[FR]: Pre-processor optimization

[jleaniz]

What is the feature you are proposing?

Evidence pre-processing runs every time a new job is executed. For some types of jobs, this is not necessary. As an example, a CompressedDirectory evidence would only need to be decompressed once for subsequent jobs to run. Currently, the pre-processor runs every single time.

What would this feature improve or what problem would it solve?

Task performance would be better, especially for larger, compressed evidence types.

What alternatives have you considered?

No response