The data used to build a bloom filter and query is is generated through the python script in scripts/kmergen.py. Version 3.7.5 of python was used. See "environment.yml" for a list of packages used.
First, I generated random k-mers of length 100 like so:
cd scripts
python kmergen.py <num_of_kmers>
For the benchmarks shown in the writeup, I generated 5 sets of kmers (1 000, 10 000, 100 000, 1 000 000, 10 000 000). The output shows up in the data/ folder under keyfile_<num_of_kmers>.txt
The next step is to build Bloom filters of the previously generated keyfiles. Here is an example call of bf.py build
python bf.py build -k <path/to/keyfile> -f <target_fpr> -n <num_distinct_keys> -o <path/to/outfile>
For the benchmarks shown in the writeup, I generated Bloom filters over all 5 values of N above as well as over 5 different false positive rates (FPRs): {0.01, 0.05, 0.10, 0.20, 0.25}. This way I could truly test that I made a functional Bloom filter.
Next, a query can be called on the recently constructed Bloom filter from a file with a set of queries in it. Here is an example call to bf.py query:
python bf.py query -i <path/to/bloomfilter/outfile> -q <path/to/queryfile>
The results of the query will be outputted to the standard out in your terminal.
All the plots made for the writeup were made in the script: benchmarks/make_plots.R. Assuming you have all necessary libraries installed, you should be able to recreate all of these plots by sourcing the R script. I suggest running the script in R studio for better user experience; otherwise the plots would have to be saved and viewed in separate steps.