__________________________________
ABOUT HYBRID_VM_2.0
This Hybrid_VM_2.0 provides three major fault localization techniques for a 
collection of 74 errors from four open source C projects. 
These techniques include:
- Dynamic Slicing
- Statistical Debugging (Ochiai, Jaccard & Tarantula)
- Hybrid approach (marriage of both dynamic slicing and Statisical debugging)

The open source c projects making up the subjects in this VM include:
-CoREBench: real open source errors from UNIX utilities - grep, coreutils and findutils. 
-Tcas: injected errors from the Siemens and SIR Benchmark

__________________________________
CONFIGURATION
This VM has gcc(gcov), frama-c slicing tool, python and pip installed. 
Python-networkx and matplotbib python libraries are also installed. 

__________________________________
USING THE HYBRID VM
the folder "~/work/hybrid" contains the fault localization script:
* faultLocalizer.py .. to execute all relevant scripts for all/each 
	technique(s) and all/each subject(s)
	-- call with "python faultLocalizer.py <options>"

Default execution format with command line option is: 
	"python faultLocalizer.py -t <technique> -s <subject>"

More detailed execution format with command line options is:
	"python faultLocalizer.py -t <technique> -s <subject> -m <mode> -v <Nvalue> -r <ranking>"

Specific command line options (<options>) include:
* <technique> (-t): all, ochiai, slicing or hybrid
* <subject> (-s): all, coreutils, grep, findutils or tcas
* <mode> (-m): abs (absolute) or per (percent)
* <NValue> (-v): a multiple of 5 in the range 5 to 95 (max value range e.g. -v 5)
* <ranking> (-r): ord (ordinal) or mod (modified competition)

_______________________________
SAMPLE USE CASES

* to execute faultLocalizer for hybrid fault localizer for TCAS subject 
	in default mode (i.e. -v 5 -m abs), execute:
	"python faultLocalizer.py -t hybrid -s tcas"

* to execute faultLocalizer for the statistical debugging technique
	with OCHIAI for the GREP subject, using ordinal ranking, execute:
	"python faultLocalizer.py -t ochiai -s grep -r ord"

* to execute faultLocalizer for the statistical debugging technique
	with TARANTULA for the FINDUTILS subject, using modified competition
	ranking, execute:
	"python faultLocalizer.py -t tarantula -s findutils -r mod"

* to execute faultLocalizer for the statistical debugging technique
	with JACCARD for the TCAS subject, using ordinal ranking, execute:
	"python faultLocalizer.py -t jaccard -s tcas -r ord"

* to execute faultLocalizer for dynamic slicing for the FINDUTILS 
	subject, execute:
	"python faultLocalizer.py -t slicing -s findutils"

* to execute faultLocalizer for the hybrid technique for the TCAS
	subject for absolute N value 5, execute:
	"python faultLocalizer.py -t hybrid -s tcas -m abs -v 5"	

* to execute faultLocalizer.py for all techniques and subjects, 
	with command line options for absolute N value 5 for 
	the hybrid approach and modified competition ranking for 
	Statistical Debugging (using Ochiai) execute: 
	"python faultLocalizer.py -t all -s all -m abs -v 5 -r mod"

_______________________________
RESULTS
* All results are stored in a csv file titled: 
	<technique>_result.csv (e.g. hybrid_result.csv)
	in the folder: ~/work/hybrid/<subject>/scripts/

_______________________________
IMPLEMENTATION DETAILS

Lots of files required for the computation of scores of techniques and subjects are
pre-computed in this VM, these files include - gdb logs, slices, call graphs, statistical
debugging ranking and coverage information.

Note:
* suspicious methods are automatically determined from gdb logs (or gcov) for HYBRID_VM_2.0.
* TCAS shortest call sequences for inter-procedutal slicing are also automatically computed.