diff --git a/README.rst b/README.rst
index 059a68f..387c8da 100644
--- a/README.rst
+++ b/README.rst
@@ -154,7 +154,8 @@ strip_document
Controls whether leading and/or trailing separation newlines are removed from
the final converted document. Supported values are ``LSTRIP`` (leading),
``RSTRIP`` (trailing), ``STRIP`` (both), and ``None`` (neither). Newlines
- within the document are unaffected.
+ within the document are unaffected. This also applies when ``convert_soup()``
+ receives an individual BeautifulSoup ``Tag`` rather than a complete soup.
Defaults to ``STRIP``.
strip_pre
diff --git a/markdownify/__init__.py b/markdownify/__init__.py
index 28cdaf6..363c1d1 100644
--- a/markdownify/__init__.py
+++ b/markdownify/__init__.py
@@ -223,7 +223,14 @@ def convert(self, html):
return self.convert_soup(soup)
def convert_soup(self, soup):
- return self.process_tag(soup, parent_tags=set())
+ text = self.process_tag(soup, parent_tags=set())
+ # A supplied Tag is also a complete conversion input. BeautifulSoup's
+ # document node already runs this finalizer during process_tag().
+ if soup.name != '[document]':
+ convert_fn = self.get_conv_fn_cached('[document]')
+ if convert_fn is not None:
+ text = convert_fn(soup, text, parent_tags={'[document]'})
+ return text
def process_element(self, node, parent_tags=None):
if isinstance(node, NavigableString):
diff --git a/tests/test_custom_converter.py b/tests/test_custom_converter.py
index 00a83fc..976132b 100644
--- a/tests/test_custom_converter.py
+++ b/tests/test_custom_converter.py
@@ -42,3 +42,49 @@ def test_soup():
html = 'test'
soup = BeautifulSoup(html, 'html.parser')
assert MarkdownConverter().convert_soup(soup) == '**test**'
+
+
+def test_tag_document_stripping():
+ from markdownify import LSTRIP, RSTRIP, STRIP
+
+ html = ''
+ for mode, expected in [(LSTRIP, 'hello\n\nworld\n\n'),
+ (RSTRIP, '\n\nhello\n\nworld'),
+ (STRIP, 'hello\n\nworld'),
+ (None, '\n\nhello\n\nworld\n\n')]:
+ for name in ['html', 'body', 'div']:
+ soup = BeautifulSoup(html, 'html.parser')
+ tag = soup.find(name)
+ before = str(soup)
+ assert MarkdownConverter(strip_document=mode).convert_soup(tag) == expected
+ assert str(soup) == before
+ assert tag.parent is not None
+
+
+def test_tag_document_invalid_strip_mode():
+ import pytest
+
+ soup = BeautifulSoup('hello
', 'html.parser')
+ with pytest.raises(ValueError, match='Invalid value for strip_document'):
+ MarkdownConverter(strip_document='invalid').convert_soup(soup.p)
+
+
+def test_document_converter_called_once():
+ class CountingConverter(MarkdownConverter):
+ calls = 0
+
+ def convert__document_(self, el, text, parent_tags):
+ self.calls += 1
+ return super().convert__document_(el, text, parent_tags)
+
+ soup = BeautifulSoup('', 'html.parser')
+ for root in [soup, soup.div, soup.p]:
+ converter = CountingConverter()
+ assert converter.convert_soup(root) == 'hello'
+ assert converter.calls == 1
+
+
+def test_tag_document_converter_exclusion():
+ soup = BeautifulSoup('hello
', 'html.parser')
+ for root in [soup, soup.p]:
+ assert MarkdownConverter(strip=['[document]']).convert_soup(root) == '\n\nhello\n\n'