#!/use/bin/env python

import logging

import numpy as num

from pyrocko import trace

from pyrocko import model

from pyrocko import pile

from pyrocko import gui_util

from pyrocko.gf import LocalEngine

from autogain import autogain, util_optic

logging.basicConfig(level=logging.INFO)

logger = logging.getLogger()

km = 1000.

if __name__ == '__main__':

# Where is your data stored?:

datapath = '/media/usb/webnet/pole_zero/restituted_displacement/2013Mar'

data_pile = pile.make_pile(datapath)

# And the stations:

stations = model.load_stations('/media/usb/webnet/meta/stations.pf')

# Station code of the trace you want to scale agains:

reference_id ='NKC'

# Frequency band to use:

fband = {'order':4, 'corner_hp':1.0, 'corner_lp':4.}

# And a taper to avoid filtering artefacts.

taper = trace.CosFader(xfrac=0.25)

# Define a window to chop traces. In this case a static length of 20

# seconds will be used and the synthetic phase arrival will be in the

# center. The relative position can be changed between 0 (phase at t=0) and

# 1 (phase at t=tmax).

window = autogain.StaticLengthWindow(static_length=20.,

phase_position=0.5)

#candidate_fn = 'candidates2013new.pf'

#candidates = [m.get_event() for m in gui_util.Marker.load_markers(candidate_fn)]

# If you have a catalog of events that you want to use, load them and make

# and event EventCollection from those events:

#candidates = model.load_events('events.pf')

#event_selector = autogain.EventCollection(events=candidates)

# I selected events that I wanted to use using snuffler and wrote the

# selected event markers into a file which is loaded here:

markers = gui_util.load_markers('candidates2013new.pf')

# Thus, the EventCollection accepts 'markers' or 'events' as kwargs:

event_selector = autogain.EventCollection(markers=markers)

# If you do not have a catalog, yet, you might want to automatically

# download events either from Geofon or Iris.

event_selector = autogain.EventSelectorCatalog(distmin=1000*km,

distmax=20000*km,

depthmin=2*km,

depthmax=600*km,

magmin=4.9)

# Setup the processing:

ag = autogain.AutoGain(data_pile, stations=stations,

event_selector=event_selector,

component='Z',

reference_nsl=reference_id,

phase_selection='first(P|p)',

scale_one=False)

# Use interpolated travel times from the fomosto store 'globalttt'

# Either you can download that store from kinherd.org but it's very large.

# What is needed is only the ttt. So you can generate it yourself using fomosto.

#phases = LocalEngine(store_superdirs=['/data/stores'],

# default_store_id='globalttt').get_store()

# ALTERNATIVELY, you can use the PhasePie which calles pyrocko's cake and

# calculates travel times on demand. Might take longer:

phases = autogain.PhasePie()

ag.set_phaser(phases)

ag.set_window(window)

# start processing:

ag.process(fband, taper)

# Store results in YAML format:

ag.save_mean('gains.txt')

optics = util_optic.Optics(ag)

optics.plot()

optics.show(file_name='gains.pdf')

optics.make_waveform_compare()

# Used events can be dumped to a file:

#event_selector.save_events('auto_events.pf')