MAINT: Skip more rows to match change in treasury data format

I'm not sure what the raw csv pulled from the federal reserve looked like before, but when trying to download fresh treasure data (data not stored in `./zipline`), there is an error that says "Time Period not in list". After checking the raw csv now, it looks like there are 5 header rows rather than just 1, so skipping those rows removes that error.

tests/test_examples.py

@@ -25,6 +25,8 @@ from zipline.testing import test_resource_path
 from zipline.testing.fixtures import WithTmpDir, ZiplineTestCase
 from zipline.testing.predicates import assert_equal
 from zipline.utils.cache import dataframe_cache
+from zipline.utils.paths import ensure_file
+
 
 # Otherwise the next line sometimes complains about being run too late.
 _multiprocess_can_split_ = False
@@ -53,6 +55,10 @@ class ExamplesTests(WithTmpDir, ZiplineTestCase):
             serialization='pickle',
         )
 
+        market_data = ('SPY_benchmark.csv', 'treasury_curves.csv')
+        for data in market_data:
+            ensure_file(cls.tmpdir.getpath('example_data/root/data/' + data))
+
     @parameterized.expand(sorted(examples.EXAMPLE_MODULES))
     def test_example(self, example_name):
         actual_perf = examples.run_example(