Model for forecasting CAD-USD exchange price.
Project description
This package contains the pipeline of the ML model - forecast_model, the requirements, tests as well as other set up files.
You can install this package and use the forecast_model to predict the future price of CAD-USD exchange.
Find below the description of the different components & their respective uses:
A. forecast_model: This is the main module of this package. It contains a number of sub modules and files.
- config: Contains 'core.py' which is used to set the configuration for all the variables & file paths needed to run the package. The variables are listed in the config.yml file
- datasets: This contains the original dataset used to train the model
- processing: This contains other modules such as:
- data_manager: contains functions which can be used to load a dataset from the datasets module, and save, load or remove a trained model.
- preprocessing: contains functions necessary to transform raw data into the format expected by the trained model. Hence, this module is the preprocessing pipeline for this model.
- trained_models: This module contains the latest version of the trained forecast_model
- connfig.yml: Contains names of all the variables & file paths used in the model
- forecast.py: for forecasting CADUSD Price
- train_pipeline.py: for training the forecast_model
- VERSION: for setting the version of the model
B. Requirements: contains dependencies required to use or test the package C. tests: contains scripts for testing the package
- conftest.py - contains a fixture fxn which is used to provide forecast period to the other test.
- test_prediction.py - used to test the predition fucntion of forecast_model D. Manifest.in: contains instructions for what to include or exclude when building the package E. pyproject.toml: this file contains basic dependencies for setting up the package and also the configuration options for pytest. F. tox.ini: this file contains the settings for using tox for automated test but I didnt use tox to test this package.
Testing:
- Because my model using Prophet and prophet can not be installed without g++ compiler, I could not test this package with tox and so I used pytest on the cmd
How to use this package: **To run the different files and test the modules, I created a conda environment for this project, installed all the dependencies in the requirements.txt except prophet. and then followed the steps in this link to install prophet. https://stackoverflow.com/questions/53178281/installing-fbprophet-python-on-windows-10 ** If testing or using the package file on your computer, you have to add the directory to your PYTHONPATH, So that python can find it. Search for 'andrei' on https://stackoverflow.com/questions/3402168/permanently-add-a-directory-to-pythonpath . If you don't do this, python will not be able to find the package and so will not be able to run the import statements
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 Cad_usd_forecast_model-0.0.1.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | 6752277b02a0432b26bca031c56dcd08c666a969263666ca154506bd8222be46 |
|
MD5 | d48a527dcf17a525d474b55e8843b578 |
|
BLAKE2b-256 | e3eeac3a3f81f71656ee616247c76ce3bdb4028b158a179b69226c0ac89096e5 |
Hashes for Cad_usd_forecast_model-0.0.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | d2eac223733257a43f3dc634761013532f6618fe77ffb9cb399a35c878d34950 |
|
MD5 | d6b79d0824013518bb4a894874b26ed9 |
|
BLAKE2b-256 | 004147520f94e809f1a9b63befddce206b12f5e4b7582be7965c6e50ed345bd7 |