making line break in some rare cases

html_telegraph_poster/html_to_telegraph.py

@@ -233,7 +233,10 @@ def preprocess_fragments(fragments):
     body = fragments[0].getparent()
 
     # remove para inside blockquote/aside/figure  (telegraph removes it anyway) and replace with line-break
-    paras_inside_quote = body.xpath('.//*[self::blockquote|self::aside|self::figure]//p[text()][following-sibling::*[text()]]')
+    paras_inside_quote = body.xpath(
+        './/*[self::blockquote|self::aside|self::figure]//'
+        'p[descendant-or-self::*[text()]][following-sibling::*[text()]]'
+    )
     for para in paras_inside_quote:
         para.tail = '\n'
 

setup.py

@@ -4,7 +4,7 @@
     long_description = fh.read()
 
 setup(name='html_telegraph_poster',
-      version='0.2.21',
+      version='0.2.22',
       description='Posts your html to telegra.ph blogging service',
       long_description=long_description,
       long_description_content_type="text/markdown",

tests/test_htp.py

@@ -279,7 +279,6 @@ def test_br_tags(self):
         )
 
     def test_iframe(self):
-        # multiple br tags should be replaced with one line break
         html = '<iframe src="//www.youtube.com/embed/abcdef">legacy text</iframe>'
         iframe_empty_src = '<iframe src=""></iframe>'
         iframe_no_src = '<iframe></iframe>'
@@ -474,9 +473,14 @@ def test_blockquote(self):
     def test_bad_para(self):
         html = '<aside><p>text inside para</p><p>another para</p></aside>'
         html2 = '<figure><figcaption><p>text inside para</p><p>another para</p></figcaption></figure>'
+        figcaption_para_with_link = '<figure><figcaption>' \
+                                    '<p><a href="https://telegram.org/">Telegram</a></p><p>Text after link</p>' \
+                                    '</figcaption></figure>'
         html3 = '<blockquote><p></p><p>second para</p></blockquote>'
         html4 = '<blockquote><p>first para</p><strong><em></em></strong><em>em text</em></blockquote>'
         html5 = '<blockquote><p>first para</p><em></em><strong></strong><em></em></blockquote>'
+        # TODO: write html6 test
+        html6 = '''<blockquote class="cut">\n<p>text inside</p>\n</blockquote>'''
         self.assertJson(
             [{'children': ['text inside para\nanother para'], 'tag': 'aside'}],
             convert_html_to_telegraph_format(html, clean_html=True)
@@ -497,6 +501,10 @@ def test_bad_para(self):
             [{'children': [u'first para'], 'tag': 'blockquote'}],
             convert_html_to_telegraph_format(html5, clean_html=True)
         )
+        self.assertJson(
+            [{"tag": "figure", "children": [{"tag": "figcaption", "children": ["Telegram\nText after link"]}]}],
+            convert_html_to_telegraph_format(figcaption_para_with_link, clean_html=True)
+        )
 
     def test_convert_without_clean(self):
         # multiple br tags should be replaced with one line break