fix pdf tester

boto3 storage returns unicode not str someplace

core/pdf.py

@@ -1,6 +1,7 @@
 """
 Utilities that manipulate pdf files
 """
+import logging
 import requests
 from xhtml2pdf import pisa             # import python module
 from PyPDF2 import PdfFileMerger,PdfFileReader
@@ -9,6 +10,7 @@ from tempfile import NamedTemporaryFile
 from django.template.loader import render_to_string
 from regluit import settings
 
+logger = logging.getLogger(__name__)
 
 # Utility function
 def ask_pdf(context={}):
@@ -35,7 +37,7 @@ def pdf_append( file1, file2, file_out ):
 def test_pdf(pdf_file):
     temp = None
     try:
-        if isinstance(pdf_file , str):
+        if isinstance(pdf_file , (str, unicode)):
             if pdf_file.startswith('http:') or pdf_file.startswith('https:'):
                 temp = NamedTemporaryFile(delete=False)
                 test_file_content = requests.get(pdf_file).content
@@ -53,7 +55,9 @@ def test_pdf(pdf_file):
         except:
             success = False
         return success
-    except:
+    except Exception:
+        pdf_file = unicode(pdf_file)
+        logger.exception('error testing a pdf: %s' % pdf_file[:100])
         return False
 
 def test_test_pdf(self):

core/tests.py

@@ -1119,7 +1119,8 @@ class EbookFileTests(TestCase):
         #test the ask-appender
         c.add_ask_to_ebfs()
         if settings.AWS_SECRET_ACCESS_KEY:
-            assert test_pdf(c.work.ebookfiles().filter(asking=True)[0].file.url)
+            askingpdfurl = c.work.ebookfiles().filter(asking=True)[0].file.url
+            assert test_pdf(askingpdfurl)
         else:
             assert test_pdf(c.work.ebookfiles().filter(asking=True)[0].file)