Problem with strategy (all)

[hn2]

Which version are you running? The lastest version is on Github. Pip is for major releases.

import pandas_ta as ta
print(ta.version)

0.3.14b0

Do you have TA Lib also installed in your environment?

$ pip list

Yes
TA-Lib 0.4.17

Upgrade.

$ pip install -U git+https://github.com/twopirllc/pandas-ta

Describe the bug
When trying to add all indicators (ta.AllStrategy) to my dataframe I get:

Traceback (most recent call last):
  File "C:\Users\hanna\Anaconda3\lib\multiprocessing\pool.py", line 119, in worker
    result = (True, func(*args, **kwds))
  File "C:\Users\hanna\Anaconda3\lib\multiprocessing\pool.py", line 44, in mapstar
    return list(map(*args))
  File "C:\Users\hanna\Anaconda3\lib\site-packages\pandas_ta\core.py", line 467, in _mp_worker
    return getattr(self, method)(*args, **kwargs)
  File "C:\Users\hanna\Anaconda3\lib\site-packages\pandas_ta\core.py", line 874, in cdl_pattern
    result = cdl_pattern(open_=open_, high=high, low=low, close=close, name=name, offset=offset, **kwargs)
  File "C:\Users\hanna\Anaconda3\lib\site-packages\pandas_ta\candles\cdl_pattern.py", line 64, in cdl_pattern
    pattern_result = Series(pattern_func(open_, high, low, close, **kwargs) / 100 * scalar)
  File "_abstract.pxi", line 352, in talib._ta_lib.Function.__call__
  File "_abstract.pxi", line 383, in talib._ta_lib.Function.__call_function
  File "C:\Users\hanna\Anaconda3\lib\site-packages\talib\__init__.py", line 24, in wrapper
    return func(*args, **kwargs)

TypeError: Argument 'open' has incorrect type (expected numpy.ndarray, got NoneType)

As you can see it is probably due to cdl_pattern error.

When I switch to ta.CommonStrategy I get:

Index(['open', 'high', 'low', 'close', 'volume', 'SMA_10', 'SMA_20', 'SMA_50',
'SMA_200', 'VOL_SMA_20'],

I get 6 indicators(sma) added to my dataframe.
Is this the expected behavior? Only 6 common indicators?

[twopirllc]

Hello @hn2,

When I switch to ta.CommonStrategy I get:

Index(['open', 'high', 'low', 'close', 'volume', 'SMA_10', 'SMA_20', 'SMA_50',
'SMA_200', 'VOL_SMA_20'],

I get 6 indicators(sma) added to my dataframe.
Is this the expected behavior? Only 6 common indicators?

Yes. This is the correct behavior for the Common Strategy (as defined below). These are the indicators used the be financial news media (CNBC, Bloomberg, et al...) You are welcome to submit more Builtin Strategies if you want to share with the community.

pandas-ta/pandas_ta/core.py

Lines 92 to 110 in b92e45c

	# All Default Strategy
	AllStrategy = Strategy(
	name="All",
	description="All the indicators with their default settings. Pandas TA default.",
	ta=None,
	)
	# Default (Example) Strategy.
	CommonStrategy = Strategy(
	name="Common Price and Volume SMAs",
	description="Common Price SMAs: 10, 20, 50, 200 and Volume SMA: 20.",
	ta=[
	{"kind": "sma", "length": 10},
	{"kind": "sma", "length": 20},
	{"kind": "sma", "length": 50},
	{"kind": "sma", "length": 200},
	{"kind": "sma", "close": "volume", "length": 20, "prefix": "VOL"}
	]
	)

Could you please share reproducible code that was requested in the bug report?

According to the Stack Trace, it is not a Pandas TA problem. The problem has to do with your "open" data type: TypeError: Argument 'open' has incorrect type (expected numpy.ndarray, got NoneType). For some reason, your "open" data type is not like your "close" and "volume" data types (since those worked fine with the Common Strategy).

Kind Regards,
KJ

[hn2]

I believe that it has to do with the fact the pandas-ta candlestick patterns are build on top of ta-lib, and ta-lib is probably expecting a different data type. Could that be the reason for the error?

[hn2]

I am still not sure how to resolve it though. I tried converting to nparray but it didn't work.

[hn2]

class technical_features:
    def __init__(self,
                 security,
                 open_name,
                 high_name,
                 low_name,
                 close_name,
                 volume_name):

        open_price = security[open_name].values
        high_price = security[high_name].values
        low_price = security[low_name].values
        close_price = security[close_name].values

        security['open'] = np.array(open_price)
        security['high'] = np.array(high_price)
        security['low'] = np.array(low_price)
        security['close'] = np.array(close_price)

        print('Type')
        print(type(security['open']))

        
    def get_indicators_all(self):
       
        security.ta.strategy('all')

        
        print(security.columns)
        print(security.ta.categories)
        print(security)

        return security

[hn2]

The type is this:
Type
<class 'pandas.core.series.Series'>
for open and for close.

[twopirllc]

I believe that it has to do with the fact the pandas-ta candlestick patterns are build on top of ta-lib, and ta-lib is probably expecting a different data type. Could that be the reason for the error?

No. It's a Type Error as the stack trace says. Again, it is not a Pandas TA error.

Thanks for sharing "some" code this time. Please share your input data or how you read your data in? Also please provide a small csv sample (around 300 rows should do).

What type is security? A Pandas DataFrame or a dict? You are not clear on that in your code since there are no type hints. When calling: security.ta.strategy('all'), security needs to be a DataFrame.

Also, you shouldn't have to use either of these:

        open_price = security[open_name].values  # This converts to numpy values
        security['open'] = np.array(open_price) # This line is redundant

Do you have yfinance installed? (pip install yfinance) If so, you should be able to do the following:

import pandas as pd
import pandas_ta as ta

security = pd.DataFrame()
security = security.ta.ticker("aapl", period="1y")
security.ta.strategy(timed=True)

print(security)
print(security.columns)

Would be interested in your output with the code I provided.

Kind Regards,
KJ

[hn2]

You code did not run for me. Same error:
TypeError: Argument 'open' has incorrect type (expected numpy.ndarray, got NoneType)
I have an older pandas version (I need it to work with quantconnect).
pandas 0.25.3
numpy 1.19.5
TA-Lib 0.4.17

[twopirllc]

Screenshot?

[hn2]

[hn2]

[twopirllc]

Is there anything in security?

import pandas as pd
import pandas_ta as ta

security = pd.DataFrame()
security = security.ta.ticker("aapl", period="1y")
print(security)
print(security.dtypes)

[hn2]

I closed visual code and reopened it. Restarted the kernal. I don't believe that there is anything in security.
I also tried upgrading pandas to 1.1.5.
Same problem.

[twopirllc]

pip install yfinance

[twopirllc]

Also upgrade Pandas TA.

$ pip install -U git+https://github.com/twopirllc/pandas-ta

Has other bug fixes and new features. I am changing strategy() to study() but both will work.

[twopirllc]

Your local environment could be corrupted... it happens sometimes.

[hn2]

That wouldn't help.
I believe that the problem has to do with TA-Lib.
Are you sure that you are getting all 61 TA-Lib candlestick patterns?
My guess is that you are not.

[twopirllc]

I am sorry solutions I have presented have not solved your issue.

I believe that the problem has to do with TA-Lib.

I agree. I would recreate the local environment, usually solves most of my problems. Also time permitting, I am trying to implement numpy versions of TA Lib and hopefully the candles so TA Lib is optional.

Are you sure that you are getting all 61 TA-Lib candlestick patterns?

Yes, I am sure. In fact, 62.

KJ

[hn2]

What version of TA-Lib are you using?

[twopirllc]

[hn2]

Ok. I have an earlier version. I will try to install version 0.4.21 (right now it is giving me an error).
I wish that you could implement everything including candle patterns in pure python without TA-Lib which is a pain to install.

[twopirllc]

I wish that you could implement everything including candle patterns in pure python without TA-Lib which is a pain to install.

I know. That's partly the reason this package exists.

[hn2]

I upgraded my ta-lib. I now get this error:

in get_indicators_all(self)
81 # self.security.ta.strategy(ta.CommonStrategy)
82
---> 83 security.ta.strategy('all')
84
85 # # Use verbose if you want to make sure it is running.

~\Anaconda3\lib\site-packages\pandas_ta\core.py in strategy(self, *args, **kwargs)
790
791 # Apply prefixes/suffixes and appends indicator results to the DataFrame
--> 792 [self._post_process(r, **kwargs) for r in results]
793
794 if verbose:

~\Anaconda3\lib\site-packages\pandas_ta\core.py in (.0)
790
791 # Apply prefixes/suffixes and appends indicator results to the DataFrame
--> 792 [self._post_process(r, **kwargs) for r in results]
793
794 if verbose:

~\Anaconda3\lib\site-packages\tqdm\std.py in iter(self)
1178
1179 try:
-> 1180 for obj in iterable:
1181 yield obj
1182 # Update and possibly print the progressbar.

~\Anaconda3\lib\multiprocessing\pool.py in (.0)
318 result._set_length
319 ))
--> 320 return (item for chunk in result for item in chunk)
321
322 def imap_unordered(self, func, iterable, chunksize=1):

~\Anaconda3\lib\multiprocessing\pool.py in next(self, timeout)
733 if success:
734 return value
--> 735 raise value
736
737 next = next # XXX

Exception: inputs are all NaN

This is my data:

open high low close volume
0 1.151420 1.151915 1.150835 1.151040 0.0
1 1.151040 1.155060 1.145575 1.146480 0.0
2 1.146480 1.149190 1.143950 1.147080 0.0
3 1.147080 1.147700 1.137870 1.139860 0.0
4 1.139860 1.143285 1.135630 1.137515 0.0
.. ... ... ... ... ...
995 1.131090 1.133190 1.128465 1.129500 0.0
996 1.129515 1.136490 1.129005 1.136155 0.0
997 1.135980 1.136255 1.134955 1.135365 0.0
998 1.135385 1.135825 1.128505 1.133005 0.0
999 1.132985 1.137550 1.131305 1.136600 0.0

[hn2]

I guess that the dataframe needs datetimeindex?
Is there a workaround?

[twopirllc]

@hn2,

I guess that the dataframe needs datetimeindex?

No. The only indicator that requires it is: vwap.

How come you use time invariant data?

Is there a workaround?
Yes. You can exclude a list of indicators you don't want from strategy.

import pandas as pd
import pandas_ta as ta

security = pd.DataFrame()
security = security.ta.ticker("aapl", period="1y")

vwap_ = ["vwap"]  # 
security.ta.strategy(exclude=vwap_, timed=True)

print(security)
print(security.columns)

# help(security.ta.strategy)  # Documentation on the strategy method

---

Also, since you are using Windows and strategy (which uses multiprocessing), you might get a freeze_support error when creating a strategy Pinned Issue #181

---

Here are others ways to get help. I recommend reading Quick Start, Help and Issues.

KJ

[hn2]

Thank you

[hn2]

Is there a way to easily exclude the entire volume group of indicators?