From a206c9132fab50bec9c70ea450bdb71cbc576648 Mon Sep 17 00:00:00 2001
From: Kernc <kerncece@gmail.com>
Date: Mon, 23 Feb 2015 04:43:58 +0100
Subject: [PATCH] Add support for YAML metadata in Markdown files

~~Python-Markdown>=2.6 and its meta extension supports YAML headers
and optional `yaml` switch which, when used, parses data with
PyYAML and hence a wee bit different metadata object gets provided.~~
Not anymore.

YAML is supported by python-markdown-yaml-meta-data extension
which uses PyYAML, which returns parsed lists of strings instead
of raw strings, datetime objects instead of string date
representations etc. Pelican needed only slight adjusting, and
now support Jekyll-like YAML headers with aforementioned Markdown
extension.

Related:
* https://github.com/teoric/python-markdown-yaml-meta-data
* https://github.com/getpelican/pelican-plugins/pull/382
* https://github.com/waylan/Python-Markdown/issues/390#issuecomment-76200467
---
 pelican/readers.py | 19 ++++++++++---------
 pelican/utils.py   |  9 ++++++++-
 2 files changed, 18 insertions(+), 10 deletions(-)

diff --git a/pelican/readers.py b/pelican/readers.py
index 3656cd96..8f57cf88 100644
--- a/pelican/readers.py
+++ b/pelican/readers.py
@@ -30,7 +30,9 @@ from pelican.utils import get_date, pelican_open, FileStampDataCacher, SafeDatet
 
 def strip_split(text, sep=','):
     """Return a list of stripped, non-empty substrings, delimited by sep."""
-    items = [x.strip() for x in text.split(sep)]
+    if not isinstance(text, list):
+        text = text.split(sep)
+    items = [x.strip() for x in text]
     return [x for x in items if x]
 
 
@@ -90,6 +92,8 @@ class BaseReader(object):
 
     def process_metadata(self, name, value):
         if name in METADATA_PROCESSORS:
+            if isinstance(value, list) and len(value) == 1:
+                value = value[0]
             return METADATA_PROCESSORS[name](value, self.settings)
         return value
 
@@ -249,16 +253,13 @@ class MarkdownReader(BaseReader):
                 summary = self._md.convert(summary_values)
                 output[name] = self.process_metadata(name, summary)
             elif name in METADATA_PROCESSORS:
-                if len(value) > 1:
-                    logger.warning('Duplicate definition of `%s` '
-                        'for %s. Using first one.', name, self._source_path)
-                output[name] = self.process_metadata(name, value[0])
-            elif len(value) > 1:
-                # handle list metadata as list of string
                 output[name] = self.process_metadata(name, value)
-            else:
-                # otherwise, handle metadata as single string
+            elif isinstance(value, list) and len(value) == 1:
+                # handle metadata as a single string
                 output[name] = self.process_metadata(name, value[0])
+            else:
+                # otherwise, handle metadata as single item
+                output[name] = self.process_metadata(name, value)
         return output
 
     def read(self, source_path):
diff --git a/pelican/utils.py b/pelican/utils.py
index caac8e61..13880753 100644
--- a/pelican/utils.py
+++ b/pelican/utils.py
@@ -222,9 +222,16 @@ def get_date(string):
 
     If no format matches the given date, raise a ValueError.
     """
-    string = re.sub(' +', ' ', string)
     default = SafeDatetime.now().replace(hour=0, minute=0,
                                         second=0, microsecond=0)
+    if isinstance(string, datetime.datetime):
+        # Return datetime object as SafeDatetime
+        return default.replace(year=string.year, month=string.month,
+                               day=string.day, hour=string.hour,
+                               minute=string.minute, second=string.second,
+                               microsecond=string.microsecond,
+                               tzinfo=string.tzinfo)
+    string = re.sub('[ _]+', ' ', string)
     try:
         return dateutil.parser.parse(string, default=default)
     except (TypeError, ValueError):