Where can I find the pre-trained vectors?

[DaWe1992]

Hi Jordan,

first of all thank you for sharing this project.

I have a question:
Where can I find the file containing the pre-trained vectors?
The Readme for RQ1 states that there should be a file named 'vectors-gensim.txt' but I cannot find it.

Also, could you briefly summarize how I can generate my own embeddings? I am given a folder of C-programs and I would like to train the embeddings from scratch. I am aware that the input source files first have to be preprocessed to get the abstracted symbolic traces, but where do I have to put the files so that it works?

Thank you very much in advance. :)
Kind regards,
Daniel

[jjhenkel]

Hi Daniel,

Thanks for your interest! To get vectors-gensim.txt you can do the following (it's actually stored in the Docker image mentioned in the README---it was to big to check into the repository):

docker pull jjhenkel/code-vectors-artifact:rq1 && \
  docker run -it --name code-vectors-rq1 --entrypoint "echo" jjhenkel/code-vectors-artifact:rq1 && \
  docker cp code-vectors-rq1:/vectors/vectors-gensim.txt . && \
  docker rm code-vectors-rq1

One key point to consider for running this on your own input corpus is that the C-programs must be in state where you can actually compile them. This is required to generate the abstracted traces. So, if you can put those C-programs in a GitHub repo and build them with gcc (say, using a basic Makefile) then I can show you exactly what you'd need to do to ingest that new GitHub repo. (Essentially, you could just copy the portions of the Makefiles used to ingest redis and replace redis with your new repo.)

[DaWe1992]

Hi Jordan,

thank you very much for your reply. I will try it and let you know how it went. :)