core.py

import math
import os
import sys
import logging
import time
import copy
import shutil
import glob
import os.path as op
from string import Template

import numpy as num

from pyrocko.guts import load, Object, String, Float, Int, Bool, List, \
    StringChoice, Dict, Timestamp
from pyrocko import orthodrome as od, gf, trace, guts, util, weeding
from pyrocko import parimap, model, marker as pmarker
from pyrocko.guts_array import Array

from grond import dataset

logger = logging.getLogger('grond.core')

guts_prefix = 'grond'


def float_or_none(x):
    if x is None:
        return x
    else:
        return float(x)


class Trace(Object):
    pass


class TraceSpectrum(Object):
    network = String.T()
    station = String.T()
    location = String.T()
    channel = String.T()
    deltaf = Float.T(default=1.0)
    fmin = Float.T(default=0.0)
    ydata = Array.T(shape=(None,), dtype=num.complex, serialize_as='list')

    def get_ydata(self):
        return self.ydata

    def get_xdata(self):
        return self.fmin + num.arange(self.ydata.size) * self.deltaf


def mahalanobis_distance(xs, mx, cov):
    imask = num.diag(cov) != 0.
    icov = num.linalg.inv(cov[imask, :][:, imask])
    temp = xs[:, imask] - mx[imask]
    return num.sqrt(num.sum(temp * num.dot(icov, temp.T).T, axis=1))


class Parameter(Object):
    name = String.T()
    unit = String.T(optional=True)
    scale_factor = Float.T(default=1., optional=True)
    scale_unit = String.T(optional=True)
    label = String.T(optional=True)

    def __init__(self, *args, **kwargs):
        if len(args) >= 1:
            kwargs['name'] = args[0]
        if len(args) >= 2:
            kwargs['unit'] = args[1]

        Object.__init__(self, **kwargs)

    def get_label(self, with_unit=True):
        l = [self.label or self.name]
        if with_unit:
            unit = self.get_unit_label()
            if unit:
                l.append('[%s]' % unit)

        return ' '.join(l)

    def get_value_label(self, value, format='%(value)g%(unit)s'):
        value = self.scaled(value)
        unit = self.get_unit_suffix()
        return format % dict(value=value, unit=unit)

    def get_unit_label(self):
        if self.scale_unit is not None:
            return self.scale_unit
        elif self.unit:
            return self.unit
        else:
            return None

    def get_unit_suffix(self):
        unit = self.get_unit_label()
        if not unit:
            return ''
        else:
            return ' %s' % unit

    def scaled(self, x):
        if isinstance(x, tuple):
            return tuple(v/self.scale_factor for v in x)
        if isinstance(x, list):
            return list(v/self.scale_factor for v in x)
        else:
            return x/self.scale_factor


class ADict(dict):
    def __getattr__(self, k):
        return self[k]

    def __setattr__(self, k, v):
        self[k] = v


class Problem(Object):
    name = String.T()
    parameters = List.T(Parameter.T())
    dependants = List.T(Parameter.T())
    apply_balancing_weights = Bool.T(default=True)
    base_source = gf.Source.T(optional=True)

    def __init__(self, **kwargs):
        Object.__init__(self, **kwargs)
        self._bootstrap_weights = None
        self._target_weights = None
        self._engine = None
        self._group_mask = None

    def get_engine(self):
        return self._engine

    def copy(self):
        o = copy.copy(self)
        o._bootstrap_weights = None
        o._target_weights = None
        return o

    def parameter_dict(self, x):
        return ADict((p.name, v) for (p, v) in zip(self.parameters, x))

    def parameter_array(self, d):
        return num.array([d[p.name] for p in self.parameters], dtype=num.float)

    @property
    def parameter_names(self):
        return [p.name for p in self.combined]

    def dump_problem_info(self, dirname):
        fn = op.join(dirname, 'problem.yaml')
        util.ensuredirs(fn)
        guts.dump(self, filename=fn)

    def dump_problem_data(self, dirname, x, ms, ns):
        fn = op.join(dirname, 'x')
        with open(fn, 'ab') as f:
            x.astype('<f8').tofile(f)

        fn = op.join(dirname, 'misfits')
        with open(fn, 'ab') as f:
            ms.astype('<f8').tofile(f)
            ns.astype('<f8').tofile(f)

    def name_to_index(self, name):
        pnames = [p.name for p in self.combined]
        return pnames.index(name)

    @property
    def nparameters(self):
        return len(self.parameters)

    @property
    def ntargets(self):
        return len(self.targets)

    @property
    def ndependants(self):
        return len(self.dependants)

    @property
    def ncombined(self):
        return len(self.parameters) + len(self.dependants)

    @property
    def combined(self):
        return self.parameters + self.dependants

    def make_bootstrap_weights(self, nbootstrap):
        ntargets = self.ntargets
        ws = num.zeros((nbootstrap, ntargets))
        rstate = num.random.RandomState(23)
        for ibootstrap in xrange(nbootstrap):
            ii = rstate.randint(0, ntargets, size=self.ntargets)
            ws[ibootstrap, :] = num.histogram(
                ii, ntargets, (-0.5, ntargets - 0.5))[0]

        return ws

    def get_bootstrap_weights(self, ibootstrap=None):
        if self._bootstrap_weights is None:
            self._bootstrap_weights = self.make_bootstrap_weights(
                self.nbootstrap)

        if ibootstrap is None:
            return self._bootstrap_weights
        else:
            return self._bootstrap_weights[ibootstrap, :]

    def set_engine(self, engine):
        self._engine = engine

    def make_group_mask(self):
        super_group_names = set()
        groups = num.zeros(len(self.targets), dtype=num.int)
        ngroups = 0
        for itarget, target in enumerate(self.targets):
            if target.super_group not in super_group_names:
                super_group_names.add(target.super_group)
                ngroups += 1

            groups[itarget] = ngroups - 1

        return groups, ngroups

    def get_group_mask(self):
        if self._group_mask is None:
            self._group_mask = self.make_group_mask()

        return self._group_mask

    def xref(self):
        return self.pack(self.base_source)


class ProblemConfig(Object):
    name_template = String.T()
    apply_balancing_weights = Bool.T(default=True)


class Forbidden(Exception):
    pass


class DirectoryAlreadyExists(Exception):
    pass


class GrondError(Exception):
    pass


class DomainChoice(StringChoice):
    choices = [
        'time_domain',
        'frequency_domain',
        'envelope',
        'absolute',
        'cc_max_norm']


class InnerMisfitConfig(Object):
    fmin = Float.T()
    fmax = Float.T()
    ffactor = Float.T(default=1.5)
    tmin = gf.Timing.T(
        help='Start of main time window used for waveform fitting.')
    tmax = gf.Timing.T(
        help='End of main time window used for waveform fitting.')
    tfade = Float.T(
        optional=True,
        help='Decay time of taper prepended and appended to main time window '
             'used for waveform fitting [s].')
    pick_synthetic_traveltime = gf.Timing.T(
        optional=True,
        help='Synthetic phase arrival definition for alignment of observed '
             'and synthetic traces.')
    pick_phasename = String.T(
        optional=True,
        help='Name of picked phase for alignment of observed and synthetic '
             'traces.')
    domain = DomainChoice.T(
        default='time_domain',
        help='Type of data characteristic to be fitted.\n\nAvailable choices '
             'are: %s' % ', '.join("``'%s'``" % s
                                   for s in DomainChoice.choices))
    tautoshift_max = Float.T(
        default=0.0,
        help='If non-zero, allow synthetic and observed traces to be shifted '
             'against each other by up to +/- the given value [s].')
    autoshift_penalty_max = Float.T(
        default=0.0,
        help='If non-zero, a penalty misfit is added for non-zero shift '
             'values.\n\nThe penalty value is computed as '
             '``autoshift_penalty_max * normalization_factor * tautoshift**2 '
             '/ tautoshift_max**2``')

    def get_full_frequency_range(self):
        return self.fmin / self.ffactor, self.fmax * self.ffactor


class TargetAnalysisResult(Object):
    balancing_weight = Float.T()


class NoAnalysisResults(Exception):
    pass


class MisfitResult(gf.Result):
    misfit_value = Float.T()
    misfit_norm = Float.T()
    processed_obs = Trace.T(optional=True)
    processed_syn = Trace.T(optional=True)
    filtered_obs = Trace.T(optional=True)
    filtered_syn = Trace.T(optional=True)
    spectrum_obs = TraceSpectrum.T(optional=True)
    spectrum_syn = TraceSpectrum.T(optional=True)
    taper = trace.Taper.T(optional=True)
    tobs_shift = Float.T(optional=True)
    tsyn_pick = Timestamp.T(optional=True)
    cc_shift = Float.T(optional=True)
    cc = Trace.T(optional=True)


class MisfitTarget(gf.Target):
    misfit_config = InnerMisfitConfig.T()
    flip_norm = Bool.T(default=False)
    manual_weight = Float.T(default=1.0)
    analysis_result = TargetAnalysisResult.T(optional=True)
    super_group = gf.StringID.T()
    group = gf.StringID.T()

    def __init__(self, **kwargs):
        gf.Target.__init__(self, **kwargs)
        self._ds = None
        self._result_mode = 'sparse'

    def string_id(self):
        return '.'.join(x for x in (
            self.super_group, self.group) + self.codes if x)

    def get_plain_target(self):
        d = dict(
            (k, getattr(self, k)) for k in gf.Target.T.propnames)
        return gf.Target(**d)

    def get_dataset(self):
        return self._ds

    def set_dataset(self, ds):
        self._ds = ds

    def set_result_mode(self, result_mode):
        self._result_mode = result_mode

    def get_combined_weight(self, apply_balancing_weights):
        w = self.manual_weight
        if apply_balancing_weights:
            w *= self.get_balancing_weight()

        return w

    def get_balancing_weight(self):
        if not self.analysis_result:
            raise NoAnalysisResults('no balancing weights available')

        return self.analysis_result.balancing_weight

    def get_taper_params(self, engine, source):
        store = engine.get_store(self.store_id)
        config = self.misfit_config
        tmin_fit = source.time + store.t(config.tmin, source, self)
        tmax_fit = source.time + store.t(config.tmax, source, self)
        tfade = 1.0/config.fmin
        if config.tfade is None:
            tfade_taper = tfade
        else:
            tfade_taper = config.tfade

        return tmin_fit, tmax_fit, tfade, tfade_taper

    def get_backazimuth_for_waveform(self):
        nslc = self.codes
        if nslc[-1] == 'R':
            backazimuth = self.azimuth + 180.
        elif nslc[-1] == 'T':
            backazimuth = self.azimuth + 90.
        else:
            backazimuth = None

        return backazimuth

    def get_freqlimits(self):
        config = self.misfit_config

        return (
            config.fmin/config.ffactor,
            config.fmin, config.fmax,
            config.fmax*config.ffactor)

    def get_pick_shift(self, engine, source):
        config = self.misfit_config
        tobs = None
        tsyn = None
        ds = self.get_dataset()

        if config.pick_synthetic_traveltime and config.pick_phasename:
            store = engine.get_store(self.store_id)
            tsyn = source.time + store.t(
                config.pick_synthetic_traveltime, source, self)

            marker = ds.get_pick(
                source.name,
                self.codes[:3],
                config.pick_phasename)

            if marker:
                tobs = marker.tmin

        return tobs, tsyn

    def get_cutout_timespan(self, tmin, tmax, tfade):
        tinc_obs = 1.0 / self.misfit_config.fmin

        tmin_obs = (math.floor(
            (tmin - tfade) / tinc_obs) - 1.0) * tinc_obs
        tmax_obs = (math.ceil(
            (tmax + tfade) / tinc_obs) + 1.0) * tinc_obs

        return tmin_obs, tmax_obs

    def post_process(self, engine, source, tr_syn):

        tr_syn = tr_syn.pyrocko_trace()
        nslc = self.codes

        config = self.misfit_config

        tmin_fit, tmax_fit, tfade, tfade_taper = \
            self.get_taper_params(engine, source)

        ds = self.get_dataset()

        tobs, tsyn = self.get_pick_shift(engine, source)
        if None not in (tobs, tsyn):
            tobs_shift = tobs - tsyn
        else:
            tobs_shift = 0.0

        tr_syn.extend(
            tmin_fit - tfade * 2.0,
            tmax_fit + tfade * 2.0,
            fillmethod='repeat')

        freqlimits = self.get_freqlimits()

        tr_syn = tr_syn.transfer(
            freqlimits=freqlimits,
            tfade=tfade)

        tr_syn.chop(tmin_fit - 2*tfade, tmax_fit + 2*tfade)

        tmin_obs, tmax_obs = self.get_cutout_timespan(
            tmin_fit+tobs_shift, tmax_fit+tobs_shift, tfade)

        try:
            tr_obs = ds.get_waveform(
                nslc,
                tmin=tmin_obs,
                tmax=tmax_obs,
                tfade=tfade,
                freqlimits=freqlimits,
                deltat=tr_syn.deltat,
                cache=True,
                backazimuth=self.get_backazimuth_for_waveform())

            if tobs_shift != 0.0:
                tr_obs = tr_obs.copy()
                tr_obs.shift(-tobs_shift)

            mr = misfit(
                tr_obs, tr_syn,
                taper=trace.CosTaper(
                    tmin_fit - tfade_taper,
                    tmin_fit,
                    tmax_fit,
                    tmax_fit + tfade_taper),
                domain=config.domain,
                exponent=2,
                flip=self.flip_norm,
                result_mode=self._result_mode,
                tautoshift_max=config.tautoshift_max,
                autoshift_penalty_max=config.autoshift_penalty_max)

            mr.tobs_shift = float(tobs_shift)
            mr.tsyn_pick = float_or_none(tsyn)

            return mr

        except dataset.NotFound, e:
            logger.debug(str(e))
            raise gf.SeismosizerError('no waveform data, %s' % str(e))


def misfit(
        tr_obs, tr_syn, taper, domain, exponent, tautoshift_max,
        autoshift_penalty_max, flip, result_mode='sparse'):

    '''
    Calculate misfit between observed and synthetic trace.

    :param tr_obs: observed trace as :py:class:`pyrocko.trace.Trace`
    :param tr_syn: synthetic trace as :py:class:`pyrocko.trace.Trace`
    :param taper: taper applied in timedomain as
        :py:class:`pyrocko.trace.Taper`
    :param domain: how to calculate difference, see :py:class:`DomainChoice`
    :param exponent: exponent of Lx type norms
    :param tautoshift_max: if non-zero, return lowest misfit when traces are
        allowed to shift against each other by up to +/- ``tautoshift_max``
    :param autoshift_penalty_max: if non-zero, a penalty misfit is added for
        for non-zero shift values. The penalty value is
        ``autoshift_penalty_max * normalization_factor * \
tautoshift**2 / tautoshift_max**2``
    :param flip: ``bool``, if set to ``True``, normalization factor is
        computed against *tr_syn* rather than *tr_obs*
    :param result_mode: ``'full'``, include traces and spectra or ``'sparse'``,
        include only misfit and normalization factor in result

    :returns: object of type :py:class:`MisfitResult`
    '''

    trace.assert_same_sampling_rate(tr_obs, tr_syn)
    tmin, tmax = taper.time_span()

    tr_proc_obs, trspec_proc_obs = _process(tr_obs, tmin, tmax, taper, domain)
    tr_proc_syn, trspec_proc_syn = _process(tr_syn, tmin, tmax, taper, domain)

    cc_shift = None
    ctr = None
    deltat = tr_proc_obs.deltat
    if domain in ('time_domain', 'envelope', 'absolute'):
        a, b = tr_proc_syn.ydata, tr_proc_obs.ydata
        if flip:
            b, a = a, b

        nshift_max = max(0, min(a.size-1,
                                int(math.floor(tautoshift_max / deltat))))

        if nshift_max == 0:
            m, n = trace.Lx_norm(a, b, norm=exponent)
        else:
            mns = []
            for ishift in xrange(-nshift_max, nshift_max+1):
                if ishift < 0:
                    a_cut = a[-ishift:]
                    b_cut = b[:ishift]
                elif ishift == 0:
                    a_cut = a
                    b_cut = b
                elif ishift > 0:
                    a_cut = a[:-ishift]
                    b_cut = b[ishift:]

                mns.append(trace.Lx_norm(a_cut, b_cut, norm=exponent))

            ms, ns = num.array(mns).T

            iarg = num.argmin(ms)
            tshift = (iarg-nshift_max)*deltat

            m, n = ms[iarg], ns[iarg]
            m += autoshift_penalty_max * n * tshift**2 / tautoshift_max**2

    elif domain == 'cc_max_norm':

        ctr = trace.correlate(
            tr_proc_syn,
            tr_proc_obs,
            mode='same',
            normalization='normal')

        cc_shift, cc_max = ctr.max()
        m = 0.5 - 0.5 * cc_max
        n = 0.5

    elif domain == 'frequency_domain':
        a, b = trspec_proc_syn.ydata, trspec_proc_obs.ydata
        if flip:
            b, a = a, b

        m, n = trace.Lx_norm(num.abs(a), num.abs(b), norm=exponent)

    if result_mode == 'full':
        result = MisfitResult(
            misfit_value=m,
            misfit_norm=n,
            processed_obs=tr_proc_obs,
            processed_syn=tr_proc_syn,
            filtered_obs=tr_obs.copy(),
            filtered_syn=tr_syn,
            spectrum_obs=trspec_proc_obs,
            spectrum_syn=trspec_proc_syn,
            taper=taper,
            cc_shift=cc_shift,
            cc=ctr)

    elif result_mode == 'sparse':
        result = MisfitResult(
            misfit_value=m,
            misfit_norm=n)
    else:
        assert False

    return result


def _process(tr, tmin, tmax, taper, domain):
    tr_proc = _extend_extract(tr, tmin, tmax)
    tr_proc.taper(taper)

    df = None
    trspec_proc = None

    if domain == 'envelope':
        tr_proc = tr_proc.envelope(inplace=False)

    elif domain == 'absolute':
        tr_proc.set_ydata(num.abs(tr_proc.get_ydata()))

    elif domain == 'frequency_domain':
        ndata = tr_proc.ydata.size
        nfft = trace.nextpow2(ndata)
        padded = num.zeros(nfft, dtype=num.float)
        padded[:ndata] = tr_proc.ydata
        spectrum = num.fft.rfft(padded)
        df = 1.0 / (tr_proc.deltat * nfft)

        trspec_proc = TraceSpectrum(
            network=tr_proc.network,
            station=tr_proc.station,
            location=tr_proc.location,
            channel=tr_proc.channel,
            deltaf=df,
            fmin=0.0,
            ydata=spectrum)

    return tr_proc, trspec_proc


def _extend_extract(tr, tmin, tmax):
    deltat = tr.deltat
    itmin_frame = int(math.floor(tmin/deltat))
    itmax_frame = int(math.ceil(tmax/deltat))
    nframe = itmax_frame - itmin_frame
    n = tr.data_len()
    a = num.empty(nframe, dtype=num.float)
    itmin_tr = int(round(tr.tmin / deltat))
    itmax_tr = itmin_tr + n
    icut1 = min(max(0, itmin_tr - itmin_frame), nframe)
    icut2 = min(max(0, itmax_tr - itmin_frame), nframe)
    icut1_tr = min(max(0, icut1 + itmin_frame - itmin_tr), n)
    icut2_tr = min(max(0, icut2 + itmin_frame - itmin_tr), n)
    a[:icut1] = tr.ydata[0]
    a[icut1:icut2] = tr.ydata[icut1_tr:icut2_tr]
    a[icut2:] = tr.ydata[-1]
    tr = tr.copy(data=False)
    tr.tmin = tmin
    tr.set_ydata(a)
    return tr


def xjoin(basepath, path):
    if path is None and basepath is not None:
        return basepath
    elif op.isabs(path) or basepath is None:
        return path
    else:
        return op.join(basepath, path)


def xrelpath(path, start):
    if op.isabs(path):
        return path
    else:
        return op.relpath(path, start)


class Path(String):
    pass


class HasPaths(Object):
    path_prefix = Path.T(optional=True)

    def __init__(self, *args, **kwargs):
        Object.__init__(self, *args, **kwargs)
        self._basepath = None
        self._parent_path_prefix = None

    def set_basepath(self, basepath, parent_path_prefix=None):
        self._basepath = basepath
        self._parent_path_prefix = parent_path_prefix
        for (prop, val) in self.T.ipropvals(self):
            if isinstance(val, HasPaths):
                val.set_basepath(
                    basepath, self.path_prefix or self._parent_path_prefix)

    def get_basepath(self):
        assert self._basepath is not None
        return self._basepath

    def change_basepath(self, new_basepath, parent_path_prefix=None):
        assert self._basepath is not None

        self._parent_path_prefix = parent_path_prefix
        if self.path_prefix or not self._parent_path_prefix:

            self.path_prefix = op.normpath(xjoin(xrelpath(
                self._basepath, new_basepath), self.path_prefix))

        for val in self.T.ivals(self):
            if isinstance(val, HasPaths):
                val.change_basepath(
                    new_basepath, self.path_prefix or self._parent_path_prefix)

        self._basepath = new_basepath

    def expand_path(self, path, extra=None):
        assert self._basepath is not None

        if extra is None:
            def extra(path):
                return path

        path_prefix = self.path_prefix or self._parent_path_prefix

        if path is None:
            return None
        elif isinstance(path, basestring):
            return extra(
                op.normpath(xjoin(self._basepath, xjoin(path_prefix, path))))
        else:
            return [
                extra(
                    op.normpath(xjoin(self._basepath, xjoin(path_prefix, p))))
                for p in path]


class RandomResponse(trace.FrequencyResponse):

    scale = Float.T(default=0.0)

    def set_random_state(self, rstate):
        self._rstate = rstate

    def evaluate(self, freqs):
        n = freqs.size
        return 1.0 + (
            self._rstate.normal(scale=self.scale, size=n) +
            0.0J * self._rstate.normal(scale=self.scale, size=n))


class SyntheticWaveformNotAvailable(Exception):
    pass


class SyntheticTest(Object):
    inject_solution = Bool.T(default=False)
    respect_data_availability = Bool.T(default=False)
    real_noise_scale = Float.T(default=0.0)
    white_noise_scale = Float.T(default=0.0)
    random_response_scale = Float.T(default=0.0)
    real_noise_toffset = Float.T(default=-3600.)
    random_seed = Int.T(optional=True)
    x = Dict.T(String.T(), Float.T())

    def __init__(self, **kwargs):
        Object.__init__(self, **kwargs)
        self._problem = None
        self._synthetics = None

    def set_problem(self, problem):
        self._problem = problem
        self._synthetics = None

    def get_problem(self):
        if self._problem is None:
            raise SyntheticWaveformNotAvailable(
                'SyntheticTest.set_problem() has not been called yet')

        return self._problem

    def get_x(self):
        problem = self.get_problem()
        if self.x:
            x = problem.preconstrain(
                problem.parameter_array(self.x))

        else:
            x = problem.preconstrain(
                problem.pack(
                    problem.base_source))

        return x

    def get_synthetics(self):
        problem = self.get_problem()
        if self._synthetics is None:
            x = self.get_x()
            results = problem.forward(x)
            synthetics = {}
            for iresult, result in enumerate(results):
                tr = result.trace.pyrocko_trace()
                tfade = tr.tmax - tr.tmin
                tr.extend(tr.tmin - tfade, tr.tmax + tfade)

                if self.random_response_scale != 0:
                    tf = RandomResponse(scale=self.random_response_scale)
                    rstate = num.random.RandomState(
                        (self.random_seed or 0) + iresult)
                    tf.set_random_state(rstate)
                    tr = tr.transfer(
                        tfade=tfade,
                        transfer_function=tf)

                if self.white_noise_scale != 0.0:
                    rstate = num.random.RandomState(
                        (self.random_seed or 0) + iresult)
                    u = rstate.normal(
                        scale=self.white_noise_scale,
                        size=tr.data_len())

                    tr.ydata += u

                synthetics[result.trace.codes] = tr

            self._synthetics = synthetics

        return self._synthetics

    def get_waveform(self, nslc, tmin, tmax, tfade=0., freqlimits=None):
        synthetics = self.get_synthetics()

        if nslc not in synthetics:
            return None

        tr = synthetics[nslc]
        tr.extend(tmin - tfade * 2.0, tmax + tfade * 2.0)

        tr = tr.transfer(
            tfade=tfade,
            freqlimits=freqlimits)

        tr.chop(tmin, tmax)
        return tr


class DatasetConfig(HasPaths):

    stations_path = Path.T(optional=True)
    stations_stationxml_paths = List.T(Path.T())
    events_path = Path.T()
    waveform_paths = List.T(Path.T())
    clippings_path = Path.T(optional=True)
    responses_sacpz_path = Path.T(optional=True)
    responses_stationxml_paths = List.T(Path.T())
    station_corrections_path = Path.T(optional=True)
    apply_correction_factors = Bool.T(default=True)
    apply_correction_delays = Bool.T(default=True)
    picks_paths = List.T(Path.T())
    blacklist_paths = List.T(Path.T())
    blacklist = List.T(
        String.T(),
        help='stations/components to be excluded according to their STA, '
             'NET.STA, NET.STA.LOC, or NET.STA.LOC.CHA codes.')
    whitelist_paths = List.T(Path.T())
    whitelist = List.T(
        String.T(),
        optional=True,
        help='if not None, list of stations/components to included according '
             'to their STA, NET.STA, NET.STA.LOC, or NET.STA.LOC.CHA codes. '
             'Note: ''when whitelisting on channel level, both, the raw and '
             'the processed channel codes have to be listed.')
    synthetic_test = SyntheticTest.T(optional=True)

    def __init__(self, *args, **kwargs):
        HasPaths.__init__(self, *args, **kwargs)
        self._ds = {}

    def get_event_names(self):
        def extra(path):
            return expand_template(path, dict(
                event_name='*'))

        def fp(path):
            return self.expand_path(path, extra=extra)

        events = []
        for fn in glob.glob(fp(self.events_path)):
            events.extend(model.load_events(filename=fn))

        event_names = [ev.name for ev in events]
        return event_names

    def get_dataset(self, event_name):
        if event_name not in self._ds:
            def extra(path):
                return expand_template(path, dict(
                    event_name=event_name))

            def fp(path):
                return self.expand_path(path, extra=extra)

            ds = dataset.Dataset(event_name)
            ds.add_stations(
                pyrocko_stations_filename=fp(self.stations_path),
                stationxml_filenames=fp(self.stations_stationxml_paths))

            ds.add_events(filename=fp(self.events_path))
            ds.add_waveforms(paths=fp(self.waveform_paths))
            if self.clippings_path:
                ds.add_clippings(markers_filename=fp(self.clippings_path))

            if self.responses_sacpz_path:
                ds.add_responses(
                    sacpz_dirname=fp(self.responses_sacpz_path))

            if self.responses_stationxml_paths:
                ds.add_responses(
                    stationxml_filenames=fp(self.responses_stationxml_paths))

            if self.station_corrections_path:
                ds.add_station_corrections(
                    filename=fp(self.station_corrections_path))

            ds.apply_correction_factors = self.apply_correction_factors
            ds.apply_correction_delays = self.apply_correction_delays

            for picks_path in self.picks_paths:
                ds.add_picks(
                    filename=fp(picks_path))

            ds.add_blacklist(self.blacklist)
            ds.add_blacklist(filenames=fp(self.blacklist_paths))
            if self.whitelist:
                ds.add_whitelist(self.whitelist)
            if self.whitelist_paths:
                ds.add_whitelist(filenames=fp(self.whitelist_paths))

            ds.set_synthetic_test(copy.deepcopy(self.synthetic_test))
            self._ds[event_name] = ds

        return self._ds[event_name]


def weed(origin, targets, limit, neighborhood=3):

    azimuths = num.zeros(len(targets))
    dists = num.zeros(len(targets))
    for i, target in enumerate(targets):
        _, azimuths[i] = target.azibazi_to(origin)
        dists[i] = target.distance_to(origin)

    badnesses = num.ones(len(targets), dtype=float)
    deleted, meandists_kept = weeding.weed(
        azimuths, dists, badnesses,
        nwanted=limit,
        neighborhood=neighborhood)

    targets_weeded = [
        target for (delete, target) in zip(deleted, targets) if not delete]

    return targets_weeded, meandists_kept, deleted


class TargetConfig(Object):

    super_group = gf.StringID.T(default='', optional=True)
    group = gf.StringID.T(optional=True)
    distance_min = Float.T(optional=True)
    distance_max = Float.T(optional=True)
    distance_3d_min = Float.T(optional=True)
    distance_3d_max = Float.T(optional=True)
    depth_min = Float.T(optional=True)
    depth_max = Float.T(optional=True)
    limit = Int.T(optional=True)
    channels = List.T(String.T())
    inner_misfit_config = InnerMisfitConfig.T()
    interpolation = gf.InterpolationMethod.T()
    store_id = gf.StringID.T()
    weight = Float.T(default=1.0)

    def get_targets(self, ds, event, default_group):

        origin = event

        targets = []
        for st in ds.get_stations():
            for cha in self.channels:
                if ds.is_blacklisted((st.nsl() + (cha,))):
                    continue

                target = MisfitTarget(
                    quantity='displacement',
                    codes=st.nsl() + (cha,),
                    lat=st.lat,
                    lon=st.lon,
                    depth=st.depth,
                    interpolation=self.interpolation,
                    store_id=self.store_id,
                    misfit_config=self.inner_misfit_config,
                    manual_weight=self.weight,
                    super_group=self.super_group,
                    group=self.group or default_group)

                if self.distance_min is not None and \
                        target.distance_to(origin) < self.distance_min:
                    continue

                if self.distance_max is not None and \
                        target.distance_to(origin) > self.distance_max:
                    continue

                if self.distance_3d_min is not None and \
                        target.distance_3d_to(origin) < self.distance_3d_min:
                    continue

                if self.distance_3d_max is not None and \
                        target.distance_3d_to(origin) > self.distance_3d_max:
                    continue

                if self.depth_min is not None and \
                        target.depth < self.depth_min:
                    continue

                if self.depth_max is not None and \
                        target.depth > self.depth_max:
                    continue

                azi, _ = target.azibazi_to(origin)
                if cha == 'R':
                    target.azimuth = azi - 180.
                    target.dip = 0.
                elif cha == 'T':
                    target.azimuth = azi - 90.
                    target.dip = 0.
                elif cha == 'Z':
                    target.azimuth = 0.
                    target.dip = -90.

                target.set_dataset(ds)
                targets.append(target)

        if self.limit:
            return weed(origin, targets, self.limit)[0]
        else:
            return targets


class AnalyserConfig(Object):
    niter = Int.T(default=1000)


class SamplerDistributionChoice(StringChoice):
    choices = ['multivariate_normal', 'normal']


class SolverConfig(Object):
    niter_uniform = Int.T(default=1000)
    niter_transition = Int.T(default=0)
    niter_explorative = Int.T(default=10000)
    niter_non_explorative = Int.T(default=0)
    sampler_distribution = SamplerDistributionChoice.T(
        default='multivariate_normal')
    scatter_scale_transition = Float.T(default=2.0)
    scatter_scale = Float.T(default=1.0)
    chain_length_factor = Float.T(default=8.0)
    compensate_excentricity = Bool.T(default=True)

    def get_solver_kwargs(self):
        return dict(
            niter_uniform=self.niter_uniform,
            niter_transition=self.niter_transition,
            niter_explorative=self.niter_explorative,
            niter_non_explorative=self.niter_non_explorative,
            sampler_distribution=self.sampler_distribution,
            scatter_scale_transition=self.scatter_scale_transition,
            scatter_scale=self.scatter_scale,
            chain_length_factor=self.chain_length_factor,
            compensate_excentricity=self.compensate_excentricity)


class EngineConfig(HasPaths):
    gf_stores_from_pyrocko_config = Bool.T(default=True)
    gf_store_superdirs = List.T(Path.T())
    gf_store_dirs = List.T(Path.T())

    def __init__(self, *args, **kwargs):
        HasPaths.__init__(self, *args, **kwargs)
        self._engine = None

    def get_engine(self):
        if self._engine is None:
            fp = self.expand_path
            self._engine = gf.LocalEngine(
                use_config=self.gf_stores_from_pyrocko_config,
                store_superdirs=fp(self.gf_store_superdirs),
                store_dirs=fp(self.gf_store_dirs))

        return self._engine


class Config(HasPaths):
    rundir_template = Path.T()
    dataset_config = DatasetConfig.T()
    target_configs = List.T(TargetConfig.T())
    problem_config = ProblemConfig.T()
    analyser_config = AnalyserConfig.T(default=AnalyserConfig.D())
    solver_config = SolverConfig.T(default=SolverConfig.D())
    engine_config = EngineConfig.T(default=EngineConfig.D())

    def __init__(self, *args, **kwargs):
        HasPaths.__init__(self, *args, **kwargs)

    def get_event_names(self):
        return self.dataset_config.get_event_names()

    def get_dataset(self, event_name):
        return self.dataset_config.get_dataset(event_name)

    def get_targets(self, event):
        ds = self.get_dataset(event.name)

        targets = []
        for igroup, target_config in enumerate(self.target_configs):
            targets.extend(target_config.get_targets(
                ds, event, 'group_%i' % igroup))

        return targets

    def setup_modelling_environment(self, problem):
        problem.set_engine(self.engine_config.get_engine())
        ds = self.get_dataset(problem.base_source.name)
        synt = ds.synthetic_test
        if synt:
            synt.set_problem(problem)
            problem.base_source = problem.unpack(synt.get_x())

    def get_problem(self, event):
        targets = self.get_targets(event)
        problem = self.problem_config.get_problem(event, targets)
        self.setup_modelling_environment(problem)
        return problem


def sarr(a):
    return ' '.join('%15g' % x for x in a)


def load_problem_info_and_data(dirname, subset=None):
    problem = load_problem_info(dirname)
    xs, misfits = load_problem_data(xjoin(dirname, subset), problem)
    return problem, xs, misfits


def load_problem_info(dirname):
    fn = op.join(dirname, 'problem.yaml')
    return guts.load(filename=fn)


def load_problem_data(dirname, problem):
    fn = op.join(dirname, 'x')
    with open(fn, 'r') as f:
        nmodels = os.fstat(f.fileno()).st_size / (problem.nparameters * 8)
        data = num.fromfile(
            f, dtype='<f8',
            count=nmodels*problem.nparameters).astype(num.float)

    nmodels = data.size/problem.nparameters
    xs = data.reshape((nmodels, problem.nparameters))

    fn = op.join(dirname, 'misfits')
    with open(fn, 'r') as f:
        data = num.fromfile(
            f, dtype='<f8', count=nmodels*problem.ntargets*2).astype(num.float)

    data = data.reshape((nmodels, problem.ntargets*2))

    combi = num.empty_like(data)
    combi[:, 0::2] = data[:, :problem.ntargets]
    combi[:, 1::2] = data[:, problem.ntargets:]

    misfits = combi.reshape((nmodels, problem.ntargets, 2))

    return xs, misfits


def get_mean_x(xs):
    return num.mean(xs, axis=0)


def get_mean_x_and_gm(problem, xs, misfits):
    gms = problem.global_misfits(misfits)
    return num.mean(xs, axis=0), num.mean(gms)


def get_best_x(problem, xs, misfits):
    gms = problem.global_misfits(misfits)
    ibest = num.argmin(gms)
    return xs[ibest, :]


def get_best_x_and_gm(problem, xs, misfits):
    gms = problem.global_misfits(misfits)
    ibest = num.argmin(gms)
    return xs[ibest, :], gms[ibest]


def get_mean_source(problem, xs):
    x_mean = get_mean_x(xs)
    source = problem.unpack(x_mean)
    return source


def get_best_source(problem, xs, misfits):
    x_best = get_best_x(problem, xs, misfits)
    source = problem.unpack(x_best)
    return source


def mean_latlondist(lats, lons):
    if len(lats) == 0:
        return 0., 0., 1000.
    else:
        ns, es = od.latlon_to_ne_numpy(lats[0], lons[0], lats, lons)
        n, e = num.mean(ns), num.mean(es)
        dists = num.sqrt((ns-n)**2 + (es-e)**2)
        lat, lon = od.ne_to_latlon(lats[0], lons[0], n, e)
        return float(lat), float(lon), float(num.max(dists))


def stations_mean_latlondist(stations):
    lats = num.array([s.lat for s in stations])
    lons = num.array([s.lon for s in stations])
    return mean_latlondist(lats, lons)


def read_config(path):
    config = load(filename=path)
    if not isinstance(config, Config):
        raise GrondError('invalid Grond configuration in file "%s"' % path)

    config.set_basepath(op.dirname(path) or '.')
    return config


def write_config(config, path):
    basepath = config.get_basepath()
    dirname = op.dirname(path) or '.'
    config.change_basepath(dirname)
    guts.dump(config, filename=path)
    config.change_basepath(basepath)


def analyse(problem, niter=1000, show_progress=False):
    if niter == 0:
        return

    wtargets = []
    for target in problem.targets:
        wtarget = copy.copy(target)
        wtarget.flip_norm = True
        wtarget.weight = 1.0
        wtargets.append(wtarget)

    groups, ngroups = problem.get_group_mask()

    wproblem = problem.copy()
    wproblem.targets = wtargets

    xbounds = num.array(wproblem.bounds(), dtype=num.float)
    npar = xbounds.shape[0]

    mss = num.zeros((niter, problem.ntargets))
    rstate = num.random.RandomState(123)

    if show_progress:
        pbar = util.progressbar('analysing problem', niter)

    isbad_mask = None
    for iiter in xrange(niter):
        while True:
            x = []
            for ipar in xrange(npar):
                v = rstate.uniform(xbounds[ipar, 0], xbounds[ipar, 1])
                x.append(v)

            try:
                x = wproblem.preconstrain(x)
                break

            except Forbidden:
                pass

        if isbad_mask is not None and num.any(isbad_mask):
            isok_mask = num.logical_not(isbad_mask)
        else:
            isok_mask = None

        _, ms = wproblem.evaluate(x, mask=isok_mask)
        mss[iiter, :] = ms

        isbad_mask = num.isnan(ms)

        if show_progress:
            pbar.update(iiter)

    if show_progress:
        pbar.finish()

    mean_ms = num.mean(mss, axis=0)

    weights = 1.0 / mean_ms
    for igroup in xrange(ngroups):
        weights[groups == igroup] /= (
            num.nansum(weights[groups == igroup]) /
            num.nansum(num.isfinite(weights[groups == igroup])))

    for weight, target in zip(weights, problem.targets):
        target.analysis_result = TargetAnalysisResult(
            balancing_weight=float(weight))


def excentricity_compensated_probabilities(xs, sbx, factor):
    inonflat = num.where(sbx != 0.0)[0]
    scale = num.zeros_like(sbx)
    scale[inonflat] = 1.0 / (sbx[inonflat] * (factor if factor != 0. else 1.0))
    #distances_all = math.sqrt(num.sum(
    #    ((xs[num.newaxis, :, :] - xs[:, num.newaxis, :]) *
    #     scale[num.newaxis, num.newaxis, :])**2, axis=2))
    distances_sqr_all = num.sum(
        ((xs[num.newaxis, :, :] - xs[:, num.newaxis, :]) *
         scale[num.newaxis, num.newaxis, :])**2, axis=2)
    probabilities = 1.0 / num.sum(distances_sqr_all < 1.0, axis=1)
    # print num.sort(num.sum(distances_sqr_all < 1.0, axis=1))
    probabilities /= num.sum(probabilities)
    return probabilities


def excentricity_compensated_choice(xs, sbx, factor):
    probabilities = excentricity_compensated_probabilities(
        xs, sbx, factor)
    r = num.random.random()
    ichoice = num.searchsorted(num.cumsum(probabilities), r)
    ichoice = min(ichoice, xs.shape[0]-1)
    return ichoice


def select_most_excentric(xcandidates, xs, sbx, factor):
    inonflat = num.where(sbx != 0.0)[0]
    scale = num.zeros_like(sbx)
    scale[inonflat] = 1.0 / (sbx[inonflat] * (factor if factor != 0. else 1.0))
    distances_sqr_all = num.sum(
        ((xcandidates[num.newaxis, :, :] - xs[:, num.newaxis, :]) *
         scale[num.newaxis, num.newaxis, :])**2, axis=2)
    #print num.sort(num.sum(distances_sqr_all < 1.0, axis=0))
    ichoice = num.argmin(num.sum(distances_sqr_all < 1.0, axis=0))
    return xcandidates[ichoice]


def solve(problem,
          rundir=None,
          niter_uniform=1000,
          niter_transition=1000,
          niter_explorative=10000,
          niter_non_explorative=0,
          scatter_scale_transition=2.0,
          scatter_scale=1.0,
          chain_length_factor=8.0,
          xs_inject=None,
          sampler_distribution='multivariate_normal',
          compensate_excentricity=True,
          status=(),
          plot=False):

    xbounds = num.array(problem.bounds(), dtype=num.float)
    npar = xbounds.shape[0]

    nlinks_cap = int(round(chain_length_factor * npar + 1))
    chains_m = num.zeros((1 + problem.nbootstrap, nlinks_cap), num.float)
    chains_i = num.zeros((1 + problem.nbootstrap, nlinks_cap), num.int)
    nlinks = 0
    mbx = None

    if xs_inject is not None and xs_inject.size != 0:
        niter_inject = xs_inject.shape[0]
    else:
        niter_inject = 0

    niter = niter_inject + niter_uniform + niter_transition + \
        niter_explorative + niter_non_explorative

    iiter = 0
    sbx = None
    mxs = None
    covs = None
    xhist = num.zeros((niter, npar))
    isbad_mask = None
    accept_sum = num.zeros(1 + problem.nbootstrap, dtype=num.int)
    accept_hist = num.zeros(niter, dtype=num.int)
    pnames = [p.name for p in problem.parameters]

    if plot:
        from matplotlib import pyplot as plt
        from grond import plot as gplot
        #plt.ion()
        #plt.show()
        solver_plot = gplot.SolverPlot(problem, plt)

    while iiter < niter:
        jchoice = None
        ichoice = None

        if iiter < niter_inject:
            phase = 'inject'
        elif iiter < niter_inject + niter_uniform:
            phase = 'uniform'
        elif iiter < niter_inject + niter_uniform + niter_transition:
            phase = 'transition'
        elif iiter < niter_inject + niter_uniform + niter_transition + \
                niter_explorative:
            phase = 'explorative'
        else:
            phase = 'non_explorative'

        factor = 0.0
        if phase == 'transition':
            T = float(niter_transition)
            A = scatter_scale_transition
            B = scatter_scale
            tau = T/(math.log(A) - math.log(B))
            t0 = math.log(A) * T / (math.log(A) - math.log(B))
            t = float(iiter - niter_uniform - niter_inject)
            factor = num.exp(-(t-t0) / tau)

        elif phase in ('explorative', 'non_explorative'):
            factor = scatter_scale

        ntries_preconstrain = 0
        ntries_sample = 0

        if phase == 'inject':
            x = xs_inject[iiter, :]
        else:
            while True:
                ntries_preconstrain += 1

                if mbx is None or phase == 'uniform':
                    x = problem.random_uniform(xbounds)
                else:
                    # jchoice = num.random.randint(0, 1 + problem.nbootstrap)
                    jchoice = num.argmin(accept_sum)

                    if phase in ('transition', 'explorative'):

                        if compensate_excentricity:
                            ichoice = excentricity_compensated_choice(
                                xhist[chains_i[jchoice, :], :], sbx, 3.)

                            xchoice = xhist[chains_i[jchoice, ichoice], :]

                        else:
                            ichoice = num.random.randint(0, nlinks)
                            xchoice = xhist[chains_i[jchoice, ichoice]]
                    else:
                        xchoice = mxs[jchoice]

                    if sampler_distribution == 'multivariate_normal':
                        ntries_sample = 0

                        ntry = 0
                        ok_mask_sum = num.zeros(npar, dtype=num.int)
                        while True:
                            ntries_sample += 1
                            vs = num.random.multivariate_normal(
                                xchoice, factor**2 * covs[jchoice])