Compression scheme using maximal cliques

1. Requirements
- The implementation requires to list all maximal cliques of a graph. We use quick_cliques, which link is avalable in the paper.
- The sdsl-light library for compact data structures.

2. Compiling
To compile run:
./compile.sh

This generates all executables, getcliquespar, timesseq, and timesrandom

3. Example

- The output/ca-CondMat/ contains the maximal cliques for the ca-CondMat graph that is also stored in text in the output/graphs directory

Generating the compressed graph. The following will create the sequences using the ranking function r_f.
This will generate files with ending in *.f.*

./frequency output/ca-CondMat/ca-CondMat.cliques F 

The following command will build the compressed representation of the graph, first argument is the 
input file prefix, the second is the output directory and file name and the third is the number of edges which it is optional (if given it computes de bpe).
./compress output/ca-CondMat/ca-CondMat.cliques.f output/caout/ca-CondMat.cliques.f

The following command will retrieve all maximal cliques from the compressed graph using 2 threads
./getcliquespar output/caout/ca-CondMat.cliques.f 2

The next command will access the compressed graph randomly according to a query file generated with randomqueries program in scripts directory
./timesseq output/caout/ca-CondMat.cliques.f queries/qca-CondMat.bin

The next command will access the compressed graph sequentially
./timesseq output/caout/ca-CondMat.cliques.f 21363

Acknowledge
- We use the huffman library of Author : Michael Dipperstein as a base, adding the functions we needed for the implementation of the byte sequence BB in the paper. (http://michael.dipperstein.com/huffman/)