[jealousGH-122] Update calculation of middle (jealous#129)

The typical price or middle price is calculated by:
```
(high + low + close) / 3
```

But it would be accurate if the data source has the `amount` that
represents the total amount of capital traded in that period.
Use `middle = amount / volume` when amout is available.

Also fix some typos in the read me file.
Add some utilities for dropping columns, head or tail.

Add python 3.11 in CI and drop 3.9.

Update version to 0.5.0

Author: jealous

.github/workflows/build-test.yml

@@ -11,7 +11,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [ "2.7", "3.9", "3.10" ]
+        python-version: [ "2.7", "3.10", "3.11" ]
     steps:
       - uses: actions/checkout@v2
       - name: Set up Python ${{ matrix.python-version }}
@@ -32,7 +32,7 @@ jobs:
         run: |
           py.test --cov=stockstats test.py
       - name: Publish Test Results
-        if: ${{ matrix.python-version == '3.10' }}
+        if: ${{ matrix.python-version == '3.11' }}
         run: codecov
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

README.md

@@ -4,7 +4,7 @@
 [![codecov](https://codecov.io/gh/jealous/stockstats/branch/master/graph/badge.svg?token=IFMD1pVJ7T)](https://codecov.io/gh/jealous/stockstats)
 [![pypi](https://img.shields.io/pypi/v/stockstats.svg)](https://pypi.python.org/pypi/stockstats)
 
-VERSION: 0.4.1
+VERSION: 0.5.0
 
 ## Introduction
 
@@ -263,7 +263,7 @@ In [22]: tp = df['middle']
 
 In [23]: df['res'] = df['middle'] > df['close']        
 
-In [24]: df[['middle', 'close', 'res', 'res_-10_c']]    
+In [24]: df[['middle', 'close', 'res', 'res_10_c']]    
 Out[24]:                                               
              middle  close    res  res_10_c            
 date                                                   
@@ -502,7 +502,7 @@ resistance levels.
 It includes three lines:
 * `df['kdjk']` - K series
 * `df['kdjd']` - D series
-* `df['kdfj']` - J series
+* `df['kdjj']` - J series
 
 The default window is 9.  Use `StockDataFrame.KDJ_WINDOW` to change it.
 Use `df['kdjk_6']` to retrieve the K series of 6 periods.
@@ -528,6 +528,9 @@ It contains 4 lines:
 It's the average of `high`, `low` and `close`.
 Use `df['middle']` to access this value.
 
+When `amount` is available, `middle = amount / volume`
+This should be more accurate because amount represents the total cash flow. 
+
 #### [Bollinger Bands](https://en.wikipedia.org/wiki/Bollinger_Bands)
 
 The Bollinger bands includes three lines

stockstats.py

@@ -485,7 +485,10 @@ def _get_cci(self, window=None):
         """ Commodity Channel Index
 
         CCI = (Typical Price  -  20-period SMA of TP) / (.015 x Mean Deviation)
-        Typical Price (TP) = (High + Low + Close)/3
+        * when amount is not available:
+          Typical Price (TP) = (High + Low + Close)/3
+        * when amount is available:
+          Typical Price (TP) = Amount / Volume
         TP is also implemented as 'middle'.
 
         :param window: number of periods
@@ -786,6 +789,8 @@ def _shifted_cr_sma(self, cr, window):
         return self._shift(cr_sma, -int(window / 2.5 + 1))
 
     def _tp(self):
+        if 'amount' in self:
+            return self['amount'] / self['volume']
         return (self['close'] + self['high'] + self['low']).divide(3.0)
 
     def _get_tp(self):
@@ -1175,6 +1180,41 @@ def init_all(self):
         for handler in self.handler.values():
             handler()
 
+    def drop_column(self, names=None, inplace=False):
+        """ drop column by the name
+
+        multiple names can be supplied in a list
+        :return: StockDataFrame
+        """
+        if self.empty:
+            return self
+        ret = self.drop(names, axis=1, inplace=inplace)
+        if inplace is True:
+            return self
+        return wrap(ret)
+
+    def drop_tail(self, n, inplace=False):
+        """ drop n rows from the tail
+
+        :return: StockDataFrame
+        """
+        tail = self.tail(n).index
+        ret = self.drop(tail, inplace=inplace)
+        if inplace is True:
+            return self
+        return wrap(ret)
+
+    def drop_head(self, n, inplace=False):
+        """ drop n rows from the beginning
+
+        :return: StockDataFrame
+        """
+        head = self.head(n).index
+        ret = self.drop(head, inplace=inplace)
+        if inplace is True:
+            return self
+        return wrap(ret)
+
     @property
     def handler(self):
         return {
@@ -1272,7 +1312,8 @@ def within(self, start_date, end_date):
     def copy(self, deep=True):
         return wrap(super(StockDataFrame, self).copy(deep))
 
-    def _ensure_type(self, obj):
+    @staticmethod
+    def _ensure_type(obj):
         """ override the method in pandas, omit the check
 
         This patch is not the perfect way but could make the lib work.
@@ -1293,10 +1334,13 @@ def retype(value, index_column=None):
         if isinstance(value, StockDataFrame):
             return value
         elif isinstance(value, pd.DataFrame):
+            name = value.columns.name
             # use all lower case for column name
             value.columns = map(lambda c: c.lower(), value.columns)
 
             if index_column in value.columns:
                 value.set_index(index_column, inplace=True)
-            return StockDataFrame(value)
+            ret = StockDataFrame(value)
+            ret.columns.name = name
+            return ret
         return value

test.py

@@ -30,7 +30,8 @@
 
 import pandas as pd
 from hamcrest import greater_than, assert_that, equal_to, close_to, \
-    contains_exactly, none, is_not, raises, has_items, instance_of
+    contains_exactly, none, is_not, raises, has_items, instance_of, \
+    not_, has_item, has_length
 from numpy import isnan
 
 from stockstats import StockDataFrame as Sdf
@@ -48,6 +49,10 @@ def near_to(value):
     return close_to(value, 1e-3)
 
 
+def not_has(item):
+    return not_(has_item(item))
+
+
 class StockDataFrameTest(TestCase):
     _stock = wrap(pd.read_csv(get_file('987654.csv')))
     _supor = Sdf.retype(pd.read_csv(get_file('002032.csv')))
@@ -140,6 +145,14 @@ def test_middle(self):
         assert_that(middle.loc[idx], near_to(12.53))
         assert_that(tp.loc[idx], equal_to(middle.loc[idx]))
 
+    def test_typical_price_with_amount(self):
+        stock = self._supor[:20]
+        tp = stock['tp']
+        assert_that(tp[20040817], near_to(11.541))
+
+        middle = stock['middle']
+        assert_that(middle[20040817], near_to(11.541))
+
     def test_cr(self):
         stock = self.get_stock_90day()
         stock.get('cr')
@@ -465,6 +478,7 @@ def test_get_wr(self):
 
     def test_get_cci(self):
         stock = self._supor.within(20160701, 20160831)
+        stock.drop('amount', axis=1, inplace=True)
         stock.get('cci_14')
         stock.get('cci')
         assert_that(stock.loc[20160817, 'cci'], near_to(50))
@@ -640,3 +654,59 @@ def test_supertrend(self):
         assert_that(st[idx], near_to(12.9021))
         assert_that(st_ub[idx], near_to(14.6457))
         assert_that(st_lb[idx], near_to(12.9021))
+
+    def test_drop_column_inplace(self):
+        stock = self._supor[:20]
+        stock.columns.name = 'Luke'
+        ret = stock.drop_column(['open', 'close'], inplace=True)
+
+        assert_that(ret.columns.name, equal_to('Luke'))
+        assert_that(ret.keys(), has_items('high', 'low'))
+        assert_that(ret.keys(), not_has('open'))
+        assert_that(ret.keys(), not_has('close'))
+        assert_that(stock.keys(), has_items('high', 'low'))
+        assert_that(stock.keys(), not_has('open'))
+        assert_that(stock.keys(), not_has('close'))
+
+    def test_drop_column(self):
+        stock = self._supor[:20]
+        stock.columns.name = 'Luke'
+        ret = stock.drop_column(['open', 'close'])
+
+        assert_that(ret.columns.name, equal_to('Luke'))
+        assert_that(ret.keys(), has_items('high', 'low'))
+        assert_that(ret.keys(), not_has('open'))
+        assert_that(ret.keys(), not_has('close'))
+        assert_that(stock.keys(), has_items('high', 'low', 'open', 'close'))
+
+    def test_drop_head_inplace(self):
+        stock = self._supor[:20]
+        ret = stock.drop_head(9, inplace=True)
+        assert_that(ret, has_length(11))
+        assert_that(ret.iloc[0].name, equal_to(20040830))
+        assert_that(stock, has_length(11))
+        assert_that(stock.iloc[0].name, equal_to(20040830))
+
+    def test_drop_head(self):
+        stock = self._supor[:20]
+        ret = stock.drop_head(9)
+        assert_that(ret, has_length(11))
+        assert_that(ret.iloc[0].name, equal_to(20040830))
+        assert_that(stock, has_length(20))
+        assert_that(stock.iloc[0].name, equal_to(20040817))
+
+    def test_drop_tail_inplace(self):
+        stock = self._supor[:20]
+        ret = stock.drop_tail(9, inplace=True)
+        assert_that(ret, has_length(11))
+        assert_that(ret.iloc[-1].name, equal_to(20040831))
+        assert_that(stock, has_length(11))
+        assert_that(stock.iloc[-1].name, equal_to(20040831))
+
+    def test_drop_tail(self):
+        stock = self._supor[:20]
+        ret = stock.drop_tail(9)
+        assert_that(ret, has_length(11))
+        assert_that(ret.iloc[-1].name, equal_to(20040831))
+        assert_that(stock, has_length(20))
+        assert_that(stock.iloc[-1].name, equal_to(20040913))