From eb3c2fc39809124f859e3b64f5c33025b85146ac Mon Sep 17 00:00:00 2001
From: Jonathan Harker <jonathan@catalyst.net.nz>
Date: Tue, 24 Nov 2015 10:09:48 +1300
Subject: [PATCH] URL title is content type header for non HTML responses.

---
 models.py | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/models.py b/models.py
index 61e8d2a..00fa79e 100644
--- a/models.py
+++ b/models.py
@@ -68,8 +68,11 @@ class Url(Model):
         if title is None:
             r = requests.get(url)
             if r.status_code == 200:
-                dom = BeautifulSoup(r.content, 'html.parser')
-                self.title = dom.title.string
+                if 'text/html' in r.headers['content-type']:
+                    dom = BeautifulSoup(r.content, 'html.parser')
+                    self.title = dom.title.string
+                else:
+                    self.title = r.headers['content-type']
             else:
                 self.title = "Error: HTTP %s" % r.status_code