[tor-bugs] #2563 [Metrics]: Add R code for processing Torperf data to the Torperf repository

[Tor Bug Tracker & Wiki]

#2563: Add R code for processing Torperf data to the Torperf repository
-------------------------+--------------------------------------------------
 Reporter:  karsten      |          Owner:  karsten
     Type:  enhancement  |         Status:  new    
 Priority:  normal       |      Milestone:         
Component:  Metrics      |        Version:         
 Keywords:               |         Parent:         
   Points:               |   Actualpoints:         
-------------------------+--------------------------------------------------

Comment(by karsten):

 This tasks contains three subtasks:
  1. Generalize filter.R from the #2543 code, so that it can be run without
 modifying the source.  This includes:
     - Pass the .data files to process as command-line arguments to the R
 script.  This script would be called by `Rscript filter.R *.data` or
 `Rscript filter.R first.data second.data`, and the arguments would be
 extracted using `args <- commandArgs(trailingOnly = TRUE)`.  If the script
 doesn't have any command-line arguments, it could simply attempt to read
 all files in the current directory ending with `.data`.
     - Change the read function to extract the `guards`, `filesize`, and
 `bytes` parameters from the filename.  For example, if we parse `torperf-
 50kb.data`, `guards` could be `torperf`, and `filesize` would be `50kb`;
 `bytes` is a bit more complicated, because we need to parse the int value
 out of `50kb` and need to know that `kb` means times 1024.
     - Remove the part that cuts off observations before and after hard-
 coded dates and have the user do this particular filter step manually.
  2. Generalize timematrix.R from the #2543 code.  I think we can use it
 unchanged, but someone should confirm that.  I don't think we can make
 this file configurable from the command line easily.  Not sure if we
 should try.
  3. Modify the #2543 HOWTO to remove the #1919 specifics.

 Tom, after giving you the quick R overview on this code last Saturday,
 would you like to tackle this?

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/2563#comment:1>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online