'''
example.py: opens and plots experimental data (channel impulse response measurements)
Copyright (c) 2020 ETH Zurich, Anton Ledergerber

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
'''
#############################################################
# python packages
#############################################################
import sys, os
import numpy as np
import pandas as pd
from scipy.interpolate import interp1d
import matplotlib.pylab as plt

#############################################################
# configurations
#############################################################
FPOffset = -3 # offset of logged samples from first path sample as determined by Decawave's leading edge detection algorithm.
numCIRSamples = 31 # number of logged CIR samples
Ts = 1/(2*(4*124.8e6)) # (s) sampling period for CIR measurement
anchorUWBIds = [0,1,2,4] # device Ids
anchorVIds = [160, 161, 162, 164] # motion capture device Ids
dynObjVId = 177 # motion capture dynamic object Id
antennaOffset = np.array([0.14, 0.063, 0.010]) # (m) antenna offset from motion capture object body
centerOffset = np.array([0.2775, 0.15, 0]) # (m) offset of center from motion capture dynamic object body
lightSpeed = 299792458 # (m/s) speed of light

#############################################################
# functions
#############################################################
def quat_to_rot_mat(quat):
    """
    convert quaternion (w,x,y,z) !hamiltonian notation! (ijk=-1) to rotation matrix
    Malcolm D. Shuster, A Survey of Attitude Representations, Equation (158)
    A rotation from frame A to frame B equals a coordinate transformation from A to B (Shuster Equation (87))
    see also Joan Sola Quaternion kinematics for the error state KF for a good overview over the different conventions
    :param quat: Nx4 (hamiltonian notation)
    :return: Nx3x3
    """
    r0 = quat[:, 0]*quat[:, 0]
    r1 = quat[:, 1]*quat[:, 1]
    r2 = quat[:, 2]*quat[:, 2]
    r3 = quat[:, 3]*quat[:, 3]

    R00 = r0 + r1 - r2 - r3
    R10 = 2 * quat[:, 1] * quat[:, 2] - 2 * quat[:, 0] * quat[:, 3]
    R20 = 2 * quat[:, 1] * quat[:, 3] + 2 * quat[:, 0] * quat[:, 2]

    R01 = 2 * quat[:, 1] * quat[:, 2] + 2 * quat[:, 0] * quat[:, 3]
    R11 = r0 - r1 + r2 - r3
    R21 = 2 * quat[:, 2] * quat[:, 3] - 2 * quat[:, 0] * quat[:, 1]

    R02 = 2 * quat[:, 1] * quat[:, 3] - 2 * quat[:, 0] * quat[:, 2]
    R12 = 2 * quat[:, 2] * quat[:, 3] + 2 * quat[:, 0] * quat[:, 1]
    R22 = r0 - r1 - r2 + r3

    R0 = np.stack((R00, R10, R20), axis=-1)
    R1 = np.stack((R01, R11, R21), axis=-1)
    R2 = np.stack((R02, R12, R22), axis=-1)
    R = np.stack((R0, R1, R2), axis=-1)
    return R

#############################################################
# load logs
#############################################################
# get directory of this file
pythonFileName = sys.argv[0]
pythonFilePath = os.path.dirname(pythonFileName)

UWBLogFile = os.path.join(pythonFilePath,"UWBDefaultCH4.pkl")
MCLogFile = os.path.join(pythonFilePath,"ViconDefaultCH4.pkl")

UWBLog = pd.read_pickle(UWBLogFile)
MCLog = pd.read_pickle(MCLogFile)

#############################################################
# process UWB logs
#############################################################

numUWBMeas = len(UWBLog)
fp = np.array(UWBLog['fp_loc'], dtype=np.float) / 64
tOffset = (fp%1)*Ts*1e9 # (ns) FP offset with respect to FP sample 

accMem = np.zeros((numUWBMeas, numCIRSamples * 4), dtype=np.uint8)
CIRReal = np.zeros((numUWBMeas, numCIRSamples))
CIRImag = np.zeros((numUWBMeas, numCIRSamples))

for j in range(numUWBMeas):
    accMem[j, :] = np.frombuffer(UWBLog['p_acc_mem'][j], dtype=np.uint8)[1:]  # discard first byte

for ii in range(numCIRSamples):
    CIRReal[:, ii] = np.int16(np.uint8(accMem[:, ii * 4]) + np.left_shift(np.uint8(accMem[:, ii * 4 + 1]),
                                                                            8 * np.ones((accMem.shape[0],),
                                                                                        dtype=np.int)))
    CIRImag[:, ii] = np.int16(
        np.uint8(accMem[:, ii * 4 + 2]) + np.left_shift(np.uint8(accMem[:, ii * 4 + 3]),
                                                         8 * np.ones((accMem.shape[0],), dtype=np.int)))
accCount = UWBLog['rxpacc']

CIRReal = CIRReal / np.expand_dims(accCount, axis=-1)
CIRImag = CIRImag / np.expand_dims(accCount, axis=-1)
CIRMag = np.sqrt(CIRReal ** 2 + CIRImag ** 2)

txId = UWBLog['txId']
txId = txId.to_numpy(dtype=np.int32)
rxId = UWBLog['rxId']
rxId = rxId.to_numpy(dtype=np.int32)
tUWB = UWBLog['time']

rxIds = np.unique(rxId)
txIds = np.unique(txId)

########################################################
# process motion capture data
########################################################

# calculate antenna position and attitude
avgAnchorPos = {}
for vid in anchorVIds:
    # position of anchor vicon body in inertial frame
    anchorLog = MCLog.loc[MCLog['vid']==vid]
    pos = anchorLog[['x','y','z']].to_numpy()
    # transformation matrix from inertial frame to vicon body frame
    A_BI = quat_to_rot_mat(anchorLog[['q0','q1','q2','q3']].to_numpy())
    A_IB = np.transpose(A_BI, axes=[0,2,1])
    # position of anchor antenna in inertial frame
    pos = pos + np.matmul(A_IB, antennaOffset)
    avgAnchorPos[vid-160] = np.mean(pos[:,0:2], axis=0)

# calculate dynamic object position and attitude
dynObjLog = MCLog.loc[MCLog['vid']==dynObjVId]
pos = dynObjLog[['x','y','z']].to_numpy()
A_BI = quat_to_rot_mat(dynObjLog[['q0', 'q1', 'q2', 'q3']].to_numpy())
A_IB = np.transpose(A_BI, axes=[0, 2, 1])
dynObjPos = pos + np.matmul(A_IB, centerOffset)

########################################################
# synchronize motion capture data to UWB data
########################################################

# interpolate based on laptop reception time
tInit = UWBLog['time'].iloc[0]
tUWB = (tUWB-tInit).to_numpy().astype(dtype=np.float)*1e-9
tDynObjMC = (dynObjLog['time']-tInit).to_numpy().astype(dtype=np.float)*1e-9

# interpolate
interpfpos_dynObj = interp1d(tDynObjMC, dynObjPos, kind='nearest', bounds_error=False,
                             fill_value=(dynObjPos[0,:], dynObjPos[-1,:]), axis=0)
posDynObjInterp = interpfpos_dynObj(tUWB)

#############################################################
# plot accumulated CIR and corresponding target path location
#############################################################

# select time, module Ids, and number of accumulated CIR measurements
tUWB0 = 23
tx = 0
rx = 1
numAccCIRMeas = 50

# calculate corresponding target path location
# based on motion capture data
idxtUWB0 = np.searchsorted(tUWB, tUWB0)
posDynObj0 = posDynObjInterp[idxtUWB0,0:2]
posTx = avgAnchorPos[tx]
posRx = avgAnchorPos[rx]
distTxDynObj0 = np.linalg.norm(posDynObj0-posTx)
distRxDynObj0 = np.linalg.norm(posDynObj0-posRx)
distTxRx = np.linalg.norm(posTx-posRx)
tau = (distRxDynObj0+distTxDynObj0-distTxRx)/lightSpeed*1e9

# select CIR measurements
txRxMask = np.logical_and(tx==txId, rx==rxId)
txRxIdx = np.nonzero(txRxMask)[0]
idxTxRx0 = np.searchsorted(tUWB[txRxMask], tUWB0)
idxTxRxAcc0 = txRxIdx[idxTxRx0:(idxTxRx0+numAccCIRMeas)]

CIRSampleTimes = np.expand_dims(np.arange(start=FPOffset, stop=FPOffset+(numCIRSamples-0.5)*Ts*1e9, step=Ts*1e9, dtype=np.float), axis=0)
tOffset = np.expand_dims(tOffset, axis=-1)

plt.figure()
plt.title('accumulated CIR at t={:.1f} for tx:{:} rx:{:}'.format(tUWB0, tx, rx))
plt.plot(np.transpose(CIRSampleTimes-tOffset[idxTxRxAcc0]), np.transpose(CIRMag[idxTxRxAcc0,:]), '.')
plt.plot([tau, tau], [0, 60], 'k-', label='target path location')
plt.legend()
plt.xlabel('(ns)')
plt.ylabel('(a.u.)')

plt.show()