gzip result files

[rkaminsk]

Hi, I have a benchmark which were generating huge output files due to large models. What do you think about gzipping output files? Maybe one could still make this optional: provide to templates, one that gzips and one that does not. The resultparser could then simple use whatever is available the gzipped or the uncompressed version. Warning I did not test it yet.

[rkaminsk]

I think, I would just write a separate open function:

def my_open(path, file_name):
    path = os.path.join(path, file_name)
    gz_path = path + ".gz"
    if os.path.isfile(gz_path):
        return gzip.open(gz_path, errors="ignore", encoding="utf-8", mode="rt")
    return open(path, errors="ignore", encoding="utf-8", mode="rt")

[rkaminsk]

Maybe we should also make the gzip version the default in the templates and examples because it might safe a lot of disk space.

[tzschmidt]

I think we should keep the non gzip version as the default for now. I had the experience, that most people get to know the tool using smaller examples, where they want to manually inspect the result files.
Once someone is familiar with the tool, it shouldn't be an issue for them to change the used template.

[rkaminsk]

I just thought it would be the safer default. Since we had space trouble on the cluster in the past (where the tool is primarily used), this might improve the situation. I won't fight for making it the default though. Just my two cents.

[tzschmidt]

I will keep it in mind.