The path of the directory assigned to a run should be relative to the project directory, and not an absolute path #5
Description
The hashmap takes hashes of run arguments as input and returns the directory in which the files of that run are found. At the moment, the path given as output is the absolute path of the results directory. However, it is preferable for this path to be saved as relative to the project directory for the following reasons:
- If the project directory happens to change its location or name, all runs would be automatically invalidated even if the relative position of the files within the project are kept the same. With a results' path relative to the project directory, retrieving results from the cache is invariant to changes in the project location or name
- In the case someone want to transfer the runs directory to another computer, it is not possible to use the saved data because all paths are saved relative to the root directory of the computer that run the experiments. If run paths are saved relative to the project directory, it is possible to retrieve results from other computers.