fix unsorted series in build_time_series function, change doc string … (#172)

* fix unsorted series in build_time_series function, change doc string to add sorted timeseries

* add unittest for build_time_series, changed build_time_series to avoid warning thrown

Author: manojbalaji1

datacommons_pandas/df_builder.py

@@ -46,11 +46,15 @@ def build_time_series(place,
       scaling_factor (`int`): Optional, the preferred `scalingFactor` value.
     Returns:
       A pandas Series with Place IDs as the index and observed statistics as
-      values, representing a time series satisfying all optional args.
+      values, representing a sorted time series satisfying all optional args.
     """
-    return pd.Series(
-        dc.get_stat_series(place, stat_var, measurement_method,
-                           observation_period, unit, scaling_factor))
+    result_dict = dc.get_stat_series(place, stat_var, measurement_method,observation_period, unit, scaling_factor)
+    
+    # Explicit dtype to avoid warning thrown by pd.Series({})
+    if not result_dict:
+        return pd.Series(result_dict,dtype=object)
+    else:
+        return pd.Series(result_dict).sort_index()
 
 
 def _group_stat_all_by_obs_options(places, stat_vars, keep_series=True):

datacommons_pandas/test/df_builder_test.py

@@ -31,6 +31,7 @@
 import unittest
 import six
 import six.moves.urllib as urllib
+import pandas as pd
 
 # Reusable parts of REST API /stat/all response.
 CA_COUNT_PERSON = {
@@ -219,6 +220,34 @@ def read(self):
     return urllib.error.HTTPError(None, 404, None, None, None)
 
 
+class TestBuildTimeSeries(unittest.TestCase):
+    """Unit tests for build_time_series."""
+    
+    @patch('six.moves.urllib.request.urlopen', side_effect=request_mock)
+    def test_basic(self, urlopen):
+        """Calling build_time_series with basic args."""
+        series = dcpd.build_time_series('geoId/06', 'Count_Person')
+        exp = pd.Series({"2000": 1, "2001": 2})
+        
+        self.assertCountEqual(series, exp)
+
+    @patch('six.moves.urllib.request.urlopen', side_effect=request_mock)
+    def test_multi_option(self, urlopen):
+        """Calling build_time_series with basic args."""
+        series = dcpd.build_time_series('geoId/06', 'Count_Person', 'CensusPEPSurvey', 'P1Y', 'RealPeople', '100')
+        exp = pd.Series({"2000": 3, "2001": 42})
+
+        self.assertCountEqual(series, exp)
+
+    @patch('six.moves.urllib.request.urlopen', side_effect=request_mock)
+    def test_no_data(self, urlopen):
+        """Error if there is no data."""
+        series = dcpd.build_time_series('geoId/06', 'Count_Person', 'DNE')
+        exp = pd.Series({}, dtype=object)
+
+        self.assertCountEqual(series, exp)
+
+
 class TestPdTimeSeries(unittest.TestCase):
     """Unit tests for _time_series_pd_input."""