MLSToolbox is composed by the following components that are included in this repository as folders.
| Component | Description |
|---|---|
| mls_toolbox_client | An Angular-based component for displaying a node-based editor to define ML pipelines |
| mls_toolbox_server | A Flask-based component for redirecting the mls_toolbox_client requests to the services provided by the mls_code_generator |
| mls_code_generator | A Python component for mainly generating Python code for the ML pipelines represented in the editor |
| mls_code_generator/mls_lib | A Python library containing object classes, used in the generated code, representing the structure and the main concepts required to instantiate any pipeline, its stages and the tasks that these stages perform |
| mls_code_generator/mls_code_generator_config | A component containing several extensible JSON files that define the graphical elements of the graphical editor representing the predefined steps of a ML pipeline |
You can find all the information you need in our WIKI!.
The technical documentation about the mls_lib is available at the following link.
This video shows how to use the MLSToolbox Pipeline Code generator to define a pipeline and generate the code to generate the model. More details about the example used in this video are available at mls_code_generator Wiki.
Pipeline.Code.Generator.Diabetes_576.mp4
You can find more videos at mls_code_generator Wiki.
Each of the folder contains their own docker_build.shand docker_run.sh files. Once both are executed, the system should be up and running!
Remember that the repository is composed of multiple repositories. After cloning the repository the following command needs to be executed:
git submodule update --init --recursive You can find the MLSToolbox Contribution Guidelines here.
You can find the MLSToolbox Code of Conduct here