Installation Guide:
Install boost c++ library 1.48 or higher
have code in the trunk directory.

Compile Guide:
(Run compile.sh)
g++ -I [boost directory] *.cpp -O2 -std=c++0x -o [executable.exe]

To Run, arguments:
-seed [#] //This is the seed for random number generation

-g [graphFile.grf] //this is the graph file. See datafiles for examples

-f [idegree number | preindexedFile.ppg ] //I suggest using a number such as 3

-gen [f] [fragment.grf] | -gen [o|c|s|n|p|e|k] [#nodesMax] [#nodesMin]
//Fragment generation method, either load file, or generate from database.
//f= load file
//p= "p-random walk", used for testing
//other arguments are other fragment generation methods

-search [9 | 0-8]
//denotes the number of neighbors to use in N-tree algorithms
//'9' denotes "all".
//other numbers denote themselves.


-sim [F|E]
//I suggest F
//F denotes "use in and out neighbors, using N-tree bipartite matching"
//E denotes "use out neighbors only, N-tree bipartite matching"
//this effects i-degree algorithm only.

-n [# of trials]
//The number of fragments to generate and run
//This should be more than 1 only if we are not loading a fragment from a file

-w 1.0
//deprecated. Feel free to ignore this value.



NOTES:
In the main file decomment out mapping and scoring functions to reenable those functions.

Compile using compile.sh

Run newtest3.sh to run on randomly generated subgraphs of the database.
Run newtest4.sh to run query files. (make sure file paths are correct)