Merge pull request #11 from sjaensch/load-csv-use-post

Use POST for loading CSV, document why we're using GET for the other tasks

tests/views/task_test.py

@@ -44,7 +44,7 @@ def test_load_employees_from_s3(self, mock_load_employees):  # noqa
 
     @mock.patch('logic.employee.load_employees_from_csv', autospec=True)
     def test_load_employees_from_csv(self, mock_load_employees_from_csv):  # noqa
-        response = self.app.get('/tasks/employees/load/csv')
+        response = self.app.post('/tasks/employees/load/csv')
 
         self.assertEqual(response.status_int, 200)
         mock_load_employees_from_csv.assert_called_with()

views/tasks.py

@@ -12,18 +12,22 @@
 from models import Love
 
 
+# All tasks that are to be executed by cron need to use HTTP GET
+# see https://cloud.google.com/appengine/docs/python/config/cron
 @app.route('/tasks/employees/load/s3', methods=['GET'])
 def load_employees_from_s3():
     logic.employee.load_employees()
     return Response(status=200)
 
 
-@app.route('/tasks/employees/load/csv', methods=['GET'])
+# This task has a web UI to trigger it, so let's use POST
+@app.route('/tasks/employees/load/csv', methods=['POST'])
 def load_employees_from_csv():
     logic.employee.load_employees_from_csv()
     return Response(status=200)
 
 
+# One-off tasks are much easier to trigger using GET
 @app.route('/tasks/employees/combine', methods=['GET'])
 def combine_employees():
     old_username, new_username = request.args['old'], request.args['new']

views/web.py

@@ -327,5 +327,5 @@ def import_employees_form():
 @admin_required
 def import_employees():
     flash('We started importing employee data in the background. Refresh the page to see it.', 'info')
-    taskqueue.add(url='/tasks/employees/load/csv', method='GET')
+    taskqueue.add(url='/tasks/employees/load/csv')
     return redirect(url_for('employees'))