Week 10(16th August — 23rdAugust)

Goals —

• Test integration of RHA on local computer with the web interface.
• Containerize the RHA
• Schedule a test batch job

Progress —

• This week began with the testing of RHA integration with the web-app using the background_runner class.
  I was able to successfully use the hider feature through the web-interface on my local computer. Testing of swapper was not possible because of limited compute resources available on my local host.For testing the swapper we will have to use the hpc.
• Executing code on the HPC (High-Performance Computing) infrastructure differs slightly from running it on a local machine. When using the HPC, job scheduling becomes necessary. As a result, some adjustments are needed in our code to enable automatic scheduling of batch jobs on the HPC environment.Keeping this in mind I started working on integration of RHA with the singularity container. This was the most challenging task of this week. I aimed to integrate the RHA in a manner similar to the way the lolcow integration has been done.
• I setup a singularity sandbox by using the previously created .sif file, which contains all the essential dependencies for running the RHA. Subsequently, I generated a directory named “rha” within “usr” and relocated all the RHA files to this location.
• After this, I proceeded to customize the singularity executable located in the root directory. This executable is executed while executing a singularity container. It’s important to note that even when operating outside the container, access to this executable remains possible. However, it’s important to understand that running this executable doesn’t equate to executing the entire singularity container. The execution of the singularity container has a lot of processes, and this script execution is just one of the process of that.
• Finally I got a complete singularity container. This container can be executed with different parameters to perform either hider anonymization or swapper anonymization.
• I have tested this container for hider anonymization and it is working fine. To test the swapper functionality, we have to transfer the container to the HPC , which is currently in progress. Despite my reliable internet connection, the file transfer process is quite time-consuming due to the container’s size of approximately 12 GBs after compression. I also tried with Globus as a transfer method, but it proved slower compared to using scp.
• This past week, the majority of my efforts were devoted to configuring the container and transferring the files between HPC and my laptop. As a result, I didn’t manage to schedule a test batch job. Moving forward, my focus for the upcoming week will be to to ensure the integration of the swapper and to schedule a test batch job.