Table of Contents

• Table of Contents
• Category Transformation
  • How to use it
    • Requirements
    • Download and install
    • Execution

Category Transformation

This repository provides support for transforming feature models (FM) into category theory.

It offers the following functionality:

• Generation of feature's attributes for an FM.
• Generation of sample of configurations from an FM.
• Generation of variant-wise attributes for configurations.
• Transformation from FM and configurations to category theory.

How to use it

Requirements

• Linux
• Python 3.9+
• Flama

Download and install

1. Install Python 3.9+

2. Clone this repository:

   git clone https://github.com/CAOSD-group/cat-transf.git

3. Create a virtual environment and activate it:

   python -m venv env

   . env/bin/activate

4. Install the dependencies:

   pip install -r requirements.txt

Execution

1. Go to the project folder: cd cat-transf

2. Run the appropriate script according to the required functionality. Normally, you should follow these steps:

• Step 1: Generate feature attribute: Given an FM, it generates an attribute for each feature in the model with random values in a domain.

  • Execution: python gen_feat_attr.py fm attribute_name -r ranges -e elements (-l)
  • Inputs:
    • fm specifies the file path of the FM in UVL format (.uvl).
    • attribute_name is the name of the new attribute.
    • The domain can be specified with one or both of the following parameters:
      • ranges specifies the domain for the values as a list of ranges with the following format: '[[1, 10], [90, 100]]' generates values from 1 to 10 and from 90 to 100.
      • elements specifies the domain for the values as a list of elements with the following format: '[True, False]' for a Boolean attribute; '["Low", "Medium", "High"]' for a string enumeration; '[1, 10.0, 100]' for specific integer or real values.
    • -l is an optional parameters indicating that the attribute will be only generated for the leaf features in the model.
  • Outputs:
    • A FM file in UVL format with the new attribute created in each feature. The output file is named as the original one ending with _X.uvl where X is the new attribute name.
  • Example: python gen_feat_attr.py Pizzas.uvl Price -r '[[1, 10]]' -l
  • Note: If there is a feature with an already existing attribute with the same name, the attribute will not be created for that feature.

• Step 2: Generate configurations: Given an FM, it generates a sample of configurations.

  • Execution: python gen_configs.py fm -s size
  • Inputs:
    • fm specifies the file path of the FM in UVL format (.uvl).
    • size is the number of configuration to be generated.
  • Outputs:
    • A csv file with the configurations. The output file is named as the original FM with _S.csv where S is the number of configurations.
  • Example: python gen_configs.py Pizzas.uvl -s 10

• Step 3: Generate variant-wise attribute: Given the configurations of an FM in a csv file, it generates an attribute for each configuration with random values in a domain.

  • Execution: python gen_config_attr.py csv attribute_name -r ranges -e elements (-l)
  • Inputs:
    • csv specifies the file path of the csv with the configurations.
    • attribute_name is the name of the new attribute.
    • The domain can be specified with one or both of the following parameters:
      • ranges specifies the domain for the values as a list of ranges with the following format: '[[1, 10], [90, 100]]' generates values from 1 to 10 and from 90 to 100.
      • elements specifies the domain for the values as a list of elements with the following format: '[True, False]' for a Boolean attribute; '["Low", "Medium", "High"]' for a string enumeration; '[1, 10.0, 100]' for specific integer or real values.
  • Outputs:
    • A csv file with the configurations and a new attribute created for each configuration. The output file is named as the original one ending with _X.csv where X is the new attribute name.
  • Example: python gen_config_attr.py Pizzas.csv Vegan -e '[True, False]'

• Step 4: Transform to category theory: Given an FM and a sample of configurations, it generates the category theory model (.cql).

  • Execution: python fm2cql.py fm csv
  • Inputs:
    • fm specifies the file path of the FM in UVL format (.uvl).
    • csv specifies the file path of the csv with the configurations.
  • Outputs:
    • A cql file with the category theory model. The output file is named as the original FM with with .cql extension.
  • Example: python fm2cql.py Pizzas.uvl Pizzas.csv