p-value caluculation script

2 files changed, 29 insertions, 1 deletions

diff --git a/data/make_reference_output.py b/data/make_reference_output.py
new file mode 100755
index 0000000..1cd31ae
--- /dev/null
+++ b/data/make_reference_output.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python
+
+import csv
+import os
+
+from fuel.iterator import DataIterator
+from fuel.schemes import SequentialExampleScheme
+from fuel.streams import DataStream
+
+from data.hdf5 import TaxiDataset
+import data
+
+dest_outfile = open(os.path.join(data.path, 'test_answer.csv'), 'w')
+dest_outcsv = csv.writer(dest_outfile)
+dest_outcsv.writerow(["TRIP_ID", "LATITUDE", "LONGITUDE"])
+
+dataset = TaxiDataset('test', 'tvt.hdf5',
+                     sources=('trip_id', 'longitude', 'latitude',
+                              'destination_longitude', 'destination_latitude'))
+it = DataIterator(DataStream(dataset), iter(xrange(dataset.num_examples)), as_dict=True)
+
+for v in it:
+    # print v
+    dest_outcsv.writerow([v['trip_id'], v['destination_latitude'],
+                                        v['destination_longitude']])
+
+dest_outfile.close()
+
diff --git a/data/transformers.py b/data/transformers.py
index f0ed44a..479afc5 100644
--- a/data/transformers.py
+++ b/data/transformers.py
@@ -187,7 +187,7 @@ class _window_helper(object):
 
         if x.shape[0] < self.window_len:
             x = numpy.concatenate(
-                [x, numpy.full((self.window_len - x.shape[0],), x[-1])])
+                [numpy.full((self.window_len - x.shape[0],), x[0]), x])
             
         y = [x[i: i+x.shape[0]-self.window_len+1][:, None]
              for i in range(self.window_len)]