merge updates from base Manishearth's repo

chatexchange/_utils.py

@@ -55,6 +55,23 @@ def html_to_text(html):
     return s.get_text()
 
 
+# Number of seconds since the user was last seen, based on <12d ago> data.
+def parse_last_seen(text):
+    suffixes = {
+        's': 1,
+        'm': 60,
+        'h': 3600,
+        'd': 86400,
+        'y': 31536000
+    }
+    splat = text.split(' ')
+    assert len(splat) == 2, "text doesn't appear to be in <x ago> format"
+    char = text[-1]
+    number = int(text[:-1])
+    assert char in suffixes, "suffix char unrecognized"
+    return number * suffixes[char]
+
+
 class LazyFrom(object):
     """
     A descriptor used when multiple lazy attributes depend on a common

chatexchange/browser.py

@@ -69,7 +69,7 @@ def _request(
         # Try again if we fail. We're blaming "the internet" for weirdness.
         MAX_HTTP_RETRIES = 5                # EGAD! A MAGIC NUMBER!
         attempt = 0
-        while attempt <= MAX_HTTP_RETRIES:      
+        while attempt <= MAX_HTTP_RETRIES:
             attempt += 1
             response = None
             try:
@@ -223,7 +223,7 @@ def _load_user(self, soup):
     @staticmethod
     def user_id_and_name_from_link(link_soup):
         user_name = link_soup.text
-        user_id = int(link_soup['href'].split('/')[2])
+        user_id = int(link_soup['href'].split('/')[-2])
         return user_id, user_name
 
     def _update_chat_fkey_and_user(self):
@@ -410,7 +410,7 @@ def get_transcript_with_message(self, message_id):
 
         room_soups = transcript_soup.select('.room-name a')
         room_soup = room_soups[-1]
-        room_id = int(room_soup['href'].split('/')[2])
+        room_id = int(room_soup['href'].split('/')[-2])
         room_name = room_soup.text
 
         messages_data = []
@@ -549,12 +549,19 @@ def get_profile(self, user_id):
         else:
             reputation = -1
 
+        stats_elements = profile_soup.select('.user-valuecell')
+        if len(stats_elements) >= 3:
+            last_seen = _utils.parse_last_seen(stats_elements[2].text)
+        else:
+            last_seen = _utils.parse_last_seen('20y ago')
+
         return {
             'name': name,
             'is_moderator': is_moderator,
             'message_count': message_count,
             'room_count': room_count,
-            'reputation': reputation
+            'reputation': reputation,
+            'last_seen': last_seen
         }
 
     def get_room_info(self, room_id):