How to resume a download?

[Thomas-Frew]

Hello everyone!

I don't have the most stable network connection, and I was wondering if it was possible to resume a data download for a large dataset if the connection is interrupted.

If not, do you have any tips to minimise losses (like downloading in stages or creating some sort of checkpoint system?)

Thanks!

[JerBouma]

Hi @Thomas-Frew,

The Toolkit class gives you the ability to provide a "custom" dataset for both historical as well as fundamental data. So as an example you can do the following:

from financetoolkit import Toolkit
import pandas as pd

ticker_list = ['AMZN', 'AAPL', 'TSLA', 'MU', 'GOOG', 'TSM']

# Only picking the first two as an example
companies = Toolkit(tickers=ticker_list[:2], api_key="API_KEY")

# Load historical data and save to a variable
historical_data = companies.get_historical_data()

# Write to Pickle
historical_data.to_pickle("saved_stocks.pickle")

# Load the Pickle (in the case your current instance is erased)
loaded_data = pd.read_pickle('saved_stocks.pickle')

# Reinitialise the Toolkit with the historical data
companies_2 = Toolkit(tickers=ticker_list[:2], api_key="API_KEY", historical=loaded_data)

# Show bollinger bands as an illustration of the capability
companies_2.technicals.get_bollinger_bands()

This gets you the following without the requirement to download any of the data again:

Therefore, if you create a loop over your tickers, write to pickle and when you got everything combine the pickles, you get what you are looking for. As an example:

Working with custom datasets is a little bit experimental however. I believe it works in pretty much every area but feel free to let me know if you run into issues.

Note: I disabled the benchmark just to prevent it from showing up every single time. It should be enough to get the Benchmark data once to make functions like Jensen's Alpha work.

[JerBouma]

Hi @Thomas-Frew, did this fix your problem?

[Thomas-Frew]

Thanks for your help! This was very useful.