Closed shreykshah closed 6 months ago
This is indeed something we want to support, this was working at one point but seems to be broken currently (#1163).
If you're using the CLI, you can experiment with using the -m
or -M
flags: this allows you to ignore the most common fingerprints in the dataset you're analyzing (see dolos --help
for more info). But this does currently seem to have a bug.
We're planning on reworking this feature and integrating this in our front-end in the coming months however, so if you're using the web application, you will need to exercise some patience.
The -m
or -M
flags are not necessarily robust, where we have to guess what the best values for these flags would be given our template. Are there any plans for explicit template support?
Yes, template support will be a port of this rework, as well as integrating this in the web application.
I plan on using the CLI. Would there be any way to approximate this for a two student course in the meantime (perhaps submitting the template as a third student and setting -M 1
)?
We have added experimental support for ignoring template code and frequent fingerprints in PR https://github.com/dodona-edu/dolos/pull/1524
Is it possible to mark certain lines as a template such that sections of code are ignored? For example, if I have handed students some starter code to work with and asked them to modify small sections, the software similarity will come back as extremely high. I am trying to ensure the starter code is ignored.