def fit(problem, method=FIT_DEFAULT_ID, verbose=False, **options):
"""
Simplified fit interface.
Given a fit problem, the name of a fitter and the fitter options,
it will run the fit and return the best value and standard error of
the parameters. If *verbose* is true, then the console monitor will
be enabled, showing progress through the fit and showing the parameter
standard error at the end of the fit, otherwise it is completely
silent.
Returns an *OptimizeResult* object containing "x" and "dx". The
dream fitter also includes the "state" object, allowing for more
detailed uncertainty analysis. Optimizer information such as the
stopping condition and the number of function evaluations are not
yet included.
To run in parallel (with multiprocessing and dream)::
from bumps.mapper import MPMapper
mapper = MPMapper.start_mapper(problem, None, cpu=0) #cpu=0 for all CPUs
result = fit(problem, method="dream", mapper=mapper)
"""
from scipy.optimize import OptimizeResult
#verbose = True
if method not in FIT_AVAILABLE_IDS:
raise ValueError("unknown method %r not one of %s"
% (method, ", ".join(sorted(FIT_ACTIVE_IDS))))
for fitclass in FITTERS:
if fitclass.id == method:
break
monitors = None if verbose else [] # default is step monitor
driver = FitDriver(
fitclass=fitclass, problem=problem, monitors=monitors,
**options)
driver.clip() # make sure fit starts within domain
x0 = problem.getp()
x, fx = driver.fit()
problem.setp(x)
dx = driver.stderr()
if verbose:
print("final chisq", problem.chisq_str())
driver.show_err()
result = OptimizeResult(
x=x, dx=driver.stderr(),
fun=fx,
success=True, status=0, message="successful termination",
#nit=0, # number of iterations
#nfev=0, # number of function evaluations
#njev, nhev # jacobian and hessian evaluations
#maxcv=0, # max constraint violation
)
if hasattr(driver.fitter, 'state'):
result.state = driver.fitter.state
return result
def bumps_fit(self,
method='dream',
pop=15,
samples=1e5,
burn=100,
steps=0,
thin=1,
alpha=0,
outliers='none',
trim=False,
monitors=[],
problem=None,
**options):
# create a fitter similar to the bumps.fitters.fit function but with option for GUI monitoring
from scipy.optimize import OptimizeResult
from bumps.fitters import FitDriver, FIT_AVAILABLE_IDS, FITTERS, FIT_ACTIVE_IDS
options['pop'] = pop
options['samples'] = samples
options['burn'] = burn
options['steps'] = steps
options['thin'] = thin
options['alpha'] = alpha
options['outliers'] = outliers
options['trim'] = trim
if problem is None:
problem = self.bumps_problem()
# verbose = True
if method not in FIT_AVAILABLE_IDS:
raise ValueError("unknown method %r not one of %s" %
(method, ", ".join(sorted(FIT_ACTIVE_IDS))))
for fitclass in FITTERS:
if fitclass.id == method:
break
driver = FitDriver(fitclass=fitclass,
problem=problem,
monitors=monitors,
**options)
driver.clip() # make sure fit starts within domain
x0 = problem.getp()
x, fx = driver.fit()
problem.setp(x)
dx = driver.stderr()
result = OptimizeResult(x=x,
dx=driver.stderr(),
fun=fx,
cov=driver.cov(),
success=True,
status=0,
message="successful termination")
if hasattr(driver.fitter, 'state'):
result.state = driver.fitter.state
return result
