The TTC 2021 OCL2PSQL Case

This is a case for the TTC 2021 on the translation of OCL queries to SQL, for a certain database schema (OCL2SQL). The case study is based on this reference implementation:

• Hoang Nguyen Phuoc Bao and Manuel Clavel, OCL2PSQL: An OCL-to-SQL Code-Generator for Model-Driven Engineering, FDSE 2019. (doi:10.1007/978-3-030-35653-8_13)

Case materials

The repository is structured as follows:

• docker:
  • Dockerfile build the MySQL image for running the benchmark.
  • init.sql initialize the database for running the benchmark.
• docs:
  • challenges.txt lists the various challenges (OCL queries) to be met by your transformation, organized by stage and challenge.
  • scenarios.txt runs some SQL queries over the various scenarios, showing what the expected results would be for them.
• metamodels includes several metamodels:
  • ocl.ecore has three EMF packages: one for data models (the database schemas), a simplified version of OCL expressions, and a simplified version of predefined OCL types.
  • sql.ecore has a metamodel for describing SQL SELECT queries.
• models includes three classes of models:
  • CarPerson.xmi describes the database schema for the sample Car-Person database.
  • StageXChallengeY.xmi has the XMI representation of the OCL query in the challenge Y of stage X.

Solution prerequisites

Reference solution

The reference solution in solutions/ReferenceXMI uses four JAR files that provide the OCL metamodel, SQL metamodel, OCL java representation, and the OCL2PSQL mapping. These JAR files are located in lib directory. You must install these dependencies into your local Maven repository before you build the reference solution. Assuming that Maven is in your PATH, you can run this Bash script:

cd solutions/ReferenceXMI
./install-jars.sh

Furthermore, the reference solution uses the MySQL database system built from the MySQL docker image in docker directory. In order to test the correctness of the reference solution, you must first build the MySQL container by using the following commands:

cd docker/
docker build --tag my-ttc2021 .
docker run -p 8083:3306 my-ttc2021

in which 8083 is the localhost port number that will be used to connect with the MySQL container. Note that, in case the port is busy, feel free to change it to another one, but remember to change it in the config.json as well.

Using the framework

The scripts directory contains the run.py script. At a first glance, invoke it without any arguments so that the solution will be built, benchmarked, running times visualized and the results compared to the reference solution's. One might fine tune the script for the following purposes:

• run.py -b -- builds the projects
• run.py -b -s -- builds the projects without testing
• run.py -m -- run the benchmark without building
• run.py -v -- visualizes the results of the latest benchmark
• run.py -c -- check results by comparing them to the reference output. The benchmark shall already been executed using -m.
• run.py -t -- build the project and run tests (usually unit tests as defined for the given solution)
• run.py -d -- runs the process in debug mode, i.e. with the Debug environment variable set to true

The config directory contains the configuration for the scripts:

• config.json -- configuration for the model generation and the benchmark
  • Note: the timeout as set in the benchmark configuration (default: 6000 seconds) applies to the gross cumulative runtime of the tool for a given changeset and update sequences. This also includes e.g. Initialization time which is not required by the benchmark framework to be measured. Timeout is only applied to the solutions' run stage (see -m for run.py), so it is not applied to e.g. the build stage (see -b for run.py).
• reporting.json -- configuration for the visualization

Running the benchmark

The script runs the benchmark for the given number of runs, for the specified tools and change sequences.

The benchmark results are stored in a CSV file. The header for the CSV file is stored in the output/header.csv file.

Reporting and visualization

Make sure you read the README.md file in the reporting directory and install all the requirements for R.

Implementing the benchmark for a new tool

To implement a tool, you need to create a new directory in the solutions directory and give it a suitable name.