Useful in situations where you are tuning system and comparing benchmarks

sysbench_parser.py

#!/usr/bin/env python3
from json import dumps

test_load = """SQL statistics:
    queries performed:
        read:                            845684
        write:                           241589
        other:                           120831
        total:                           1208104
    transactions:                        60398  (1506.89 per sec.)
    queries:                             1208104 (30141.30 per sec.)
    ignored errors:                      8      (0.20 per sec.)
    reconnects:                          0      (0.00 per sec.)

General statistics:
    total time:                          40.0802s
    total number of events:              60398

Latency (ms):
         min:                                    6.38
         avg:                                   33.14
         max:                                  369.33
         95th percentile:                       51.94
         sum:                              2001450.92

Threads fairness:
    events (avg/stddev):           1207.9600/52.97
    execution time (avg/stddev):   40.0290/0.02
"""

class SysbenchParser:
    def __init__(self):
        self.hold = {}
        self.current = []

    def parse(self, row):
        if ':' not in row:
            return
        # We are in a tree branch state
        if row.rstrip().endswith(':'):
            if row[0] != ' ':
                self.current = [row.split(':')[0].lower()]
            else: 
                self.current.append(row.split(':')[0].lstrip()[:-1])
        # Leaf node
        else:
            t = [a.strip() for a in row.strip().split(':')]
            key = t[0].replace(' ', '_').lower()
            value = self.process_value(t[1])

            key = '_'.join([i.replace(' ', '_') for i in self.current]) + '_' + key
            self.hold[key] = value
        return

    def process_value(self, row):
        if row.endswith(')'):
            t = row.split(' ')
            total = float(row[0])
            e = None
            for i in t[1:]:
                if i.startswith('('):
                    e = i[1:]
            ps = float(e)
            return {'total': total, 'per_second': ps}
        elif row.endswith('s'):
            return {'second': float(row[:-1])}
        elif '/' in row:
            t = row.split('/')
            return {'values': [float(v) for v in t]}
        else: 
            return {'value': float(row)}



    def payload(self):
        return dumps(self.hold)


if __name__ == '__main__':
    d = SysbenchParser()
    import sys
    for i in sys.stdin.readlines():
        d.parse(i)
    print(d.payload())