LAVA Custom Test JSON File Parse Pattern Format

I'm currently working on getting a custom test job (lmbench) implemented with LAVA. I'm currently stuck on trying to get LAVA to parse my results correctly. I'm currently not familiar with how I need to implement the regular expression in the .json file in order for lava to parse it correctly. The correct regular expression that works with python is as follows:

   '(?P<test_case_id>\w+)\s+:\s+(?P<measurement>[\d\.]+)\s+:\s+(?P<units>[^\s]+)'

Here is my attempt at creating a json file to make my test work. The test runs fine and says it passes but it just isn't capturing the data I want. What do I need to do in order for the pattern to work correctly? Also is there any documentation on the format that this pattern needs to be in (since I haven't been able to find any)?

{
   "format": "Lava-Test Test Definition 1.0",
   "test_id": "lmbench",
   "run": {
      "steps": ["/usr/bin/python /root/lava_metric_wrapper.py /root/test_metrics.csv"]
   },
   "parse": {
      "pattern": "^(?P<test_case_id>\\w+)\\s+:\\s+(?P<measurement>[\\d\\.]+)\\s+:\\s+(?P<units>[^\\s]+)",
      "appendall": {
         "result": "pass"
      }
    }
}

Example output that LAVA captures

    2+0 records in
    2+0 records out
    read_throughput_64kB_blocks : 170.97 : MB/sec
    read_data_to_proc_throughput : 270.10 : MB/sec
    write_data_to_proc_throughput : 1303.54 : MB/sec
    write_latency_metric : 1.7994 : usec
    read_latency_metric : 1.1746 : usec
    open-close_latency_metric : 25.6473 : usec

Thanks,
-Ryan