Optimizer not working locally

dark.yellowjacket

optimizer does not work on my local platform.

whether i set workers to 1 or to os.cpu_count which is 20.

dark.yellowjacket

additionally
the following complaint

The environment variable DATA_BASE_URL was not specified. The default value is 'https://data-api.quantiacs.io/'
NOTICE: The environment variable CACHE_RETENTION was not specified. The default value is '7'
NOTICE: The environment variable CACHE_DIR was not specified. The default value is 'data-cache'

dark.yellowjacket

This post is deleted!

dark.yellowjacket

additional information after trying to optimize THIS IS ON WINDOWS

KeyboardInterrupt Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_9644\3607707927.py in <module>
6 strategy_long,
7 qno.full_range_args_generator(ma_period=range(10, 200, 10)),
----> 8 workers = 20 #cpu_count # you can set more workers on your local PC to speed up default =1
9 )
10 end = time.time()

~\anaconda3\envs\qntdev\lib\site-packages\qnt\optimizer.py in optimize_strategy(data, output_function, argument_generator, stats_function, stats_to_weight, workers)
38 argument_generator,
39 stats_to_weight,
---> 40 workers
41 )
42

~\anaconda3\envs\qntdev\lib\site-packages\qnt\optimizer.py in optimize(target_function, argument_generator, result_to_weight, workers)
61 for w in workers:
62 w.daemon = True
---> 63 w.start()
64
65 with progressbar.ProgressBar(max_value=length, poll_interval=1) as bar:

~\anaconda3\envs\qntdev\lib\multiprocessing\process.py in start(self)
110 'daemonic processes are not allowed to have children'
111 _cleanup()
--> 112 self._popen = self._Popen(self)
113 self._sentinel = self._popen.sentinel
114 # Avoid a refcycle if the target function holds an indirect

~\anaconda3\envs\qntdev\lib\multiprocessing\context.py in _Popen(process_obj)
221 @staticmethod
222 def _Popen(process_obj):
--> 223 return _default_context.get_context().Process._Popen(process_obj)
224
225 class DefaultContext(BaseContext):

~\anaconda3\envs\qntdev\lib\multiprocessing\context.py in _Popen(process_obj)
320 def _Popen(process_obj):
321 from .popen_spawn_win32 import Popen
--> 322 return Popen(process_obj)
323
324 class SpawnContext(BaseContext):

~\anaconda3\envs\qntdev\lib\multiprocessing\popen_spawn_win32.py in init(self, process_obj)
87 try:
88 reduction.dump(prep_data, to_child)
---> 89 reduction.dump(process_obj, to_child)
90 finally:
91 set_spawning_popen(None)

~\anaconda3\envs\qntdev\lib\multiprocessing\reduction.py in dump(obj, file, protocol)
58 def dump(obj, file, protocol=None):
59 '''Replacement for pickle.dump() using ForkingPickler.'''
---> 60 ForkingPickler(file, protocol).dump(obj)
61
62 #

KeyboardInterrupt:

support

Hello, sorry for the basic question, did you have some warnings when using the toolbox locally (no optimizer)?

dark.yellowjacket

@support
No just the complaint indicated on my second comment.

support

@dark-yellowjacket Ok, thanks. We are aware of optimizer issues. Please see also:

https://quantiacs.com/community/topic/249/strategy-optimization-in-local-development-environment-is-not-working

https://quantiacs.com/community/topic/233/local-development-error-could-not-convert-string-to-float-nas

We are working on that. Sorry for the problems.

dark.yellowjacket

@support
Hello I resolved the issue

I moved the function and copied pasted the appropriate import s to a ,py file and named it strategy .py
i commented out the strategy_long(...) function from my notebook
then :... inserted in the notebook: from strategy import * before running the optimizer.

4 ) and it works FYI for asset by asset 20 cores about 2,5 times faster counting wall time

NOTE : I think this procedure gets around Windows Python GIL. however, it does not appear to be explained in the python Docs

support

@dark-yellowjacket Thanks a lot for posting the solution!

captain.nidoran

Thanks @dark-yellowjacket for post your solution for the optimizer, it works fine for me but could you please go deeper explaining how do you implement the optimized obtained parameter values in the final strategy. I have tried several methods (even in several computers) , obtaining always errors.
For example using this template provided by Quantiacs (where i call to "Strategy" I use the same previouslly optimized, and the config.json is the output asset by asset obtained after optimization):


import json
import xarray as xr
import qnt.backtester as qnbk
from Strategy import *

def optmized_strategy(data, config):
    results = []
    for c in config:
        results.append(strategy_long(data, **c))
    # align and join results
    results = xr.align(*results, join='outer')
    results = [r.fillna(0) for r in results]
    output = sum(results) / len(results)
    return output


config = json.load(open('config.json', 'r'))

# multi-pass
# It may look slow, but it is ok. The evaluator will run only one iteration per day.
qnbk.backtest(
    competition_type='stocks_nasdaq100',
    lookback_period=365,
    strategy=lambda d: optmized_strategy(d, config),
 # strategy=strategy_long, # you can check the base strategy too
    start_date='2006-01-01')

It rises the following error:

Reloaded modules: Estrategia
fetched chunk 1/5 0s
fetched chunk 2/5 0s
fetched chunk 3/5 0s
fetched chunk 4/5 0s
fetched chunk 5/5 0s
Data loaded 1s
Run last pass...
Load data...
fetched chunk 1/1 0s
Data loaded 0s
Run strategy...
Traceback (most recent call last):

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\spyder_kernels\py3compat.py", line 356, in compat_exec
    exec(code, globals, locals)

  File "c:\users\luispc\desktop\quantiacs\q18\prueba.py", line 39, in <module>
    start_date='2006-01-01')

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\qnt\backtester.py", line 291, in backtest
    result = strategy_wrap(data, state)

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\qnt\backtester.py", line 270, in <lambda>
    strategy_wrap = (lambda d, s: strategy(d)) if args_count < 2 else strategy

  File "c:\users\luispc\desktop\quantiacs\q18\prueba.py", line 37, in <lambda>
    strategy=lambda d: optmized_strategy(d, config),

  File "c:\users\luispc\desktop\quantiacs\q18\prueba.py", line 22, in optmized_strategy
    results.append(strategy_long(data, **c))

  File "C:\Users\LuisPC\Desktop\quantiacs\Q18\Estrategia.py", line 16, in strategy_long
    data = data.sel(asset=[asset])

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\xarray\core\dataarray.py", line 1337, in sel
    **indexers_kwargs,

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\xarray\core\dataset.py", line 2505, in sel
    self, indexers=indexers, method=method, tolerance=tolerance

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\xarray\core\coordinates.py", line 422, in remap_label_indexers
    obj, v_indexers, method=method, tolerance=tolerance

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\xarray\core\indexing.py", line 120, in remap_label_indexers
    idxr, new_idx = index.query(labels, method=method, tolerance=tolerance)

  File "C:\Users\LuisPC\.conda\envs\qntdev\lib\site-packages\xarray\core\indexes.py", line 242, in query
    raise KeyError(f"not all values found in index {coord_name!r}")

KeyError: "not all values found in index 'asset'"

I tried to explain this problem about the implementation in the Backtester in another forum thread time ago, without getting a valid answer, therefore I would appreciate any idea you can give me on this matter.
...Otherwise I am afraid that I will not present algorithms for this contest (Sadly )
Regards.
Luis G.