def _make(self, ep):
project = "Minna Bluff"
db = self.db
with db.session_ctx():
prj = db.get_project(project)
Ar40, Ar39, Ar38, Ar37, Ar36 = [], [], [], [], []
for dev in (("Eurotherm", "Furnace"), ("CO2")):
for si in prj.samples:
for li in si.labnumbers:
self.debug("blanks for {},{}".format(si.name, li.identifier))
for ai in li.analyses:
if ai.extraction.extraction_device.name in dev:
bs = self._extract_blanks(ai)
if bs is not None:
r = make_runid(li.identifier, ai.aliquot, ai.step)
# self.debug('blanks for {} {}'.format(r,bs))
Ar40.append(bs[0])
Ar39.append(bs[1])
Ar38.append(bs[2])
Ar37.append(bs[3])
Ar36.append(bs[4])
reg = WeightedMeanRegressor()
print "blanks for {}".format(dev)
for iso in (Ar40, Ar39, Ar38, Ar37, Ar36):
ys, es = zip(*iso)
reg.trait_set(ys=ys, yserr=es)
print reg.predict()
def _mean_regress(self, scatter, r, fit):
from pychron.core.regression.mean_regressor import MeanRegressor, WeightedMeanRegressor
if hasattr(scatter, 'yerror') and fit=='weighted mean':
if r is None or not isinstance(r, WeightedMeanRegressor):
r = WeightedMeanRegressor()
else:
if r is None or not isinstance(r, MeanRegressor):
r = MeanRegressor()
self._set_regressor(scatter, r)
# r.trait_setq(fit=fit)
r.calculate()
self._set_excluded(scatter, r)
return r
def _mean_regress(self, scatter, r, fit):
from pychron.core.regression.mean_regressor import MeanRegressor, WeightedMeanRegressor
if hasattr(scatter, 'yerror') and fit == 'weighted mean':
if r is None or not isinstance(r, WeightedMeanRegressor):
r = WeightedMeanRegressor()
else:
if r is None or not isinstance(r, MeanRegressor):
r = MeanRegressor()
self._set_regressor(scatter, r)
# r.trait_setq(fit=fit)
r.calculate()
self._set_excluded(scatter, r)
return r
def setUp(self):
n = 1000
ys = np.ones(n) * 5
# es = np.random.rand(n)
es = np.ones(n)
ys = np.hstack((ys, [5.1]))
es = np.hstack((es, [1000]))
# print es
self.reg = WeightedMeanRegressor(ys=ys, errors=es)
def _make(self, ep):
project = 'Minna Bluff'
db = self.db
with db.session_ctx():
prj = db.get_project(project)
Ar40, Ar39, Ar38, Ar37, Ar36 = [], [], [], [], []
for dev in (('Eurotherm', 'Furnace'), ('CO2')):
for si in prj.samples:
for li in si.labnumbers:
self.debug('blanks for {},{}'.format(
si.name, li.identifier))
for ai in li.analyses:
if ai.extraction.extraction_device.name in dev:
bs = self._extract_blanks(ai)
if bs is not None:
r = make_runid(li.identifier, ai.aliquot,
ai.step)
# self.debug('blanks for {} {}'.format(r,bs))
Ar40.append(bs[0])
Ar39.append(bs[1])
Ar38.append(bs[2])
Ar37.append(bs[3])
Ar36.append(bs[4])
reg = WeightedMeanRegressor()
print 'blanks for {}'.format(dev)
for iso in (Ar40, Ar39, Ar38, Ar37, Ar36):
ys, es = zip(*iso)
reg.trait_set(ys=ys, yserr=es)
print reg.predict()
def _mean_regress(self, scatter, r, fit):
if hasattr(scatter, 'yerror'):
if r is None or not isinstance(r, WeightedMeanRegressor):
r = WeightedMeanRegressor()
else:
if r is None or not isinstance(r, MeanRegressor):
r = MeanRegressor()
self._set_regressor(scatter, r)
r.trait_set(fit=fit, trait_change_notify=False)
r.calculate()
self._set_excluded(scatter, r)
return r
def _mean_regress(self, scatter, r, fit):
if hasattr(scatter, 'yerror'):
if r is None or not isinstance(r, WeightedMeanRegressor):
r = WeightedMeanRegressor()
else:
if r is None or not isinstance(r, MeanRegressor):
r = MeanRegressor()
self._set_regressor(scatter, r)
r.trait_set(fit=fit, trait_change_notify=False)
r.calculate()
self._set_excluded(scatter, r)
return r
def _get_regressor(self, fit, error_type, xs, ys, es):
fit = fit.lower()
if fit in ('linear', 'parabolic', 'cubic'):
reg = WeightedPolynomialRegressor(fit=fit)
elif fit in [
'preceding', 'bracketing interpolate', 'bracketing average'
]:
reg = InterpolationRegressor(kind=fit)
elif fit == 'weighted mean':
reg = WeightedMeanRegressor()
else:
print 'fit {} not valid'.format(fit)
raise NotImplementedError
mi = min(xs)
xs = [xi - mi for xi in xs]
reg.trait_set(error_type=error_type, xs=xs, ys=ys, yserr=es)
reg.calculate()
return reg
