4 files changed, 66 insertions, 0 deletions
diff --git a/visualizer/extractor/destinations.py b/visualizer/extractor/destinations.py
new file mode 100755
index 0000000..967e766
--- /dev/null
+++ b/visualizer/extractor/destinations.py
@@ -0,0 +1,19 @@
+#!/usr/bin/env python
+
+from data.hdf5 import taxi_it
+from visualizer import Vlist, Point
+
+
+_sample_size = 5000
+
+if __name__ == '__main__':
+    points = Vlist(cluster=True)
+    for line in taxi_it('train'):
+        if len(line['latitude'])>0:
+            points.append(Point(line['latitude'][-1], line['longitude'][-1]))
+            if len(points) >= _sample_size:
+                break
+    points.save('destinations (cluster)')
+    points.cluster = False
+    points.heatmap = True
+    points.save('destinations (heatmap)')
diff --git a/visualizer/extractor/stands.py b/visualizer/extractor/stands.py
new file mode 100755
index 0000000..9224143
--- /dev/null
+++ b/visualizer/extractor/stands.py
@@ -0,0 +1,14 @@
+#!/usr/bin/env python
+
+from data.hdf5 import taxi_it
+from visualizer import Vlist, Point
+
+
+if __name__ == '__main__':
+    it = taxi_it('stands')
+    next(it) # Ignore the "no stand" entry
+
+    points = Vlist()
+    for (i, line) in enumerate(it):
+        points.append(Point(line['stands_latitude'], line['stands_longitude'], 'Stand (%d): %s' % (i+1, line['stands_name'])))
+    points.save('stands')
diff --git a/visualizer/extractor/test_positions.py b/visualizer/extractor/test_positions.py
new file mode 100755
index 0000000..a84d2ba
--- /dev/null
+++ b/visualizer/extractor/test_positions.py
@@ -0,0 +1,12 @@
+#!/usr/bin/env python
+
+from data.hdf5 import taxi_it
+from visualizer import Vlist, Point
+
+
+if __name__ == '__main__':
+    points = Vlist(heatmap=True)
+    for line in taxi_it('test'):
+        for (lat, lon) in zip(line['latitude'], line['longitude']):
+            points.append(Point(lat, lon))
+    points.save('test positions')
diff --git a/visualizer/extractor/train_poi.py b/visualizer/extractor/train_poi.py
new file mode 100755
index 0000000..a4ccbca
--- /dev/null
+++ b/visualizer/extractor/train_poi.py
@@ -0,0 +1,21 @@
+#!/usr/bin/env python
+
+import os
+
+import data
+from data.hdf5 import TaxiDataset
+from visualizer import Path
+
+
+poi = {
+    'longest': 1492417
+}
+
+if __name__ == '__main__':
+    prefix = os.path.join(data.path, 'visualizer', 'Train POI')
+    if not os.path.isdir(prefix):
+        os.mkdir(prefix)
+
+    d = TaxiDataset('train')
+    for (k, v) in poi.items():
+        Path(d.extract(v)).save(os.path.join('Train POI', k))