download report: just skip bad download entries

We get these about once a week where a single download entry will have
some invalid utf8 junk in it. Its almost always just a single row, so
lets just discard it for now. We'll still get a single email so we could
in theory add the entry into the DB if we think its legit.

Change-Id: Icb8eebcae0f7b3f04756a3e22ea447a07a210bc2
Reviewed-on: https://review.linaro.org/20505
Reviewed-by: Ben Copeland <ben.copeland@linaro.org>

1 files changed, 8 insertions, 4 deletions

diff --git a/license_protected_downloads/management/commands/report_process.py b/license_protected_downloads/management/commands/report_process.py
index 4382b14..288a6d0 100644
--- a/license_protected_downloads/management/commands/report_process.py
+++ b/license_protected_downloads/management/commands/report_process.py
@@ -1,7 +1,7 @@
 from django.conf import settings
 from django.core.management.base import BaseCommand
 from license_protected_downloads.models import Download
-from django.db import DatabaseError
+from django.db import DataError, DatabaseError
 import os
 import time
 import glob
@@ -45,9 +45,13 @@ class Command(BaseCommand):
                     # This looks odd, but we sometimes get URLs with newlines
                     # in them and we need the real file name
                     download = row[1].replace('\n', '')
-                    Download.objects.create(
-                        ip=row[0], name=download, link=str2bool(row[2]),
-                        ref=row[3])
+                    try:
+                        Download.objects.create(
+                            ip=row[0], name=download, link=str2bool(row[2]),
+                            ref=row[3])
+                    except DataError:
+                        logging.error(
+                            'Skipping invalid download entry: %s', row)
                 os.remove(name)
             except (csv.Error, DatabaseError):
                 logging.exception('unable to process csv %s', name)