#PropL project

PropL is a project of bugs prediction with machine learning, following works of Vincenzo Musco.
The PropL software is written with Python2.7 and Python3.4. The stochastic software is being developed.

• Results
• Scientific papers
• Files
• Dependencies
• How to use it?
• Contacts

###Results

Java Project	Mutant operator	F-score init.	F-score learn.
Codec	ABS	0.00	0.56
Codec	AOR	0.19	0.75
Codec	LCR	0.11	0.66
Codec	ROR	0.19	0.74
Codec	UOI	0.20	0.77
Collections	ABS	0.26	0.20
Collections	AOR	0.67	0.27
Collections	LCR	0.25	0.29
Collections	ROR	0.24	0.23
Collections	UOI	0.63	0.36
Gson	ABS	0.00	0.79
Gson	AOR	0.03	0.73
Gson	LCR	0.02	0.76
Gson	ROR	0.05	0.82
Gson	UOI	0.05	0.83
Io	ABS	0.06	0.57
Io	AOR	0.32	0.66
Io	LCR	0.29	0.74
Io	ROR	0.50	0.71
Io	UOI	0.42	0.7
Lang	ABS	0.00	0.39
Lang	AOR	0.60	0.60
Lang	LCR	0.40	0.53
Lang	ROR	0.67	0.56
Lang	UOI	0.67	0.79
Shindig	ABS	0.00	0.55
Shindig	AOR	0.50	0.62
Shindig	LCR	0.33	0.57
Shindig	ROR	0.40	0.66
Shindig	UOI	0.50	0.65
Sonar	ABS	0.40	0.00
Sonar	AOR	0.17	0.60
Sonar	LCR	0.67	0.54
Sonar	ROR	0.67	0.00
Sonar	UOI	0.17	0.67
Spojo	ABS	0.00	0.00
Spojo	AOR	0.00	0.00
Spojo	LCR	0.32	0.64
Spojo	ROR	0.29	0.80
Spojo	UOI	0.00	0.65

###Scientific papers

• Initial work of Vincenzo Musco
• The learning approach - soon

###Files

• compute-basename.py : Program to compute f-scores with the algorithm of Vincenzo Musco (F-score init.).
• propl.py : Program to compute f-scores with the learning approach (F-score learn.).
• libs/
  • basic_stat_lib.py : Some functions to make some stats (compute precision, recall, f-score, etc...)
  • graph_visualization.py : Program (Python2.7) to visualize impacted nodes in a .graphml file
  • learning_lib.py : Some learning algorithms
  • testing_lib.py : Some functions to test the learning algorithm choose on tests sample
  • tex_lib.py : Library to write results in a tex file
  • use_graph_lib.py : Object which represents a use graph (see the definition of a use graph in the recent paper of Vincenzo Musco)
  • utils_lib.py : Some functions to chunk a list, write into a CSV file, etc...
  • xml_lib.py : Personal XML library to load and give an appreciation on the XML document (valid or not)
  • xml_parsing_lib.py : Personal XML library to parse XML documents from Vincenzo Musco
  • exceptions/
    • FailToLoad.py
    • NoArgument.py
    • RunError.py

###Dependencies

• networkx (Python2.7/Python3.4)
• numpy (Python2.7/Python3.7)
• matplotlib (Python2.7 only)

###How to use it?

You can use the script written in bash : chmod +x run.sh && ./run.sh directory_test usegraph_x.graphml directory_to_store_results, or...

• In the root directory : python3.4 propl.py <your_test_directory> <number_of_tests> [--option]
• For help : python3.4 prop.py --help

###Contacts

• Developer : Antonin Carette (antonin[dot]carette[at]etudiant[dot]univ-lille1[dot]fr)
• First supervisor : Philippe Preux (philippe[dot]preux[at]inria[dot]fr)
• Second supervisor : Martin Monperrus (martin[dot]monperrus[at]univ-lille1[dot]fr)
• Third supervisor : Vincenzo Musco (vincenzo[dot]musco[at]inria[dot]fr)