Reinstate Yahoo API (actions, price history)

Author: David Stephens

pandas_datareader/data.py

@@ -32,6 +32,7 @@
     RobinhoodQuoteReader
 from pandas_datareader.stooq import StooqDailyReader
 from pandas_datareader.tiingo import TiingoDailyReader, TiingoQuoteReader
+from pandas_datareader.yahoo.actions import (YahooActionReader, YahooDivReader)
 from pandas_datareader.yahoo.components import _get_data as \
     get_components_yahoo
 from pandas_datareader.yahoo.daily import YahooDailyReader
@@ -79,11 +80,11 @@ def get_quote_av(*args, **kwargs):
 
 
 def get_data_yahoo_actions(*args, **kwargs):
-    raise ImmediateDeprecationError(DEP_ERROR_MSG.format('Yahoo Actions'))
+    return YahooActionReader(*args, **kwargs).read()
 
 
 def get_quote_yahoo(*args, **kwargs):
-    raise ImmediateDeprecationError(DEP_ERROR_MSG.format('Yahoo Actions'))
+    raise ImmediateDeprecationError(DEP_ERROR_MSG.format('Yahoo Quotes'))
     return YahooQuotesReader(*args, **kwargs).read()
 
 
@@ -305,7 +306,6 @@ def DataReader(name, data_source=None, start=None, end=None,
     ff = DataReader("F-F_ST_Reversal_Factor", "famafrench")
     """
     if data_source == "yahoo":
-        raise ImmediateDeprecationError(DEP_ERROR_MSG.format('Yahoo Daily'))
         return YahooDailyReader(symbols=name, start=start, end=end,
                                 adjust_price=False, chunksize=25,
                                 retry_count=retry_count, pause=pause,
@@ -401,6 +401,18 @@ def DataReader(name, data_source=None, start=None, end=None,
                                  retry_count=retry_count, pause=pause,
                                  session=session,
                                  api_key=access_key).read()
+
+    elif data_source == "yahoo-actions":
+        return YahooActionReader(symbols=name, start=start, end=end,
+                                 retry_count=retry_count, pause=pause,
+                                 session=session).read()
+
+    elif data_source == "yahoo-dividends":
+        return YahooDivReader(symbols=name, start=start, end=end,
+                              adjust_price=False, chunksize=25,
+                              retry_count=retry_count, pause=pause,
+                              session=session, interval='d').read()
+
     else:
         msg = "data_source=%r is not implemented" % data_source
         raise NotImplementedError(msg)

pandas_datareader/tests/yahoo/test_yahoo.py

@@ -21,7 +21,6 @@ class TestYahoo(object):
 
     @classmethod
     def setup_class(cls):
-        pytest.skip('Skip all Yahoo! tests.')
         pytest.importorskip("lxml")
 
     @skip_on_exception(RemoteDataError)
@@ -30,7 +29,8 @@ def test_yahoo(self):
         start = datetime(2010, 1, 1)
         end = datetime(2013, 1, 25)
 
-        assert web.DataReader('F', 'yahoo', start, end)['Close'][-1] == 13.68
+        assert round(web.DataReader('F', 'yahoo', start, end)['Close'][-1],
+                     2) == 13.68
 
     def test_yahoo_fails(self):
         start = datetime(2010, 1, 1)
@@ -39,13 +39,15 @@ def test_yahoo_fails(self):
         with pytest.raises(Exception):
             web.DataReader('NON EXISTENT TICKER', 'yahoo', start, end)
 
+    @pytest.mark.skip('Yahoo quotes deprecated')
     def test_get_quote_series(self):
         try:
             df = web.get_quote_yahoo(pd.Series(['GOOG', 'AAPL', 'GOOG']))
         except ConnectionError:
             pytest.xfail(reason=XFAIL_REASON)
         tm.assert_series_equal(df.iloc[0], df.iloc[2])
 
+    @pytest.mark.skip('Yahoo quotes deprecated')
     def test_get_quote_string(self):
         _yahoo_codes.update({'MarketCap': 'j1'})
         try:
@@ -55,13 +57,15 @@ def test_get_quote_string(self):
 
         assert not pd.isnull(df['MarketCap'][0])
 
+    @pytest.mark.skip('Yahoo quotes deprecated')
     def test_get_quote_stringlist(self):
         try:
             df = web.get_quote_yahoo(['GOOG', 'AAPL', 'GOOG'])
         except ConnectionError:
             pytest.xfail(reason=XFAIL_REASON)
         tm.assert_series_equal(df.iloc[0], df.iloc[2])
 
+    @pytest.mark.skip('Yahoo quotes deprecated')
     def test_get_quote_comma_name(self):
         _yahoo_codes.update({'name': 'n'})
         try:
@@ -113,6 +117,10 @@ def test_get_data_single_symbol(self):
         # just test that we succeed
         web.get_data_yahoo('GOOG')
 
+    @skip_on_exception(RemoteDataError)
+    def test_data_with_no_actions(self):
+        web.get_data_yahoo('TSLA')
+
     @skip_on_exception(RemoteDataError)
     def test_get_data_adjust_price(self):
         goog = web.get_data_yahoo('GOOG')
@@ -165,11 +173,11 @@ def test_get_data_null_as_missing_data(self, adj_pr):
     def test_get_data_multiple_symbols_two_dates(self):
         pan = web.get_data_yahoo(['GE', 'MSFT', 'INTC'], 'JAN-01-12',
                                  'JAN-31-12')
-        result = pan.Close['01-18-12'].T
+        result = pan.Close.loc['01-18-12'].T
         assert result.size == 3
 
         # sanity checking
-        assert result.dtypes.all() == np.floating
+        assert result.dtypes == np.floating
 
         expected = np.array([[18.99, 28.4, 25.18],
                              [18.58, 28.31, 25.13],
@@ -179,7 +187,6 @@ def test_get_data_multiple_symbols_two_dates(self):
         result = df[(df.index >= 'Jan-15-12') & (df.index <= 'Jan-20-12')]
         assert expected.shape == result.shape
 
-    @pytest.mark.xfail(reason="failing after #355")
     def test_get_date_ret_index(self):
         pan = web.get_data_yahoo(['GE', 'INTC', 'IBM'], '1977', '1987',
                                  ret_index=True)
@@ -193,7 +200,6 @@ def test_get_date_ret_index(self):
         # sanity checking
         assert np.issubdtype(pan.values.dtype, np.floating)
 
-    @pytest.mark.xfail(reason="failing after #355")
     def test_get_data_yahoo_actions(self):
         start = datetime(1990, 1, 1)
         end = datetime(2000, 4, 5)
@@ -228,7 +234,6 @@ def test_yahoo_reader_class(self):
         r = YahooDailyReader('GOOG', session=session)
         assert r.session is session
 
-    @pytest.mark.xfail(reason="failing after #355")
     def test_yahoo_DataReader(self):
         start = datetime(2010, 1, 1)
         end = datetime(2015, 5, 9)
@@ -241,6 +246,22 @@ def test_yahoo_DataReader(self):
                                     '2013-08-08', '2013-05-09',
                                     '2013-02-07', '2012-11-07',
                                     '2012-08-09'])
+
+        exp = pd.DataFrame({'action': ['DIVIDEND', 'DIVIDEND', 'DIVIDEND',
+                                       'DIVIDEND', 'SPLIT', 'DIVIDEND',
+                                       'DIVIDEND', 'DIVIDEND',
+                                       'DIVIDEND', 'DIVIDEND', 'DIVIDEND',
+                                       'DIVIDEND', 'DIVIDEND'],
+                            'value': [0.52, 0.47, 0.47, 0.47, 0.14285714,
+                                      3.29, 3.05, 3.05, 3.05,
+                                      3.05, 2.65, 2.65, 2.65]},
+                           index=exp_idx)
+        exp.index.name = 'Date'
+        tm.assert_frame_equal(result.reindex_like(exp).round(5), exp.round(5))
+
+        result = web.get_data_yahoo_actions('AAPL', start, end,
+                                            adjust_dividends=True)
+
         exp = pd.DataFrame({'action': ['DIVIDEND', 'DIVIDEND', 'DIVIDEND',
                                        'DIVIDEND', 'SPLIT', 'DIVIDEND',
                                        'DIVIDEND', 'DIVIDEND',
@@ -251,13 +272,11 @@ def test_yahoo_DataReader(self):
                                       0.43571, 0.37857, 0.37857, 0.37857]},
                            index=exp_idx)
         exp.index.name = 'Date'
-
-        tm.assert_frame_equal(result.reindex_like(exp), exp)
+        tm.assert_frame_equal(result.reindex_like(exp).round(5), exp.round(5))
 
     @skip_on_exception(RemoteDataError)
     def test_yahoo_DataReader_multi(self):
         start = datetime(2010, 1, 1)
         end = datetime(2015, 5, 9)
-
         result = web.DataReader(['AAPL', 'F'], 'yahoo-actions', start, end)
         assert isinstance(result, pd.Panel)

pandas_datareader/yahoo/actions.py

@@ -0,0 +1,59 @@
+from pandas import (Panel, concat, DataFrame)
+from pandas_datareader.yahoo.daily import YahooDailyReader
+
+
+class YahooActionReader(YahooDailyReader):
+    """
+    Returns DataFrame of historical corporate actions (dividends and stock
+    splits) from symbols, over date range, start to end. All dates in the
+    resulting DataFrame correspond with dividend and stock split ex-dates.
+    """
+
+    def read(self):
+        data = super(YahooActionReader, self).read()
+        if isinstance(data, Panel):
+            data = data.swapaxes('minor', 'items')
+            actions = {}
+            for s in data:
+                actions[s] = _get_one_action(data[s])
+            return Panel(actions).swapaxes('items', 'minor')
+        else:
+            return _get_one_action(data)
+
+    @property
+    def get_actions(self):
+        return True
+
+
+def _get_one_action(data):
+    actions = DataFrame(columns=['action', 'value'])
+
+    if 'Dividends' in data.columns:
+        # Add a label column so we can combine our two DFs
+        dividends = DataFrame(data['Dividends']).dropna()
+        dividends["action"] = "DIVIDEND"
+        dividends = dividends.rename(columns={'Dividends': 'value'})
+        actions = concat([actions, dividends]).sort_index(ascending=False)
+
+    if 'Splits' in data.columns:
+        # Add a label column so we can combine our two DFs
+        splits = DataFrame(data['Splits']).dropna()
+        splits["action"] = "SPLIT"
+        splits = splits.rename(columns={'Splits': 'value'})
+        actions = concat([actions, splits]).sort_index(ascending=False)
+
+    return actions
+
+
+class YahooDivReader(YahooActionReader):
+
+    def read(self):
+        data = super(YahooDivReader, self).read()
+        return data[data['action'] == 'DIVIDEND']
+
+
+class YahooSplitReader(YahooActionReader):
+
+    def read(self):
+        data = super(YahooSplitReader, self).read()
+        return data[data['action'] == 'SPLIT']