xoelop

import datetime
import requests
import pandas as pd

def daily_price_historical(symbol, comparison_symbol, limit=0, aggregate=1, exchange='CCCAGG', print_url=False):
    """Returns a pandas.Dataframe containing OHLC daily data for the specified
    symbol

    Parameters
    ----------

xoelop

Atom Beautify - Debugging information

The following debugging information was generated by Atom Beautify on Mon Apr 16 2018 14:42:01 GMT+0800 (WITA).

---

Table Of Contents

• Versions
• Original file to be beautified
  • Original File Contents

xoelop

import numpy as np

# To substitute inf value by the max non inf value

a = np.array([1, np.inf, 10, -5, -np.inf])
print(a)
# [  1.  inf  10.  -5. -inf] 

a[np.isposinf(a)] = a[~np.isposinf(a)].max()
a[np.isneginf(a)] = a[~np.isneginf(a)].min()

xoelop

viridis_colorscale =  [
    [0.0, "#440154"],
    [0.0627450980392, "#48186a"],
    [0.125490196078, "#472d7b"],
    [0.188235294118, "#424086"],
    [0.250980392157, "#3b528b"], 
    [0.313725490196, "#33638d"], 
    [0.376470588235, "#2c728e"], 
    [0.439215686275, "#26828e"], 
    [0.501960784314, "#21918c"], 

xoelop

Figure({
    'data': [{'colorscale': 'Viridis',
              'hoverinfo': 'text+name',
              'name': 'Training set',
              'showscale': False,
              'text': array(['max_features: 0.05<br>n_estimators: 10<br>min_samples_split: 13<br>Test score: 0.8402<br>Train score: 0.9668<br>Fit time: 52.99s',
                             'max_features: 0.07<br>n_estimators: 10<br>min_samples_split: 13<br>Test score: 0.8845<br>Train score: 0.9715<br>Fit time: 74.24s',
                             'max_features: 0.09<br>n_estimators: 10<br>min_samples_split: 13<br>Test score: 0.8954<br>Train score: 0.9731<br>Fit time: 92.01s',
                             'max_features: 0.11<br>n_estimators: 10<br>min_samples_split: 13<br>Test score: 0.8851<br>Train score: 0.9758<br>Fit time: 116.09s',
                             'max_features: 0.13<br>n_estimators: 10<br>min_samples_split: 13<br>Test score: 0.8973<br>Train score: 0.9761<br>Fit time: 135.92s',

xoelop

model = RandomForestRegressor(n_jobs=-1, random_state=42, verbose=2)

grid = {'n_estimators': [10, 13, 18, 25, 33, 45, 60, 81, 110, 148, 200],
        'max_features': [0.05, 0.07, 0.09, 0.11, 0.13, 0.15, 0.17, 0.19, 0.21, 0.23, 0.25],
        'min_samples_split': [2, 3, 5, 8, 13, 20, 32, 50, 80, 126, 200]}

rf_gridsearch = GridSearchCV(estimator=model, param_grid=grid, n_jobs=4, 
                             cv=cv, verbose=2, return_train_score=True)

rf_gridsearch.fit(X1, y1)

xoelop

max_scores = df_gridsearch.groupby(['param_min_samples_split', 
                                    'param_max_features']).max()
max_scores = max_scores.unstack()[['mean_test_score', 'mean_train_score']]
sns.heatmap(max_scores.mean_test_score, annot=True, fmt='.4g');

xoelop

df_gridsearch['size'] = (df_gridsearch.mean_test_score / 
                         df_gridsearch.mean_test_score.max()) ** 100 * 20 + 1

xoelop

from tqdm import tqdm
from concurrent.futures import ThreadPoolExecutor, ProcessPoolExecutor, as_completed


def parallel_process(array,
                     function,
                     type_pool: str = 'multithreading',
                     use_kwargs=False,
                     n_jobs=16,
                     front_num=3,

xoelop

import pandas as pd
import numpy as np

def ratio(ret1: pd.Series,
          ret2: pd.Series = 0,
          ratio: str = 'sharpe',
          log: bool = True) -> float:
    """log: if True, convert ret1 and ret2 to log returns"""
    if log:
        ret1 = np.log1p(ret1)