github: create(): convert links and images to markdown

fixes snarfed/bridgy#794

granary/github.py

@@ -141,6 +141,11 @@ class GitHub(source.Source):
   POST_ID_RE = re.compile('^[0-9]+$')
   # https://github.com/moby/moby/issues/679#issuecomment-18307522
   REPO_NAME_RE = re.compile('^[A-Za-z0-9_.-]+$')
+  # https://github.com/Alir3z4/html2text/blob/master/docs/usage.md#available-options
+  HTML2TEXT_OPTIONS = {
+    'ignore_images': False,
+    'ignore_links': False,
+  }
 
   def __init__(self, access_token=None):
     """Constructor.

granary/source.py

@@ -66,14 +66,19 @@ def __init__(self, *args, **kwargs):
     super(RateLimited, self).__init__(*args, **kwargs)
 
 
-def html_to_text(html):
-  """Converts string html to string text with html2text."""
+def html_to_text(html, **kwargs):
+  """Converts string html to string text with html2text.
+
+  Args:
+    **kwargs: html2text options
+      https://github.com/Alir3z4/html2text/blob/master/docs/usage.md#available-options
+  """
   if html:
     h = html2text.HTML2Text()
     h.unicode_snob = True
     h.body_width = 0  # don't wrap lines
-    h.ignore_links = True
-    h.ignore_images = True
+    h.ignore_links = kwargs.get('ignore_links', True)
+    h.ignore_images = kwargs.get('ignore_images', True)
 
     # hacky monkey patch fix for html2text escaping sequences that are
     # significant in markdown syntax. the X\\Y replacement depends on knowledge
@@ -169,10 +174,13 @@ class Source(object):
     placeholder for the post URL and (optionally) a %(content)s placeholder
     for the post content.
   * POST_ID_RE: regexp, optional, matches valid post ids. Used in post_id().
+  * HTML2TEXT_OPTIONS: dict mapping string html2text option names to values
+    https://github.com/Alir3z4/html2text/blob/master/docs/usage.md#available-options
   """
   __metaclass__ = SourceMeta
 
   POST_ID_RE = None
+  HTML2TEXT_OPTIONS = {}
 
   def user_url(self, user_id):
     """Returns the URL for a user's profile."""
@@ -917,7 +925,7 @@ def _content_for_create(self, obj, ignore_formatting=False, prefer_name=False,
     is_html = (bool(BeautifulSoup(content, 'html.parser').find()) or
                HTML_ENTITY_RE.search(content))
     if is_html and not ignore_formatting:
-      content = html_to_text(content)
+      content = html_to_text(content, **self.HTML2TEXT_OPTIONS)
     elif not is_html and ignore_formatting:
       content = re.sub(r'\s+', ' ', content)
 

granary/test/test_github.py

@@ -616,6 +616,29 @@ def _test_create_issue(self, in_reply_to):
       'url': 'https://github.com/foo/bar/issues/123',
     }, result.content)
 
+  def test_create_with_image_and_link(self):
+    self.expect_requests_post(github.REST_API_CREATE_ISSUE % ('foo', 'bar'), json={
+        'title': 'an issue title',
+        'body': '[bar](http://foo/) ![](https://baz/)',
+      }, response={
+        'html_url': 'https://github.com/foo/bar/issues/123',
+      }, headers=EXPECTED_HEADERS)
+    self.mox.ReplayAll()
+
+    result = self.gh.create({
+      'url': 'https://github.com/foo/bar/issues/333',
+      'title': 'an issue title',
+      'content': """
+<a href="http://foo/">bar</a>
+<img src="https://baz/" />
+""",
+      'inReplyTo': [{'url': 'https://github.com/foo/bar/issues'}],
+    })
+    self.assertIsNone(result.error_plain, result)
+    self.assert_equals({
+      'url': 'https://github.com/foo/bar/issues/123',
+    }, result.content)
+
   def test_preview_issue(self):
     for i in range(2):
       rendered = self.expect_markdown_render(ISSUE_OBJ['content'].strip())

requirements.freeze.txt

@@ -5,7 +5,7 @@ chardet==3.0.4
 coverage==4.0.3
 -e git+https://github.com/snarfed/gdata-python-client-1.git@1df4e1efea7e5cf2754bc7eec6c1ab48ab09e3b1#egg=gdata
 google-api-python-client==1.6.3
-html2text==2016.9.19
+html2text==2018.1.9
 html5lib==0.9999999
 httplib2==0.10.3
 humanize==0.5.1