docs, url parsing regexes and a plan for tomorrow

event_feed.py

@@ -29,13 +29,26 @@ template = env.get_template('event_template.md')
 existing_posts = os.listdir(output_dir)
 
 def findURLs(string):
-        
+    """
+    return all URLs in a given string
+    """
     regex = r"(?i)\b((?:https?://|www\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}/)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:'\".,<>?«»“”‘’]))"
     url = re.findall(regex,string)      
     return [x[0] for x in url]     
 
+def find_imageURLS(string):
+    """
+    return all image URLS in a given string
+    """
+    regex = r"(?:http\:|https\:)?\/\/.*?\.(?:png|jpg|jpeg|gif|svg)" 
+
+    img_urls = re.findall(regex, string, flags=re.IGNORECASE)
+    return img_urls 
+
 def create_metadata(event):
-    #construct a formatted dict of event metadata for in a post
+    """
+    construct a formatted dict of event metadata for use as frontmatter for HUGO post
+    """
 
     if event.location:
         location_urls = findURLs(event.location)
@@ -57,9 +70,16 @@ def create_metadata(event):
     'uid': event.uid
     }
 
+    #TODO:  1 find image urls in description
+    #       2 download them to hugo post
+    #       3 replace image url with local path
+
     return event_metadata
 
 def localize_time(date):
+    """
+    Turn a given date into various timezones
+    """
     # Dates need to be displayed for the various TZs
     # takes arrow objects
 
@@ -104,6 +124,9 @@ def create_event_post(post_dir, event):
 
 
 def update_event_post(post_dir, event):
+    """
+    Update a post based on the VCARD event 'created' field which changes when updated
+    """ 
     if os.path.exists(post_dir):
         old_timestamp = open(os.path.join(post_dir,'.timestamp')).read()
         if event.created > arrow.get(old_timestamp):