UnbindingBCDashboardSection not reporting warps correctly

[alexrd]

No warps show up in the dashboard reporter even though there are warping events in the HDF5.

[wpena1]

Hi, Using the example lisozyme system in implicit solvent I noticed this issue too, no warps are reported and the rate (tau) is reported as 0.0. I ran with 48 walkers, 1000 cycles and for 10000 steps per cycle (dt = 0.002 ps). At the end of the run there were 52 exit points. I tried to calculate the unbinding rate using the following.

from wepy.resampling.resamplers.wexplore import WExploreResampler
from wepy.boundary_conditions.receptor import UnbindingBC
from wepy.hdf5 import WepyHDF5
from wepy.analysis.contig_tree import ContigTree
from wepy.analysis.profiles import ContigTreeProfiler

h5_path = './result.wepy.h5'
wepy_h5 = WepyHDF5(h5_path, mode='r+')
with wepy_h5:
    contigtree = ContigTree(wepy_h5, decision_class=WExploreResampler.DECISION,
                                                            boundary_condition_class=UnbindingBC)
from wepy.analysis.rates import contig_warp_rates
with wepy_h5:
    sample = contigtree.make_contig(contigtree.spanning_contig_traces()[0])
    results = contig_warp_rates(sample,20) #  20 ps per cycle 

total_weight = results[-1][0][0]
flux = results[-1][0][1]
sampling_time = results[-1][0][2]

rate = flux

is that the right approach to get an unbinding rate? Thanks

[alexrd]

I tried this on one of my files and the answer was off. I'm not 100% sure why since I haven't used those functions. Instead I would use this much simpler approach (it's faster too):

import numpy as np
wepy_h5.open()

time_per_cycle = 20 # ps

# get the total elapsed time of the simulation
tot_time = wepy_h5.num_traj_frames(0,0) * time_per_cycle

# get the sum of all the warped walkers
tot_weight = np.array(wepy_h5.h5['runs/0/warping/weight']).sum()

rate = tot_weight/tot_time  # in ps^-1
rate_s = rate*1e12          # rate in s^-1

This assumes that you only have one run in your h5 file. If you have more than one you can add up the weights and elapsed times in each run.

[wpena1]

Will give this a try. Thanks a lot.