To use this CAFA benchmarking software perform the following 4 steps:

1. Generate Data (use the dataLoadingScripts directory):

usage: main.py [-h] [-o FILE] [-n FILE] [-g FILE] [-d DATAOUTPUT]

The purpose of this step is generate all swissprot data required to load into database for querying

optional arguments:

-h, --help show this help message and exit

-o FILE, --older FILE

                    Older swissprot input file.

-n FILE, --newer FILE

                    Newer swissprot input file.

-g FILE, --goterm FILE

                    Go term file, tab separated.

-d DATAOUTPUT, --dataoutput DATAOUTPUT

                    Provide path for redirecting your output.

2. Create database and load schema

   1.Create the database in mysql command line:

   create [dbname];

   exit

   2.mysql -u [username] -p [dbname] < cafa.sql

   (Note: cafa.sql located in schema directory)

3. Loading data generated in Step (1) into the database created in Step (2):

   Edit file: /dbCAFA/dataLoaderScripts/cafaDBLoading.txt to reflect dbNAME and file paths of all the data files generated in Step 1.

   Run the following command in shell:

   mysql -u [username] -p < /dbCAFA/dataLoaderScripts/cafaDBLoading.txt

   Note: Sometimes the path of the file is not found by mysql due to some problems, here is what I did to solve the issue:

     sudo apt-get install apparmor-utils
   
     sudo aa-complain /usr/sbin/mysqld
   
     sudo /etc/init.d/apparmor reload
   

4. Benchmark generation (use folder benchmarkGeneration):

   To generate benchmarks firstly modify 2nd line in cafaModel/modelCAFA.py to reflect your username, password and database name.

   Next run the script in benchmarkGeneration folder

   usage: generateBenchmarks.py [-h] [--chooseType CHOOSETYPE]

                            [--chooseOntology CHOOSEONTOLOGY]
   

   optional arguments:

   -h, --help show this help message and exit

   --chooseType CHOOSETYPE, -t CHOOSETYPE

                    Either noKnowledge or partialKnowledge
   

   --chooseOntology CHOOSEONTOLOGY, -o CHOOSEONTOLOGY

                    Only if partialKnowledge choose:
   
                    molecular_function OR
   
                    biological_process OR
   
                    cellular_component; 
                    
                    skip this option if noKnowledge
   

TODO:

1. The data loading scripts for Protein, GO, Protein_GO and Evidence need editing so that they write AI of each table in the files generated. For consistency purposes Protein file needs to include a header. Parameterize the dates in prepareProtein.py

2. cafaDBLoading.txt needs: auto increments reset before you begin loading. Database name, file path should be parameterized. The infile command loading protein data should ignore the header once above point is completed.

3. Editing the main.py in dataLoaderScripts directory so that it can run the files from anywhere.