As trendet is currently just intended to be used combined with investpy, in order to extend its usage, coverage to identify trends from any time-series pandas.DataFrame is proposed.

So on, an additional function is proposed where the pandas.DataFrame column should be specified as the data is not structured as the one used by investpy. This will lead to extend trendet usage so that it can be used for every time-series dataset.

Really appreciate for your great and wonderful efforts on developing "trendet".
I'm exploring the usage of it on my strategy.
May I clarify one thing, will trendet "peep" the next data point (forward looking)?
Is the window size for backward looking (t-) instead of forward looking (t+)?
Thank you!

Hi I was analysing your lib
an look like good idea
but I could not install it on my pc or even on google colabs

I tried
pip install trendet==0.6
and also
pip install trendet --upgrade

On collab:

on local:

line 14, in requirements
with io.open('requirements.txt', encoding='utf-8') as f:
FileNotFoundError: [Errno 2] No such file or directory: 'requirements.txt'
----------------------------------------
ERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.

Hi,
I've tried to run trendet script but it gives me an attribute error:
AttributeError: partially initialized module 'trendet' has no attribute 'identify_all_trends' (most likely due to a circular import)
Do you have any idea what that could be?

pip install trendet Collecting trendet Downloading trendet-0.6.tar.gz (10 kB) ERROR: Command errored out with exit status 1: command: 'd:\programas\python37\python.exe' -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'C:\\Users\\x\\AppData\\Local\\Temp\\pip-install-bq8c2po4\\trendet\\setup.py'"'"'; __file__='"'"'C:\\Users\\x\\AppData\\Local\\Temp\\pip-install-bq8c2po4\\trendet\\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' egg_info --egg-base 'C:\Users\x\AppData\Local\Temp\pip-pip-egg-info-tk2aj33h' cwd: C:\Users\x\AppData\Local\Temp\pip-install-bq8c2po4\trendet\ Complete output (7 lines): Traceback (most recent call last): File "<string>", line 1, in <module> File "C:\Users\x\AppData\Local\Temp\pip-install-bq8c2po4\trendet\setup.py", line 32, in <module> install_requires=requirements(), File "C:\Users\x\AppData\Local\Temp\pip-install-bq8c2po4\trendet\setup.py", line 14, in requirements with io.open('requirements.txt', encoding='utf-8') as f: FileNotFoundError: [Errno 2] No such file or directory: 'requirements.txt' ---------------------------------------- ERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.

Simillar to alvarobartt/investpy#85

I have tried:

• python -m pip install trendet
• python -m pip install trendet==0.6
• python -m pip install trendet==0.6 --upgrade --user
• python -m pip install trendet-0.6.tar.gz

I also tried to pre-install the requirements
sphinx==2.4.4 sphinx_rtd_theme==0.4.3 recommonmark==0.6.0

As previous build referenced in commit 5ab3e82 failed due to codecov configuration, the addition of .codecov.yml file, in order to avoid build failing when it comes to code coverage even when the CI build succeeded, is proposed.

Reference: https://docs.codecov.io/docs/commit-status

Hello,
I'm trying to understand the logic of your software, can you explain how you set the trends? Why use the mean() and how do you append the up_trends and down_trends lists.
Thank you for your response and your work,
Pascal

Create a function to check that trends do not overlap, so the trend which last longer is the one to keep and the otherone is discarded.

For example, inside of a Down Trend between 01/01/2018 and 01/02/2018 there is an Up Trend between 05/01/2018 and 13/01/2018; and as the Down Trend lasts longer, it is expected to be the predominant trend on that market range of dates.

This issue was identified when developing the function identify_all_trends, where an Up Trend was overlapped inside a Down Trend, as shown below:

in the function "identify_df_trends" the parameter "column" is useless. The algorithm always uses the "Close" column regardless of the input parameter

look it here

I tried to use AAPL data to define trend for the last two years.
I set up window_size = 5 it gave, an error below. I changed it to 10, it ran ok. Then, changed the ticker to AMZAN it gave me the error for 5 and 10.

Please, let me know if you have any suggestions

AttributeError Traceback (most recent call last)
File G:\My Drive\development\swing\main.py:392
385 print('script is completed')
390 if name == 'main':
--> 392 main()
402 # WMA https://corporatefinanceinstitute.com/resources/career-map/sell-side/capital-markets/weighted-moving-average-wma/

File G:\My Drive\development\swing\main.py:284, in main()
280 df['day'] = df['timestamp'].dt.strftime('%Y-%m-%d')
281 # trend_df = alpaca_paral_data_col([ticker], start_date_ms, end_date_ms, timespan = 'day', multiplier=1)
282 # print(f'{ticker} i = {i} trend_df\n', trend_df.head())
283 # trend_df = sf.trend_define(ticker, start_date_ms, end_date_ms)
--> 284 trend_df = sf.trend_define_trendet(ticker, start_date_ms, end_date_ms)
285 print(f'{ticker} i = {i} trend_df\n', trend_df.head())
286 df = df.merge(trend_df[['ticker', 'day', 'trend']], on=['ticker', 'day'], how='inner')

File G:\My Drive\development\swing\swing_functions.py:197, in trend_define_trendet(ticker, start_date_ms, end_date_ms)
195 trend_df = alpaca_paral_data_col([ticker], start_date_ms, end_date_ms, timespan = 'day', multiplier=1)
196 trend_df['close'] = trend_df['close'].astype('float64')
--> 197 trend_df = trendet.identify_df_trends(df=trend_df,
198 column='close',
199 window_size=5,
200 identify='both')#.rename(columns={'Up Trend': 'trend'})
201 trend_df['trend'] = trend_df['Up Trend'].notnull().replace({True: 'uptrend', False: None})
202 trend_df['Down Trend'] = trend_df['Down Trend'].notnull().replace({True: 'downtrend', False: None})

File ~\Anaconda3\envs\VENV_SWING_P3_9_11\lib\site-packages\trendet\identification.py:589, in identify_df_trends(df, column, window_size, identify)
587 for down in results['Down Trend']:
588 if down['from'] < up['from'] < down['to'] or down['from'] < up['to'] < down['to']:
--> 589 if (up['to'] - up['from']).days > (down['to'] - down['from']).days:
590 flag = True
591 else:

AttributeError: 'int' object has no attribute 'days'

Stock values doesn't fluctuate on the weekends, therefore they should not be included in the trends plot. It distorts the output as some trends seem to be more consistent than others, just because they contain a weekend (or several of them).

I love the description of the module - but I am having problems implementing this example.

https://trendet.readthedocs.io/usage.html#identify-trends-of-custom-dataframe

As you can see above - the bold code is of my own making. I'm taking a column, predicted_stock_data[:,0], from a dataframe to use a the closing price. I then add the 'Date'. I can't grasp the problem with 'Date'.

res = predicted_stock_data[:,0]
des = pd.DataFrame(res, columns = ['Close'])
start = datetime(2018, 9, 29)
dates = pd.date_range(start, periods=100).to_pydatetime().tolist()
des['Dates'] = dates
res = trendet.identify_df_trends(df=des, column='Close')

with plt.style.context('classic'):
plt.figure(figsize=(20, 10))
ax = sns.lineplot(x=res['Date'], y=res['Close'])

labels = res['Up Trend'].dropna().unique().tolist()

for label in labels:
    sns.lineplot(x=res[res['Up Trend'] == label]['Date'],
                 y=res[res['Up Trend'] == label]['Close'],
                 color='green')

    ax.axvspan(res[res['Up Trend'] == label]['Date'].iloc[0],
               res[res['Up Trend'] == label]['Date'].iloc[-1],
               alpha=0.2,
               color='green')

labels = res['Down Trend'].dropna().unique().tolist()

for label in labels:
    sns.lineplot(x=res[res['Down Trend'] == label]['Date'],
                 y=res[res['Down Trend'] == label]['Close'],
                 color='red')

    ax.axvspan(res[res['Down Trend'] == label]['Date'].iloc[0],
               res[res['Down Trend'] == label]['Date'].iloc[-1],
               alpha=0.2,
               color='red')

plt.show()

As trendet has been updated the update of trendet gist is proposed, since country param has been added to identify_trends() and identify_all_trends(), and a new function has been added identify_df_trends().

Documentation needs to be checked and updated before releasing trendet 0.5.

Hi,
I'm using this great python library.
I was wondering why I come up with this error:
'int' object has no attribute 'days'
my code is here:

import trendet

import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd

sns.set(style='darkgrid')

dff = pd.read_csv('setran.csv')


df = trendet.identify_df_trends(df=dff , column='Close')


print(df['Up Trend'].values)
df.reset_index(inplace=True)

plt.figure(figsize=(20, 10))

ax = sns.lineplot(x=df.index, y=df['Close'])
ax.set(xlabel='Date')

labels = df['Up Trend'].dropna().unique().tolist()

for label in labels:
    sns.lineplot(x=df[df['Up Trend'] == label].index,
                 y=df[df['Up Trend'] == label]['Close'],
                 color='green')

    ax.axvspan(df[df['Up Trend'] == label].index[0],
               df[df['Up Trend'] == label].index[-1],
               alpha=0.2,
               color='green')

labels = df['Down Trend'].dropna().unique().tolist()

for label in labels:
    sns.lineplot(x=df[df['Down Trend'] == label].index,
                 y=df[df['Down Trend'] == label]['Close'],
                 color='red')

    ax.axvspan(df[df['Down Trend'] == label].index[0],
               df[df['Down Trend'] == label].index[-1],
               alpha=0.2,
               color='red')
               
locs, _ = plt.xticks()
labels = []

for position in locs[1:-1]:
    labels.append(str(df['Date'].loc[position])[:-9])

plt.xticks(locs[1:-1], labels)
plt.show()

Can you please answer me to solve this issue (or just my problem) ?

Since in the last trendet commit 5ab3e82 a new function has been created in order to allow trendet usage for any custom pandas.DataFrame object, the pytest need to be updated in order to fix code coverage.

What makes an algorithm like this attractive is being able to identify the current trend of the market.
However, in all tests I've done, the algorithm ignores the final intervals.
Below I show you the result of the algorithm with smoothed data, to further accentuate the phenomenon that I mentioned.