UBUMLaaS

---

This application is the continuation of UBUMLaaS developed by the ADMIRABLE research group of the University of Burgos.

This new version provides support for Semi-Supervised Learning algorithms and new instance selection filters.

It also has a renewed and modernized interface. With a whole new administration section of the application itself, with statistics for users and general system statistics.

Machine Learning as a Service (MLaaS) platform based on ADMIRABLE and BEST-AI research groups methods.

---

Installation (Linux)

1. Clone this repository

   $ git clone https://github.com/dpr1005/UBUMLaaS.git

2. Go to UBUMLaaS repository's folder

   $ cd UBUMLaaS

3. Create a conda environment

   $ conda env create -f UBUMLaaS_env.yml

4. Activate environment

   $ conda activate UBUMLaaS

5. Modify env_variables.sh with properly values

   export SECRET_KEY=<app secret key>
   export EMAIL_AC=<email>
   export EMAIL_PASS=<email-password>
   export EMAIL_URL=<email-url>
   export FLASK_ENV=development #development or production
   LIBFOLDER=/absolute/path/to/UBUMLaaS

6. With the conda environment UBUMLaaS, execute the script to export environment variables when activate conda env.

   $ source env_vars_to_conda.sh

7. Create database

   $ mv data_base.sqlite ubumlaas/data.sqlite

   Opt: Download a database and put it in ./ubumlaas/
8. Install Redis-Server

   $ sudo apt install redis-server
   $ sudo service redis-server start
   $ sudo systemctl enable redis-server #If you want to initialize the service in startup

   Caution: Close all workers of RQ before stop redis-server

Execution

1. Inside the UBUMLaaS repository's folder, activate conda environment if not activated.

   $ conda activate UBUMLaaS

2. Execute to run the server

   ./run.sh

   Remember to give the required permissions in case they are needed.

Update database

1. Execute migrate.py

   $ python migrate.py

---

GitHub @dpr1005  ·  Twitter @callmednx  ·  LinkedIn Daniel Puente Ramírez