How to use

Install dependencies

To use this benchmark you need to install the dependencies of python.

pip install -r requirements.txt

Build project

The first time when the benchmark are launched, you need to build the jar. To make this you need to build the project with the flag -b.

python3 benchmark.py -b

Run Benchmark

To run a benchmark you can use a various flags.

Flags

-b (--build): If present, build the project and exits
-d (--debug): If present, print log to stdout
-v (--verbose): If present, print with more details
-n (--num_rel_block): Number of the relationships loaded from the dataset file
-f (--file): Path of the dataset file of relationships
-q (--queries): Number of the queries to run
-dp (--depth): Depth of queries between 1 and 10, default is r=random
-lp (--log_path): The path where the logs are saved, default the path of execution of the benchmark
-a (--algorithms): Select algorithm to use with specific depth (j=join, b=traverse-bfs, d=traverse-dfs). Default value: bbbbjjjjddd
-ln (--limit_node_research): Number of the maximum id node to research, default 3000

Benchmark Examples

If you want to run a benchmark with depth 5 using only CSR-BFS, you can use the following command:

python3 benchmark.py -v -n 1000 -dp 5 -q 100 -a bbbbbbbbbb

With this command you run a benchmark loading only the first 1000 relationships from the dataset file and exec 100 queries with depth 5 resolving that with CSR-BFS algorithm.

If you want to run a benchmark with queries with random depth and with a specific algorithm in a determined depth, you can use:

python3 benchmark.py -v -n 1000 -dp r -q 100 -a bbbbjjjjdd

The -a flag specific in a determined position what kind of algorithm that you want to use. In this case, with depth <= 4 to resolve the query is used the CSR-BFS algorithm and for depth >= 9 is used the CSR-DFS. For depths between 5 and 8, the query is resolved with HashJoin.