Python module for loading time series data.
Project description
Python module for downloading time series data
This repository gets data from the Alpha Vantage API: https://www.alphavantage.co/. You need an API key in order to use the utilities in this repo.
You can retrieve price data from many stocks traded in stock markets.
Load Data
To use this module, you will need to have your own Alpha Vantage API key.
from loaders import loader
data_loader = loader.LoaderAlphaV(symbol="AAPL", interval=1)
df = data_loader.ts_intraday()
df_ext = data_loader.ts_intraday_extended(interval='5min&slice=year1month5')
Preprocess
This module is focused on producing data that is ingestible by pytorch models. The results is to return a Dataloader iteratior that can easily be used by the pytorch-framework. The preprocess module contains code for preprocessing of time series data.
The following steps are recommended for time series data:
- Split in train-valid-test sets
- Separate continous from categorical data
- Make sequential data
- Normalize data, where
X(t) = x(t -1 , ..., t - n); X(t - 1) = x(t - 2), ..., x(t - n); y(t) = x(t); y(t - 1) = x(t - 1)
, etc. - Transform to pytorch tensors.
There might be other options for the preprocessing that are equally well suited for your problem.
The class ContCatSplit
can be used on a financial time series. It can create date-features and then we can split data in time series and date features.
Packaging
For details on packaging and folder structure, see the link https://packaging.python.org/en/latest/tutorials/packaging-projects/
On windows, install latest version of pip:
py -m pip install --ugrade pip
Create distribution package, make sure you have latest PyPA's build installed
py -m pip install --upgrade build
Build
py -m build
You need to have Twine installed:
py -m pip install --upgrade twine
To publish on PyPi do the following:
twine upload dist/*
Install your newly created packege with:
python3 -m pip install [your-package]
For publishing in Test PyPI do the following:
py -m pip install --index-url https://test.pypi.org/simple/ --no-deps example-package-YOUR-USERNAME-HERE
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for financial_ts-0.1.39-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 40da50c5346c9ac05c5296bfebd407d25834aac6296feb1c2a8a5bb601e366e6 |
|
MD5 | 4746fd0d9a5b3c5fa201dc191b6cafb0 |
|
BLAKE2b-256 | 81303c11e8ddb09d82cb8c9de8e7633e25d074214e24a8be793f5eb333e52843 |