Feature/custom file names

[dpastoor]

This cleans up the file handling around summary file needs significantly, including some nomenclature fixes, adjustment of flag positions and a fix when using SAEM/IMP to not fail with missing gradient file (it shouldn't be present then).

[dpastoor]

previously didn't have cpu file:

⋊> ~/m/t/r/m/pk on review/devin ⨯ bbi nonmem summary 200/200 --json --no-cov-file | jq .run_details.output_files_used                                                                   17:40:35
[
  "200.lst",
  "200.ext",
  "200.cor"
]

now picks up cpu file:

⋊> ~/m/t/r/m/pk on review/devin ⨯ bbi nonmem summary 200/200 --json --no-cov-file | jq .run_details.output_files_used                                                                   17:40:46
[
  "200.lst",
  "200.cpu",
  "200.ext",
  "200.cor"
]

no need to suppress grd file for SAEM/IMP + now includes cpu_time

⋊> ~/m/t/r/m/pk on review/devin ⨯ bbi nonmem summary 200/200 --json | jq .run_details                                                                                                   17:52:27
{
  "version": "7.4.4",
  "run_start": "-999999999",
  "run_end": "Sat Jun 13 13:55:26 EDT 2020",
  "estimation_time": 47.36,
  "covariance_time": 1.33,
  "cpu_time": 2906.315,
  "function_evaluations": -999999999,
  "significant_digits": -999999999,
  "problem_text": "LEM 200 104 + COV-effects(CRCL, AGE) on CL: (ALB) on V2",
  "mod_file": "-999999999",
  "estimation_method": [
    "Stochastic Approximation Expectation-Maximization",
    "Objective Function Evaluation by Importance Sampling"
  ],
  "data_set": "../../../data/derived/analysis1.csv",
  "number_of_patients": 160,
  "number_of_obs": 3142,
  "number_of_data_records": 4292,
  "output_files_used": [
    "200.lst",
    "200.cpu",
    "200.ext",
    "200.cov",
    "200.cor"
  ]
}

which matches up to the total CPU time:

⋊> ~/m/t/r/m/pk on review/devin ⨯ cat 200/200.lst | rg 2906                                                                                                                             17:52:32
 #CPUT: Total CPU Time in Seconds,     2906.315

[dpastoor]

Running a model with no grd and shk files we can see:

⋊> ~/D/psp412422-sup-0008-supinfo8 bbi nonmem summary urine --no-grd-file --json                                                                                                        20:27:15
FATA[0000] No file present at /Users/devinp/Downloads/psp412422-sup-0008-supinfo8/urine.shk
you can suppress bbi searching for the file using --no-shk-file

[seth127]

Looks good to me. I don't see where you tested giving it a custom ext name, but I see it in the code.

Not sure what the testing/merging regime is for bbi/Go but I guess merge and build this if you feel good about it and I'll test it further through rbabylon.