Make standalone scripts for collecting results

[percyfal]

Data output is now collected via rules, making heavy use of pandas. This functionality should also be present in standalone scripts, along the following lines:

collect_results.py --samples SAMPLE1 SAMPLE2... --input path/to/S1.csv path/to/S2.csv ... --application star --output path/to/output.csv

Either one script is made for every application, or one script manages all applications, as above - this assumes most output is similar enough that few modifications are necessary. Streaming input/output is a requirement.