pyyeti.cla.DR_Results.split

DR_Results.split()

Split results apart into a new DR_Results structure

Returns:

results (DR_Results instance) – This is a new, two-level instance of DR_Results (each entry value is a base-level instance of DR_Results). For example, if:

self['SC_atm'].cases == ['ff1', 'ff2', 'ff3']

then results would have:

results['ff1']['SC_atm'].cases = ['ff1']
results['ff2']['SC_atm'].cases = ['ff2']
results['ff3']['SC_atm'].cases = ['ff3']

Notes

This routine is sort of an inverse of merge(). Splitting allows easy comparisons of one sub-case to another. It also allows the deletion of some cases: split the results, delete the entries you don’t want, re-merge the results back together, and form a new set of extreme values (with form_extreme()).

To illustrate what this routine does, consider this ‘SC_ifa’ entry in an initial results structure. In this case, the CLA event is “Stage 1 / Stage 2 Separation” and there are 3 stage one / stage two separation cases.

PP(results[‘SC_ifa’]):

<class 'types.SimpleNamespace'>[n=16]
    .cases  : [n=3]: ['Sep 1', 'Sep 2', 'Sep 3']
    .domain : 'time'
    .drminfo: <class 'types.SimpleNamespace'>[n=20]
    .event  : 'Stage 1 / Stage 2 Separation'
    .ext    : float64 ndarray: (12, 2)
    .ext_x  : None
    .hist   : float64 ndarray: (3, 12, 10001)
    .maxcase: None
    .mx_x   : float64 ndarray: (12, 3)
    .mincase: None
    .mn_x   : float64 ndarray: (12, 3)
    .mission: 'Rocket / Spacecraft VLC'
    .mn     : float64 ndarray: (12, 3)
    .mx     : float64 ndarray: (12, 3)
    .srs    : <class 'types.SimpleNamespace'>[n=5]
        .ext  : <class 'dict'>[n=2]
            25: float64 ndarray: (12, 990)
            50: float64 ndarray: (12, 990)
        .frq  : float64 ndarray: (990,)
        .srs  : <class 'dict'>[n=2]
            25: float64 ndarray: (3, 12, 990)
            50: float64 ndarray: (3, 12, 990)
        .type : 'eqsine'
        .units: 'G, rad/sec^2'
    .time   : float32 ndarray: (10001,)

This routine would return a new structure with the three cases split up into three events. For example, sp_res = results.split(), would give the following. Note that there would typically be more categories than just ‘SC_ifa’, and they would all be split in the same way.

sp_res['Sep 1']['SC_ifa']
sp_res['Sep 2']['SC_ifa']
sp_res['Sep 3']['SC_ifa']

And sp_res['Sep 2']['SC_ifa'] would be (notice the 3’s become 1’s):

<class 'types.SimpleNamespace'>[n=16]
    .cases  : [n=1]: ['Sep 2']
    .domain : 'time'
    .drminfo: <class 'types.SimpleNamespace'>[n=20]
    .event  : 'Sep 2'
    .ext    : float64 ndarray: (12, 2)
    .ext_x  : None
    .hist   : float64 ndarray: (1, 12, 10001)
    .maxcase: None
    .mx_x   : float64 ndarray: (12, 1)
    .mincase: None
    .mn_x   : float64 ndarray: (12, 1)
    .mission: 'Rocket / Spacecraft VLC'
    .mn     : float64 ndarray: (12, 1)
    .mx     : float64 ndarray: (12, 1)
    .srs    : <class 'types.SimpleNamespace'>[n=5]
        .ext  : <class 'dict'>[n=2]
            25: float64 ndarray: (12, 990)
            50: float64 ndarray: (12, 990)
        .frq  : float64 ndarray: (990,)
        .srs  : <class 'dict'>[n=2]
            25: float64 ndarray: (1, 12, 990)
            50: float64 ndarray: (1, 12, 990)
        .type : 'eqsine'
        .units: 'G, rad/sec^2'
    .time   : float32 ndarray: (10001,)

Example usage 1:

# compare sub-cases 'MECO  1' and 'MECO 10':
sp = results.split()
sp['MECO  1'].rptpct(
    sp['MECO 10'], names=('MECO  1', 'MECO 10'),
    direc='m1_vs_m10')

Example usage 2:

# delete case 'MECO 15' from the results and form a new
# set of statistical extreme results:
from pyyeti.stats import ksingle

# split results and delete 'MECO 15':
sp = results.split()
del sp['MECO 15']

# merge remaining cases back together:
new_res = cla.DR_Results()
new_res.merge(sp.values())

# form non-statistical extrema (completes the merge):
new_res.form_extreme()

# change to statistical extrema:
ncases = len(new_res['extreme'].cases)
new_res['extreme'].calc_stat_ext(
    stats.ksingle(0.99, 0.90, ncases))

# compare new extrema to original:
new_res['extreme'].rptpct(
    results, names=('W/O MECO 15', 'Original'),
    direc='no_m15_vs_all')

Raises:

TypeError – When self[cat] is not a SimpleNamespace. This usually happens when self is a multiple level DR_Results instance. In that case, instead of res.split(), try something like res[event].split().