DatabaseStats.py

from ElsevierAPI import open_api_session
import time
import argparse
import textwrap


ps_api = open_api_session()
NON_DIRECTIONAL = ['Binding','FunctionalAssociation','CellExpression','Metabolization','Paralog']

def IsDirectional(relType:str):
    return relType not in NON_DIRECTIONAL

def CountRelations (EntityIN, EntityOUT, RelType, isDirectional=True, TextRefs:list=None):
    if isDirectional == True:
        base_GOQLquery = 'SELECT Relation WHERE objectType = '+ RelType +' AND NeighborOf downstream (SELECT Entity WHERE objectType = '+EntityIN+') AND NeighborOf upstream (SELECT Entity WHERE objectType = '+EntityOUT+')'
    elif EntityIN != EntityOUT:
        base_GOQLquery = 'SELECT Relation WHERE objectType = '+ RelType +' AND NeighborOf (SELECT Entity WHERE objectType = '+EntityIN+') AND NeighborOf (SELECT Entity WHERE objectType = '+EntityOUT+')'
    else:
        base_GOQLquery = 'SELECT Relation WHERE objectType = '+ RelType +' AND NeighborOf downstream (SELECT Entity WHERE objectType = '+EntityIN+') AND NeighborOf downstream (SELECT Entity WHERE objectType = '+EntityOUT+')'

    textref_counts = {'Total':ps_api.get_result_size(base_GOQLquery)}
    if isinstance(TextRefs,list):
        for t in TextRefs:
            texref_clause = " AND TextRef LIKE '%{textref_type}%'".format(textref_type=t)
            textref_counts[texref_clause] = ps_api.get_result_size(base_GOQLquery + texref_clause)

    return textref_counts


def GetStats(Triples:list, TextRefs:list=None):
    non_directional_counted = set()
    start_time = time.time()
    with open('DBstats.txt', 'w', encoding='utf-8') as f:
        additional_columns = str()
        if isinstance(TextRefs,list):
            additional_columns = '\t'.join(TextRefs)
        f.write('Regulator\tTarget\tRelation\tRelation counts'+additional_columns+'\n')
        for triple in Triples:
            EntityIN = triple[0]
            rel_type = triple[1]
            EntityOUT = triple[2]

            is_directional = IsDirectional(rel_type)
            if is_directional == False:
                if EntityOUT+EntityIN in non_directional_counted: 
                    print('skipping: %s-%s-%s was already counted' % (EntityIN,rel_type,EntityOUT))
                    continue
                else: non_directional_counted.update([EntityIN+EntityOUT])

            rel_counts_dict = CountRelations(EntityIN,EntityOUT,rel_type,is_directional,TextRefs)
            if len(rel_counts_dict) == 0:
                print('%s\t%s\t%s triple has no relations' % (EntityIN,EntityOUT,rel_type))
            else:
                for stat, rel_count in rel_counts_dict.items():
                    print('%s\t%s\t%s:' % (EntityIN,EntityOUT,rel_type))
                    print('%s - %d' % stat,rel_count)
                    f.write('%s\t%s\t%s' % (EntityIN,EntityOUT,rel_type))
                    f.write('\t%d' % rel_count)
                f.write('\n')

    print("Time to retreive database stats: %s" % (ps_api.execution_time(start_time)))


if __name__ == "__main__":
    instructions = '''
    infile - 3 column tab-delimted file with triplets description: Regulator<>Relation<tab>Target.
    TextRefs - comma-separated list of TextRef substrings to obtain document sections stats.\n
    Possible TextRefs values: "abs", "body", "pmid", "doi"
    If infile is not supplied counts for all possible triplets for all entity types and all relation types will be outputed

    '''
    parser = argparse.ArgumentParser(formatter_class=argparse.RawDescriptionHelpFormatter,epilog=textwrap.dedent(instructions))
    parser.add_argument('-i', '--infile', default='')
    parser.add_argument('-t', '--TextRefs', default='')
    args = parser.parse_args()


    DBRelTypes = list()
    DBEntities = list()
    DBtriples_types = list()
    TextRefs  = None
    if args.TextRefs:
        TextRefs = str(args.TextRefs).split(',')

    if not args.infile:
        DBRelTypes = ps_api.get_relation_types()
        DBEntities = ps_api.get_entity_types()
        for rel_type in DBRelTypes:
            for EntityIN in DBEntities:
                for EntityOUT in DBEntities:
                    DBtriples_types.append([EntityIN,rel_type,EntityOUT])
    else:
        with open(args.infile, 'r', encoding='utf-8') as f:
            for line in f:
                triple = line.strip().split('\t')
                DBtriples_types.append(triple)

    GetStats(DBtriples_types,TextRefs)