FLEXIS

FLEXIS: FLEXible Frequent Subgraph Mining using Maximal Independent Sets

Prerequisties

• Install CMake
• Install Ninja
• Install vcpkg and set VCPKG_ROOT environment

Generating

Debug build

cmake --preset debug

Release build

cmake --preset release

Compiling the executable

Debug build

cmake --build --preset debug

Release build

cmake --build --preset release

Download and process Datasets

bash ./scripts/processDatasets.sh --grami

Running the flexis system

Flexis command skeleton

./build/release/bin/flexis --input <vf3.bin file> --support <int> --slider [0.0...1.0]

Flexis executable --help

flexis


./build/release/bin/flexis [OPTIONS]


OPTIONS:
  -h,     --help              Print this help message and exit
          --input TEXT:FILE REQUIRED
                              Input bin file
          --output TEXT       Output json file
  -s,     --support INT [10]  support
  -k,     --max-size INT:INT in [2 - 10] [6]
                              max size of candidate pattern
  -a,     --slider FLOAT:FLOAT in [0 - 1] [0.98]
                              slider
          --verbose [0]       verbose mode
          --print-output [0]  print output
          --output-json [0]   output json
          --print-patterns [0]
                              print patterns

slider argument

Tells the slider used for Maximal Independent Set metric

max-size argument

Used to tell the FLEXIS the max-size vertex pattern to explore>

Command example

Example command 1

./build/release/bin/flexis --input ./data/processed/p2p-Gnutella08.vf3.bin --support 55 --slider 1.0

Example command 2

./build/release/bin/flexis --input ./data/processed/mico.vf3.bin --support 8000 --slider 0.9

Example with get the time and patterns count for frequent and merging

./build/release/bin/flexis --input ./data/processed/p2p-Gnutella08.vf3.bin --support 55 --slider 1.0 --print-output

Initial generation took 1 ms, generated 125 patterns
Step 2: starting with 125 patterns
Matching took 3 ms, reduced 125 patterns to 125 frequent patterns
Generation took 1230 ms, extended 125 patterns to 27225 new patterns
Step 3: starting with 27225 patterns
Matching took 2298 ms, reduced 27225 patterns to 2 frequent patterns
Generation took 0 ms, extended 2 patterns to 6 new patterns
Step 4: starting with 6 patterns
Matching took 0 ms, reduced 6 patterns to 0 frequent patterns
Generation took 0 ms, extended 0 patterns to 0 new patterns
Finished mining 4 steps out of 6 steps, 0 patterns left

Example with verbose flag

Prints the threashold and filename ./build/release/bin/flexis --input ./data/processed/p2p-Gnutella08.vf3.bin --support 55 --slider 1.0 --verbose

Filename ./data/processed/p2p-Gnutella08.vf3.bin
threshold is set to 55
Initial generation took 1 ms, generated 125 patterns
Step 2: starting with 125 patterns
Matching took 5 ms, reduced 125 patterns to 125 frequent patterns
Generation took 1272 ms, extended 125 patterns to 27225 new patterns
Step 3: starting with 27225 patterns
Matching took 2427 ms, reduced 27225 patterns to 2 frequent patterns
Generation took 0 ms, extended 2 patterns to 6 new patterns
Step 4: starting with 6 patterns
Matching took 0 ms, reduced 6 patterns to 0 frequent patterns
Generation took 0 ms, extended 0 patterns to 0 new patterns
Finished mining 4 steps out of 6 steps, 0 patterns left

Example with printing patterns

./build/release/bin/flexis --input ./data/processed/p2p-Gnutella08.vf3.bin --support 55 --slider 1.0 --print-output --print-patterns

It prints the patterns with the following format:

p <num_vertices> <num_edges>
v <vid> <vlabel>
...
e <src> <dst> <elabel>
...

The candidate and frequent patterns are printed.

• Candidate patterns are printed as: Merged patterns: <num_of_patterns> followed by each candidate pattern

• Frequent patterns are printed as: Frequent patterns: <num_of_patterns> followed by each frequent pattern

Initial generation took 1 ms, generated 125 patterns
Merged patterns: 125
p 2 1
v 0 2
v 1 1
e 0 1 1
p 2 1
v 0 2
v 1 0
e 0 1 0
p 2 1
v 0 4
v 1 2
e 0 1 2
p 2 1
v 0 1
v 1 3
e 0 1 0
p 2 1
v 0 0
v 1 3
e 0 1 1
...

p 2 1
v 0 3
v 1 1
e 0 1 1
p 2 1
v 0 1
v 1 2
e 0 1 2
Step 2: starting with 125 patterns
Matching took 3 ms, reduced 125 patterns to 125 frequent patterns
Frequent patterns: 125
p 2 1
v 0 2
v 1 1
e 0 1 1
...

e 0 1 0
p 2 1
v 0 3
v 1 1
e 0 1 1
p 2 1
v 0 1
v 1 2
e 0 1 2
Generation took 1275 ms, extended 125 patterns to 27225 new patterns
Merged patterns: 27225
p 3 3
v 0 0
v 1 3
v 2 3
e 2 1 2
e 0 1 4
e 0 2 3
...

p 3 3
v 0 4
v 1 3
v 2 0
e 2 0 3
e 1 2 1
e 0 1 4
Step 3: starting with 27225 patterns
Matching took 2389 ms, reduced 27225 patterns to 2 frequent patterns
Frequent patterns: 2
p 3 2
v 0 1
v 1 3
v 2 0
e 0 1 2
e 0 2 4
p 3 2
v 0 0
v 1 3
v 2 2
e 0 1 4
e 0 2 0
Generation took 0 ms, extended 2 patterns to 6 new patterns
Merged patterns: 6
p 4 3
v 0 1
v 1 3
v 2 0
v 3 0
e 0 1 2
e 0 2 4
e 0 3 4