This ML project is about inference for patients that are suspects for COVID-19 desease.
- Python
- Ludwig
- Flask
- Angular
https://github.com/criscojocaru/VintilaSiFormatia/wiki
- Python script will read the training data set and preprocess it
- After preprocessing, we will train the model using Ludwig
- The user accesses the website and loads the dataset
- Call from frontend towards backend is made to evaluate the dataset on the trained model
- Results are returned and showed in an easy to understand way
-
Install Python 3
-
Run the following commands in an elevated PowerShell window / CMD window:
python -m pip install --upgrade pip
pip install ludwig
- (Optional step) If you're using Windows, on build 2004 you have to run:
pip install numpy==1.19.3
- Also run the following commands:
python -m spacy download en
pip install flask
- Install Node.js, LTS. Make sure to also install npm and select Add to PATH
-
Navigate to /sources/Angular_UI in a PowerShell elevated prompt
-
Run the following command:
npm ci
- python .\sources\main.py train .\input\training.dataset.csv .\output\preprocessed.dataset.csv .\sources\model_definition_train.yaml
- python .\sources\main.py evaluate .\results\covid_inference_train\model .\input\testing.dataset.csv .\output\preprocessed-exp.dataset.csv .\sources\model_definition_evaluate.yaml
-
Navigate to the project's root in a PowerShell prompt
-
Run the following command:
python sources/server.py
-
Navigate to the /sources/Angular_UI folder in a PowerShell prompt
-
Run the following command:
npm start
- Access the website application at the link: http://localhost:4200/covid-inference
- Bengescu Paul, 341C5, Developer
- Ciudin Andreea-Alexandra, 341C5, Developer
- Cojocaru Cristina-Gabriela, 342C3, Team Leader
- Petrescu Ioana-Raluca, 341C5, Analyst
- Vintila Marius-Mihai, 341C5, Project Manager