def dB(day, peak_datetimes, instrument, current_dif, windows, probe_list, plot = False, lowpass = False, rand_noise=True): #for only one instrument
if day == 1:
if windows:
path_fol_A = r'C:\Users\jonas\MSci-Data\day_one\A'
path_fol_B = r'C:\Users\jonas\MSci-Data\day_one\B'
else:
path_fol_A = os.path.expanduser("~/Documents/MSciProject/Data/day_one/A")
path_fol_B = os.path.expanduser("~/Documents/MSciProject/Data/day_one/B")
elif day == 2:
if windows:
path_fol_A = r'C:\Users\jonas\MSci-Data\day_two\A'
path_fol_B = r'C:\Users\jonas\MSci-Data\day_two\B'
else:
path_fol_A = os.path.expanduser("~/Documents/MSciProject/Data/day_two/A")
path_fol_B = os.path.expanduser("~/Documents/MSciProject/Data/day_two/B")
#set start and end time to first and last current jump time with a 3 minute buffer either side
start_dt = peak_datetimes[0] - pd.Timedelta(minutes = 3)
end_dt = peak_datetimes[-1] + pd.Timedelta(minutes = 3)
sampling_freq = 1 #do we want to remove the high freq noise?
start_csv_A, end_csv_A = processing.which_csvs(True, day ,start_dt, end_dt, tz_MAG = True)
start_csv_B, end_csv_B = processing.which_csvs(False, day ,start_dt, end_dt, tz_MAG = True)
print(start_csv_A, end_csv_A)
all_files_A = [0]*(end_csv_A + 1 - start_csv_A)
#finding all the csv files that we need from the start and end time
if day == 1:
for index, j in enumerate(range(start_csv_A, end_csv_A + 1)): #this will loop through and add the csv files that contain the start and end time set above
if windows:
all_files_A[index] = path_fol_A + f'\SoloA_2019-06-21--08-10-10_{j}.csv'
else:
all_files_A[index] = path_fol_A + os.path.expanduser(f'/SoloA_2019-06-21--08-10-10_{j}.csv') #need to change path_fol_A to the path where your A folder is
all_files_B = [0]*(end_csv_B + 1 - start_csv_B)
for index, j in enumerate(range(start_csv_B, end_csv_B + 1)):
if windows:
all_files_B[index] = path_fol_B + f'\SoloB_2019-06-21--08-09-10_{j}.csv'
else:
all_files_B[index] = path_fol_B + os.path.expanduser(f'/SoloB_2019-06-21--08-09-10_{j}.csv') #need to change path_f
if day == 2:
for index, j in enumerate(range(start_csv_A, end_csv_A + 1)): #this will loop through and add the csv files that contain the start and end time set above
if windows:
all_files_A[index] = path_fol_A + f'\SoloA_2019-06-24--08-14-46_{j}.csv'
else:
all_files_A[index] = path_fol_A + os.path.expanduser(f'/SoloA_2019-06-24--08-14-46_{j}.csv') #need to change path_fol_A to the path where your A folder is
all_files_B = [0]*(end_csv_B + 1 - start_csv_B)
for index, j in enumerate(range(start_csv_B, end_csv_B + 1)):
if windows:
all_files_B[index] = path_fol_B + f'\SoloB_2019-06-24--08-14-24_{j}.csv'
else:
all_files_B[index] = path_fol_B + os.path.expanduser(f'/SoloB_2019-06-24--08-14-24_{j}.csv') #need to change path_f
vect_dict = {}
for i in probe_list:
#looping through each sensor
if i < 8:
soloA_bool = True
all_files = all_files_A
else:
soloA_bool = False
all_files = all_files_B
if i < 9:
num_str = f'0{i+1}'
else:
num_str = i+1
collist = ['time', f'Probe{num_str}_X', f'Probe{num_str}_Y', f'Probe{num_str}_Z']
#reads csv files and rotates to spacecraft frame
if soloA_bool:
df = processing.read_files(all_files, soloA_bool, sampling_freq, collist, day=day, start_dt = start_dt, end_dt = end_dt)
if day == 1:
rotate_mat = processing.rotate_21(soloA_bool)[i]
else:
rotate_mat = processing.rotate_24(soloA_bool)[i]
else:
df = processing.read_files(all_files, soloA_bool, sampling_freq, collist, day=day, start_dt = start_dt, end_dt = end_dt)
if day == 1:
rotate_mat = processing.rotate_21(soloA_bool)[i-8]
else:
rotate_mat = processing.rotate_24(soloA_bool)[i-8]
df.iloc[:,0:3] = np.matmul(rotate_mat, df.iloc[:,0:3].values.T).T
df = processing.shifttime(df, soloA_bool, day) # must shift MFSA data to MAG/spacecraft time
df = df.between_time(start_dt.time(), end_dt.time())
#lowpass filter to remove high frequency signals
if lowpass:
def butter_lowpass(cutoff, fs, order=10):
nyq = 0.5 * fs
normal_cutoff = cutoff / nyq
b, a = butter(order, normal_cutoff, btype='low', analog=False)
return b, a
def butter_lowpass_filter(data, cutoff, fs, order=10):
b, a = butter_lowpass(cutoff, fs, order=order)
y = lfilter(b, a, data)
return y
cutoff = 15
fs = sampling_freq
for axis in ['X','Y','Z']:
df[f'Probe{num_str}_{axis}'] = butter_lowpass_filter(df[f'Probe{num_str}_{axis}'], cutoff, fs)
step_dict = processing.calculate_dB(df, peak_datetimes)
#get data for dI
xdata = list(current_dif[:len(peak_datetimes)])
print(xdata)
#get data for dB
probe_x_tmp = step_dict.get(f'Probe{num_str}_X')
probe_y_tmp = step_dict.get(f'Probe{num_str}_Y')
probe_z_tmp = step_dict.get(f'Probe{num_str}_Z')
probe_x_tmp_err = step_dict.get(f'Probe{num_str}_X err')
probe_y_tmp_err = step_dict.get(f'Probe{num_str}_Y err')
probe_z_tmp_err = step_dict.get(f'Probe{num_str}_Z err')
#force zero forces the fir through the origin by adding this as a point - should theoretically go through the origin
force_zero = False
if force_zero:
xdata.append(0.0)
probe_x_tmp.append(0.0)
probe_y_tmp.append(0.0)
probe_z_tmp.append(0.0)
probe_x_tmp_err.append(0.0) #error on bonus point should be zero, but curve_fit requires finite error - and this forces the line through the origin anyway
probe_y_tmp_err.append(0.0)
probe_z_tmp_err.append(0.0)
if rand_noise:
#takes random noise and adds it in quadrature with the standard error of averaging
if day == 1:
err_path = f'..\\Results\\day1_mfsa_probe_vars.csv'
elif day == 2:
err_path = f'..\\Results\\day2_mfsa_probe_vars.csv'
df_err_correction = pd.read_csv(err_path)
df_err = df_err_correction.iloc[i]
#print(df_err, i, num_str)
probe_x_tmp_err = [df_err['Bx_var'] for k in range(len(xdata))]#[np.sqrt(k**2 + df_err['Bx_var']**2) for k in probe_x_tmp_err]
probe_y_tmp_err = [df_err['By_var'] for k in range(len(xdata))]#[np.sqrt(k**2 + df_err['By_var']**2) for k in probe_y_tmp_err]
probe_z_tmp_err = [df_err['Bz_var'] for k in range(len(xdata))]#[np.sqrt(k**2 + df_err['Bz_var']**2) for k in probe_z_tmp_err]
def line(x,a,b):
return a*x + b
params_x,cov_x = spo.curve_fit(line, xdata, probe_x_tmp[:], sigma = probe_x_tmp_err[:], absolute_sigma = True)
params_y,cov_y = spo.curve_fit(line, xdata, probe_y_tmp[:], sigma = probe_y_tmp_err[:], absolute_sigma = True)
params_z,cov_z = spo.curve_fit(line, xdata, probe_z_tmp[:], sigma = probe_z_tmp_err[:], absolute_sigma = True)
perr_x = np.sqrt(np.diag(cov_x))
perr_y = np.sqrt(np.diag(cov_y))
perr_z = np.sqrt(np.diag(cov_z))
print('spo.curve_fit')
print('Slope = ', params_x[0], '+/-', perr_x[0], 'Intercept = ', params_x[1], '+/-', perr_x[1])
print('Slope = ', params_y[0], '+/-', perr_y[0], 'Intercept = ', params_y[1], '+/-', perr_y[1])
print('Slope = ', params_z[0], '+/-', perr_z[0], 'Intercept = ', params_z[1], '+/-', perr_z[1])
else:
X = spstats.linregress(xdata, probe_x_tmp) #adding bonus point has little effect on grad - only changes intercept
Y = spstats.linregress(xdata, probe_y_tmp)
Z = spstats.linregress(xdata, probe_z_tmp)
print('sps.linregress')
print('Slope = ', X.slope, '+/-', X.stderr, ' Intercept = ', X.intercept)
print('Slope = ', Y.slope, '+/-', Y.stderr, ' Intercept = ', Y.intercept)
print('Slope = ', Z.slope, '+/-', Z.stderr, ' Intercept = ', Z.intercept)
if plot:
plt.figure()
if rand_noise:
plt.plot(xdata, params_x[0]*np.array(xdata) + params_x[1], 'b:', label = f'curve_fit - X grad: {round(params_x[0],2)} ± {round(perr_x[0],2)} int: {round(params_x[1],2)} ± {round(perr_x[1],2)}')
plt.plot(xdata, params_y[0]*np.array(xdata) + params_y[1], 'r:', label = f'curve_fit - Y grad: {round(params_y[0],2)} ± {round(perr_y[0],2)} int: {round(params_y[1],2)} ± {round(perr_y[1],2)}')
plt.plot(xdata, params_z[0]*np.array(xdata) + params_z[1], 'g:', label = f'curve_fit - Z grad: {round(params_z[0],2)} ± {round(perr_z[0],2)} int: {round(params_z[1],2)} ± {round(perr_z[1],2)}')
elif rand_noise == False:
plt.plot(xdata, X.intercept + X.slope*np.array(xdata), 'b-', label = f'X grad: {round(X.slope,2)} ± {round(X.stderr,2)} int: {round(X.intercept, 2)}')
plt.plot(xdata, Y.intercept + Y.slope*np.array(xdata), 'r-', label = f'Y grad: {round(Y.slope,2)} ± {round(Y.stderr,2)} int: {round(Y.intercept, 2)}')
plt.plot(xdata, Z.intercept + Z.slope*np.array(xdata), 'g-', label = f'Z grad: {round(Z.slope,2)} ± {round(Z.stderr,2)} int: {round(Z.intercept, 2)}')
plt.errorbar(xdata, probe_x_tmp, yerr = probe_x_tmp_err, fmt = 'bs', markeredgewidth = 2)
plt.errorbar(xdata, probe_y_tmp, yerr = probe_y_tmp_err, fmt = 'rs', markeredgewidth = 2)
plt.errorbar(xdata, probe_z_tmp, yerr = probe_z_tmp_err, fmt = 'gs', markeredgewidth = 2)
plt.legend(loc="best")
plt.title(f'{instrument} - Probe {num_str} - MFSA - Day {day_number}')
plt.xlabel('dI [A]')
plt.ylabel('dB [nT]')
plt.show()
save_all = False
if save_all:
dBdI = {}
for dI in range(len(xdata)):
dBdI[f'{dI+1}'] = [xdata[dI],probe_x_tmp[dI],probe_x_tmp_err[dI],probe_y_tmp[dI],probe_y_tmp_err[dI],probe_z_tmp[dI],probe_z_tmp_err[dI]]
w = csv.writer(open(f"..\\Results\\dBdI_data\\Day{day_number}\\1Hz_with_err\\{instrument}\\{instrument}_probe{i+1}_vect_dict_1Hz_day{day_number}.csv", "w"))
w.writerow(["key","dI","dB_X","dB_X_err","dB_Y","dB_Y_err","dB_Z","dB_Z_err"])
for key, val in dBdI.items():
w.writerow([key,val[0],val[1],val[2],val[3],val[4],val[5],val[6]])#,val[9],val[10],val[11]])
if rand_noise:
vect_dict[f'{i+1}'] = [params_x[0], params_y[0], params_z[0], perr_x[0], perr_y[0], perr_z[0], params_x[1], params_y[1], params_z[1], perr_x[1], perr_y[1], perr_z[1]]
else:
vect_dict[f'{i+1}'] = [X.slope, Y.slope, Z.slope,X.stderr,Y.stderr,Z.stderr, X.intercept ,Y.intercept, Z.intercept]
##,params_x[0],params_y[0],params_z[0],perr_x[0],perr_y[0],perr_z[0]] #atm linear regression gradient - or should it be curve_fit?
return vect_dict
def day_two(windows,
probe_num_list,
start_dt,
end_dt,
*,
alt=False,
sampling_freq=None,
plot=True,
spectrogram=False,
powerspec=False,
inst_name=None):
"""day_two(windows, probe_num_list, start_dt, end_dt, *, alt = False, sampling_freq = None, plot = True, spectrogram = False, powerspec = False, inst_name=None) --> used to analyse day 2 MFSA data
Parameters:
windows: first arg. Determines if using Windows or Mac path
probe_num_list: second arg. List of probe numbers as integers that are desired
start_dt: third arg. Start datetime for timespan interested in
end_dt: fourth arg. End datetime for timespan interested in
Keyword Args:
alt: fifth arg. If True uses brute force fft method rather than inbuilt scipy periodogram to compute power spectra
sampling_freq: sixth arg. Determines the desired sampling freqeuncy. If None defaults to 1kHz native.
plot: seventh arg. If true plots timeseries of the desired data
spectrogram: eigth arg. If true plots spectrogram of the desired data
powerspec: ninth arg. If true plots powerspectrum, x,y,z and trace of desired data
inst_name: tenth arg. Provides the name of the instrument to go in powerspec suptitle
"""
if windows:
path_fol_A = r'C:\Users\jonas\MSci-Data\day_two\A'
path_fol_B = r'C:\Users\jonas\MSci-Data\day_two\B'
else:
path_fol_A = os.path.expanduser(
"~/Documents/MsciProject/Data/day_two/A")
path_fol_B = os.path.expanduser(
"~/Documents/MsciProject/Data/day_two/B")
#here select which probe is desired, only one at a time
#num = 12
b_noise = []
for num in probe_num_list:
print('num = ', num)
if num < 9:
soloA_bool = True
else:
soloA_bool = False
if num < 10:
num_str = f'0{num}'
else:
num_str = num
collist = [
'time', f'Probe{num_str}_X', f'Probe{num_str}_Y',
f'Probe{num_str}_Z'
]
#finding the correct MFSA data files
start_csv, end_csv = processing.which_csvs(
soloA_bool, 2, start_dt, end_dt, tz_MAG=False
) #this function (in processing.py) finds the number at the end of the csv files we want
print(start_csv, end_csv)
all_files = [0] * (end_csv + 1 - start_csv)
for index, i in enumerate(
range(start_csv, end_csv + 1)
): #this will loop through and add the csv files that contain the start and end time set above
if soloA_bool:
if windows:
all_files[
index] = path_fol_A + f'\SoloA_2019-06-24--08-14-46_{i}.csv'
else:
all_files[index] = path_fol_A + os.path.expanduser(
f'/SoloA_2019-06-24--08-14-46_{i}.csv'
) #need to change path_fol_A to the path where your A folder is
else:
if windows:
all_files[
index] = path_fol_B + f'\SoloB_2019-06-24--08-14-24_{i}.csv'
else:
all_files[index] = path_fol_B + os.path.expanduser(
f'/SoloB_2019-06-24--08-14-24_{i}.csv'
) #need to change path_fol_B to the path where your B folder is
#set this to the directory where the data is kept on your local computer
if soloA_bool:
df = processing.read_files(all_files,
soloA_bool,
sampling_freq,
collist,
day=2,
start_dt=start_dt,
end_dt=end_dt)
rotate_mat = processing.rotate_24(soloA_bool)[num - 1]
else:
df = processing.read_files(all_files,
soloA_bool,
sampling_freq,
collist,
day=2,
start_dt=start_dt,
end_dt=end_dt)
rotate_mat = processing.rotate_24(soloA_bool)[num - 9]
df.iloc[:, 0:3] = np.matmul(rotate_mat, df.iloc[:, 0:3].values.T).T
#find the df of the exact time span desired
df2 = df.between_time(start_dt.time(), end_dt.time())
dflen = len(df2)
#df3 = df2.resample('1s').mean()
#shitfing time so the axes are in spacecraft time to compare with current data
df2 = processing.shifttime(df2, soloA_bool, 2)
if plot: #plotting the raw probes results
#plt.figure()
df3 = df2.resample('1s').mean()
tmp = []
for col in collist[1:]:
df2[col] = df2[col] - df2[col].mean()
plt.plot(df2.index.time, df2[col], label=str(col))
var_1hz = np.std(df3[col])
var_1khz = np.std(df2[col])
print('std - 1Hz', col, var_1hz)
print('std - 1kHz', col, var_1khz)
tmp.append(var_1hz)
tmp.append(var_1khz)
#print(df2[col].abs().idxmax())
plt.xlabel('Time (s)')
plt.ylabel('dB (nT)')
plt.title(f'Probe {num} @ {sampling_freq}Hz, {start_dt.date()}')
plt.legend(loc="best")
plt.show()
return tmp
if spectrogram:
x = np.sqrt(df2[collist[1]]**2 + df2[collist[2]]**2 +
df2[collist[3]]**2)
fs = sampling_freq
div = dflen / 1000
#f, Pxx = sps.periodogram(x,fs)
#div = 500
nff = dflen // div
wind = sps.hamming(int(dflen // div))
f, t, Sxx = sps.spectrogram(x,
fs,
window=wind,
noverlap=int(dflen // (2 * div)),
nfft=nff) #,nperseg=700)
ax = plt.figure()
plt.pcolormesh(t, f, Sxx, vmin=0., vmax=0.03)
plt.semilogy()
plt.ylabel('Frequency [Hz]')
plt.xlabel('Time [sec]')
plt.title(
f'Spectrogram: Probe {num} @ {sampling_freq}Hz, {start_dt.date()}'
)
plt.ylim((10**0, sampling_freq / 2))
plt.clim()
fig = plt.gcf()
cbar = plt.colorbar()
#cbar.ax.set_yticklabels(fontsize=8)
cbar.set_label('Normalised Power/Frequency') #, rotation=270)
#fig, ax2 = plt.subplots()
#Pxx, freqs, bins, im = ax2.specgram(x, Fs=sampling_freq)#, noverlap=900)
#ax2.set_yscale('log')
#ax2.set_ylim((10**0,sampling_freq/2))
plt.show()
if powerspec:
if inst_name == None:
processing.powerspecplot(df2,
sampling_freq,
collist,
alt,
save=False)
else:
processing.powerspecplot(df2,
sampling_freq,
collist,
alt,
inst=inst_name,
save=True)
def day_one(windows,
probe_num_list,
start_dt,
end_dt,
alt,
sampling_freq=None,
plot=False,
spectrogram=False,
powerspec=False):
"""
all_files - set this to the directory where the data is kept on your local computer
collist - list of columns desired
soloA_bool - set to True if desire sensors in SoloA channel, else False
num - set to the number of the probe desired
start_dt - start datetime desired
end_dt - end datetime desired
alt - boolean - set to True if desire the 'brute force' method for power spectrum rather than periodogram method
sampling_freq - set to desired sampling frequency - default = None
"""
if windows:
path_fol_A = r'C:\Users\jonas\MSci-Data\day_one\A'
path_fol_B = r'C:\Users\jonas\MSci-Data\day_one\B'
else:
path_fol_A = os.path.expanduser(
"~/Documents/MSciProject/Data/day_one/A")
path_fol_B = os.path.expanduser(
"~/Documents/MSciProject/Data/day_one/B")
#alt - set to true if you want to see power spec using the stnadard method - not the inbuilt funciton
#num = 5
#b_noise = []
for num in probe_num_list:
print('num = ', num)
if num < 9:
soloA_bool = True
else:
soloA_bool = False
if num < 10:
num_str = f'0{num}'
else:
num_str = num
collist = [
'time', f'Probe{num_str}_X', f'Probe{num_str}_Y',
f'Probe{num_str}_Z'
]
day = 1
start_csv, end_csv = processing.which_csvs(
soloA_bool, day, start_dt, end_dt, tz_MAG=True
) #this function (in processing.py) finds the number at the end of the csv files we want
print(start_csv, end_csv)
all_files = [0] * (end_csv + 1 - start_csv)
for index, i in enumerate(
range(start_csv, end_csv + 1)
): #this will loop through and add the csv files that contain the start and end time set above
if soloA_bool:
if windows:
all_files[
index] = path_fol_A + f'\SoloA_2019-06-21--08-10-10_{i}.csv'
else:
all_files[index] = path_fol_A + os.path.expanduser(
f'/SoloA_2019-06-21--08-10-10_{i}.csv'
) #need to change path_fol_A to the path where your A folder is
else:
if windows:
all_files[
index] = path_fol_B + f'\SoloB_2019-06-21--08-09-10_{i}.csv'
else:
all_files[index] = path_fol_B + os.path.expanduser(
f'/SoloB_2019-06-21--08-09-10_{i}.csv'
) #need to change path_fol_B to the path where your B folder is
if soloA_bool:
df = processing.read_files(all_files,
soloA_bool,
sampling_freq,
collist,
day=1,
start_dt=start_dt,
end_dt=end_dt)
rotate_mat = processing.rotate_21(soloA_bool)[num - 1]
else:
df = processing.read_files(all_files,
soloA_bool,
sampling_freq,
collist,
day=1,
start_dt=start_dt,
end_dt=end_dt)
rotate_mat = processing.rotate_21(soloA_bool)[num - 9]
df.iloc[:, 0:3] = np.matmul(rotate_mat, df.iloc[:, 0:3].values.T).T
print(len(df))
df2 = df.between_time(start_dt.time(), end_dt.time())
dflen = len(df2)
if plot: #plotting the raw probes results
plt.figure()
tmp = []
df3 = df2.resample('1s').mean()
for col in collist[1:]:
#plt.plot(df2.index.time, df2[col], label=str(col))
#print(df2[col].abs().idxmax())
var_1hz = np.std(df3[col])
var_1khz = np.std(df2[col])
print('std - 1Hz', col, var_1hz)
print('std - 1kHz', col, var_1khz)
tmp.append(var_1hz)
tmp.append(var_1khz)
plt.xlabel('Time (s)')
plt.ylabel('B (nT)')
plt.title(f'Probe {num} @ {sampling_freq}Hz, {start_dt.date()}')
plt.legend(loc="best")
plt.show()
return tmp
if spectrogram:
x = np.sqrt(df2[collist[1]]**2 + df2[collist[2]]**2 +
df2[collist[3]]**2)
fs = sampling_freq
div = dflen / 1000
#f, Pxx = sps.periodogram(x,fs)
#div = 500
nff = dflen // div
wind = sps.hamming(int(dflen // div))
f, t, Sxx = sps.spectrogram(x,
fs,
window=wind,
noverlap=int(dflen // (2 * div)),
nfft=nff) #,nperseg=700)
ax = plt.figure()
plt.pcolormesh(t, f, Sxx, vmin=0., vmax=0.01)
plt.semilogy()
plt.ylabel('Frequency [Hz]')
plt.xlabel('Time [sec]')
plt.title(
f'Spectrogram: Probe {num} @ {sampling_freq}Hz, {start_dt.date()}'
)
plt.ylim((10**0, sampling_freq / 2))
plt.clim()
fig = plt.gcf()
cbar = plt.colorbar()
#cbar.ax.set_yticklabels(fontsize=8)
cbar.set_label('Normalised Power/Frequency') #, rotation=270)
# fig, ax2 = plt.subplots()
# Pxx, freqs, bins, im = ax2.specgram(x, Fs=sampling_freq)#, noverlap=900)
# ax2.set_yscale('log')
# ax2.set_ylim((10**0,sampling_freq/2))
plt.show()
if powerspec:
processing.powerspecplot(df2,
sampling_freq,
collist,
alt,
save=False)
def get_data(self, windows=True):
start = time.time()
if windows:
os.environ['MFSA_raw'] = 'C:\\Users\\jonas\\MSci-Data'
else:
os.environ['MFSA_raw'] = os.path.expanduser(
'~/Documents/MsciProject/Data')
mfsa_init_path = os.environ.get('MFSA_raw')
#print(mfsa_init_path)
if self.day == 1:
mfsa_init_path = os.path.join(mfsa_init_path, 'day_one')
elif self.day == 2:
mfsa_init_path = os.path.join(mfsa_init_path, 'day_two')
#print(mfsa_init_path)
mfsa_fol_A = os.path.join(mfsa_init_path, 'A')
mfsa_fol_B = os.path.join(mfsa_init_path, 'B')
#print(mfsa_fol_A)
if self.probe < 9:
soloA_bool = True
else:
soloA_bool = False
if self.probe < 10:
num_str = f'0{self.probe}'
else:
num_str = self.probe
self.collist = [
'time', f'Probe{num_str}_X', f'Probe{num_str}_Y',
f'Probe{num_str}_Z'
]
#finding the correct MFSA data files
start_csv, end_csv = processing.which_csvs(
soloA_bool, day, self.start, self.end, tz_MAG=False
) #this function (in processing.py) finds the number at the end of the csv files we want
print(start_csv, end_csv)
all_files = [0] * (end_csv + 1 - start_csv)
if self.day == 2:
for index, i in enumerate(
range(start_csv, end_csv + 1)
): #this will loop through and add the csv files that contain the start and end time set above
if soloA_bool:
all_files[index] = os.path.join(
mfsa_fol_A, f'SoloA_2019-06-24--08-14-46_{i}.csv')
else:
all_files[index] = os.path.join(
mfsa_fol_B, f'SoloB_2019-06-24--08-14-24_{i}.csv')
elif self.day == 1:
for index, i in enumerate(
range(start_csv, end_csv + 1)
): #this will loop through and add the csv files that contain the start and end time set above
if soloA_bool:
all_files[index] = os.path.join(
mfsa_fol_A, f'SoloA_2019-06-21--08-10-10_{i}.csv')
else:
all_files[index] = os.path.join(
mfsa_fol_B, f'SoloB_2019-06-21--08-09-10_{i}.csv')
if soloA_bool:
self.df = processing.read_files(all_files,
soloA_bool,
self.fs,
self.collist,
self.day,
start_dt=self.start,
end_dt=self.end)
if self.day == 2:
rotate_mat = processing.rotate_24(soloA_bool)[self.probe - 1]
elif self.day == 1:
rotate_mat = processing.rotate_21(soloA_bool)[self.probe - 1]
else:
self.df = processing.read_files(all_files,
soloA_bool,
self.fs,
self.collist,
self.day,
start_dt=self.start,
end_dt=self.end)
if self.day == 2:
rotate_mat = processing.rotate_24(soloA_bool)[self.probe - 9]
elif self.day == 1:
rotate_mat = processing.rotate_21(soloA_bool)[self.probe - 9]
self.df.iloc[:, 0:3] = np.matmul(rotate_mat,
self.df.iloc[:, 0:3].values.T).T
#find the df of the exact time span desired
self.df = self.df.between_time(self.start.time(), self.end.time())
self.df = processing.shifttime(self.df, soloA_bool, self.day)
self.dflen = len(self.df)
print(self.df.head())
print(self.df.tail())
print('Data Loaded - Execution Time: ', round(time.time() - start, 2),
'seconds')