pop("sanitize", _actually_empty)
         self.config = kwargs.pop("config", {})
         super().__init__(*args, **kwargs)
 

django_prose_editor/sanitized.py

@@ -1,10 +1,14 @@
-from django_prose_editor.fields import ProseEditorField
+from django_prose_editor.fields import ProseEditorField, _actually_empty
+
+
+def _nh3_sanitizer():
+    from nh3 import clean
+
+    return lambda x: _actually_empty(clean(x))
 
 
 class SanitizedProseEditorField(ProseEditorField):
     def __init__(self, *args, **kwargs):
         if "sanitize" not in kwargs:
-            from nh3 import clean
-
-            kwargs["sanitize"] = clean
+            kwargs["sanitize"] = _nh3_sanitizer()
         super().__init__(*args, **kwargs)

tests/testapp/test_prose_editor.py

@@ -2,12 +2,33 @@
 from django.db import models
 from django.test.utils import isolate_apps
 
+from django_prose_editor.fields import ProseEditorField
 from django_prose_editor.sanitized import SanitizedProseEditorField
 
 
 class Test(test.TestCase):
     @isolate_apps()
-    def test_field(self):
+    def test_standard_field(self):
+        class Model(models.Model):
+            description = ProseEditorField()
+
+            def __str__(self):
+                return self.description
+
+        m = Model(description="<p></p>")
+        m.full_clean()
+        self.assertEqual(m.description, "")
+
+        m = Model(description="<h1></h1>")
+        m.full_clean()
+        self.assertEqual(m.description, "")
+
+        m = Model(description="<p>hello</p>")
+        m.full_clean()
+        self.assertEqual(m.description, "<p>hello</p>")
+
+    @isolate_apps()
+    def test_sanitized_field(self):
         class Model(models.Model):
             description = SanitizedProseEditorField()
 
@@ -17,3 +38,15 @@ def __str__(self):
         m = Model(description="<style>h1{color:red}</style><h1>Hello</h1>")
         m.full_clean()
         self.assertEqual(m.description, "<h1>Hello</h1>")
+
+        m = Model(description="<p></p>")
+        m.full_clean()
+        self.assertEqual(m.description, "")
+
+        m = Model(description="<h1></h1>")
+        m.full_clean()
+        self.assertEqual(m.description, "")
+
+        m = Model(description="<p>hello</p>")
+        m.full_clean()
+        self.assertEqual(m.description, "<p>hello</p>")