123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230 |
- # synchronizer : converts Pulseq (.seq) files into sequences of amplitude, time and synchro sets.
- # Output is given by MRI machine
- # Babich Nikita, Kozin Roman, Karsakov Grigory
- # March 2024
- import numpy as np
- from matplotlib import pyplot as plt
- # from pulseq_fixed import sequence_fixed as puls_fix
- import pypulseq as puls_fix
- def output_seq(dict):
- """
- The interpretation from pypulseq format of sequence to the files needed to analog part of MRI
- Интерпретация последовательности из формата pypulseq в файлы, необходимые для аналоговой части МРТ.
- :param dict: Dictionary of the impulse sequence pypulseq provided
- Словарь импульсных последовательностей, предоставленный pypulseq
- :return: files in "data_output_seq/" directory of every type of amplitudes and time points
- Запись в файлы в директории "data_output_seq/" амплитуд и времени для всех типов
- """
- loc_t_adc = dict['t_adc']
- loc_t_rf = dict['t_rf']
- loc_t_rf_centers = dict['t_rf_centers']
- loc_t_gx = dict['t_gx']
- loc_t_gy = dict['t_gy']
- loc_t_gz = dict['t_gz']
- loc_adc = dict['adc']
- loc_rf = dict['rf']
- loc_rf_centers = dict['rf_centers']
- loc_gx = dict['gx']
- loc_gy = dict['gy']
- loc_gz = dict['gz']
- with open('data_output_seq/t_adc.txt', 'w') as f:
- data = str(tuple(loc_t_adc))
- f.write(data)
- with open('data_output_seq/t_rf.txt', 'w') as f:
- data = str(tuple(loc_t_rf))
- f.write(data)
- with open('data_output_seq/t_rf_centers.txt', 'w') as f:
- data = str(tuple(loc_t_rf_centers))
- f.write(data)
- with open('data_output_seq/t_gx.txt', 'w') as f:
- data = str(tuple(loc_t_gx))
- f.write(data)
- with open('data_output_seq/t_gy.txt', 'w') as f:
- data = str(tuple(loc_t_gy))
- f.write(data)
- with open('data_output_seq/t_gz.txt', 'w') as f:
- data = str(tuple(loc_t_gz))
- f.write(data)
- with open('data_output_seq/adc.txt', 'w') as f:
- data = str(tuple(loc_adc))
- f.write(data)
- with open('data_output_seq/rf.txt', 'w') as f:
- data = str(tuple(loc_rf))
- f.write(data)
- with open('data_output_seq/rf_centers.txt', 'w') as f:
- data = str(tuple(loc_rf_centers))
- f.write(data)
- with open('data_output_seq/gx.txt', 'w') as f:
- data = str(tuple(loc_gx))
- f.write(data)
- with open('data_output_seq/gy.txt', 'w') as f:
- data = str(tuple(loc_gy))
- f.write(data)
- with open('data_output_seq/gz.txt', 'w') as f:
- data = str(tuple(loc_gz))
- f.write(data)
- def adc_correction():
- """
- Helper function that rise times for correction of ADC events
- Вспомогательная функция получения времён для коррекции АЦП событий
- :return: rise_time: float, stores in pulseq, related to exact type of gradient events
- хранится в pulseq, связан с конкретным типом градиентного события
- fall_time: float, same as rise_time
- аналогично rise_time
- """
- rise_time, fall_time = None, None
- is_adc_inside = False
- for j in range(blocks_number - 1):
- iterable_block = seq_input.get_block(block_index=j + 1)
- if iterable_block.adc is not None:
- is_adc_inside = True
- rise_time = iterable_block.gx.rise_time
- fall_time = iterable_block.gx.fall_time
- if not is_adc_inside:
- raise Exception("No ADC event found inside sequence")
- return rise_time, fall_time
- def adc_event_edges(local_gate_adc):
- """
- Helper function that rise numbers of blocks of border correction of ADC events
- Вспомогательная функция для получения номеров блоков границ коррекции АЦП событий
- :return: num_begin_l: int, number of time block when adc event starts
- номер временного блока начала АЦП события
- num_finish_l: int, same but ends
- то же, но для окончания
- """
- num_begin_l = 0
- flag_begin = False
- flag_finish = False
- num_finish_l = 1
- for k in range(len(local_gate_adc) - 1):
- if local_gate_adc[k] != 0 and not flag_begin:
- num_begin_l = k
- flag_begin = True
- if local_gate_adc[k] != 0 and local_gate_adc[k + 1] == 0 and not flag_finish:
- num_finish_l = k
- flag_finish = True
- return num_begin_l, num_finish_l
- def synchronization(N_samples):
- ### MAIN LOOP ###
- ### ОСНОВНОЙ ЦИКЛ###
- for i in range(N_samples):
- # delaying of RF event for time period of local delay
- # задержка RF события на период времени локальной задержки
- if RF_assintant[0] - RF_raster < time_sample[i] < RF_assintant[0] + RF_raster:
- RF_stop = int(RF_assintant[1] / time_step)
- gate_rf[i:RF_stop] = 1.0
- var = 1
- # mandatory disabling of RF gate due to ADC work same time
- # принудительное отключение RF-шлюза из-за одновременной работы АЦП
- gate_rf_2 = map(lambda x: time_sample[i] - ADC_raster < x < time_sample[i] + ADC_raster and 1 or 0,
- seq_output_dict['t_adc'])
- if np.any(np.array(list(gate_rf_2)) > 0):
- gate_rf[i] = 0.0
- # TR switch with own delay before ADC turning
- # TR перключение с собственной задержкой перед включением АЦП
- gate_tr_1 = map(lambda x: time_sample[i] - ADC_raster < x < time_sample[i] + ADC_raster and 1 or 0,
- seq_output_dict['t_adc'])
- if np.any(np.array(list(gate_tr_1)) > 0):
- block_delay_tr = int(local_delay_tr / time_step)
- gate_tr_switch[i - block_delay_tr:i + 1] = 0.0
- # first step of ADC gate - enabling
- # первый шак АЦП шлюза - включение
- gate_adc_1 = map(lambda x: time_sample[i] - ADC_raster < x < time_sample[i] + ADC_raster and 1 or 0,
- seq_output_dict['t_adc'])
- if np.any(np.array(list(gate_adc_1)) > 0):
- gate_adc[i] = 1.0
- # adc correction sue to rise and fall time of gradient
- # АЦП коррекция в зависимости от времени нарастания или спада градиента
- # defining time that ADC need to be disabled during of
- # определение премени, когда АЦП необходимо отключить
- rise_time_loc, fall_time_loc = adc_correction()
- num_beg, num_fin = adc_event_edges(gate_adc)
- rise_time_tick = int(rise_time_loc / time_step)
- fall_time_tick = int(rise_time_loc / time_step)
- gate_adc[num_beg:num_beg + rise_time_tick] = 0.0
- gate_adc[num_fin - fall_time_tick:num_fin + 1] = 0.0
- gates_release = {"adc": gate_adc,
- "rf": gate_rf,
- "tr_switch": gate_tr_switch,
- "gx": gate_gx,
- "gy": gate_gy,
- "gz": gate_gz}
- # gates_output(gates_release)
- if __name__ == '__main__':
- print('')
- seq_file = "seq_store/SE_rfdeath_5000.seq"
- seq_input = puls_fix.Sequence()
- seq_input.read(file_path=seq_file)
- seq_output_dict = seq_input.waveforms_export(time_range=(0, 3))
- # artificial delays due to construction of the MRI
- # искусственные задержки из-за тех. особенностей МРТ
- RF_dtime = 100 * 1e-6
- TR_dtime = 100 * 1e-6
- time_info = seq_input.duration()
- blocks_number = time_info[1]
- time_dur = time_info[0]
- time_step = 20 * 1e-9
- N_samples = int(time_dur / time_step)
- time_sample = np.linspace(0, time_dur, N_samples)
- # output interpretation. all formats of files defined in method
- # интерпретация выхода. Все форматы файлов определены в методе
- output_seq(seq_output_dict)
- # defining constants of the sequence
- # определение констант последовательности
- local_definitions = seq_input.definitions
- ADC_raster = local_definitions['AdcRasterTime']
- RF_raster = local_definitions['RadiofrequencyRasterTime']
- gate_adc = np.zeros(N_samples)
- gate_rf = np.zeros(N_samples)
- gate_tr_switch = np.ones(N_samples)
- gate_gx = np.zeros(N_samples)
- gate_gy = np.zeros(N_samples)
- gate_gz = np.zeros(N_samples)
- local_delay_rf = RF_dtime
- local_delay_tr = TR_dtime
- local_raster_time = time_step
- RF_assintant = [seq_output_dict['t_rf'][0] - RF_dtime, seq_output_dict['t_rf'][-1]]
- synchronization(N_samples)
- # testing plots for synchronization
- # графики тестов синхрозации
- plt.plot(seq_output_dict['t_gx'][:int(N_samples)], seq_output_dict['gx'][:int(N_samples)])
- plt.plot(seq_output_dict['t_gy'][:int(N_samples)], seq_output_dict['gy'][:int(N_samples)])
- plt.plot(seq_output_dict['t_gz'][:int(N_samples)], seq_output_dict['gz'][:int(N_samples)])
- plt.savefig("plots_output/gradients.png")
- plt.show()
- plt.plot(seq_output_dict['t_gx'][:int(N_samples)], seq_output_dict['gx'][:int(N_samples)] / 720)
- plt.plot(time_sample[:int(N_samples)], gate_adc[:int(N_samples)], label='ADC gate')
- plt.plot(time_sample[:int(N_samples)], gate_tr_switch[:int(N_samples)], label='TR switch')
- plt.plot(seq_output_dict['t_rf'], seq_output_dict['rf'] / 210, label='RF signal')
- plt.plot(time_sample[:int(N_samples)], gate_rf[:int(N_samples)], label='RF gate')
- plt.legend()
- plt.savefig("plots_output/synchro_pulse.png")
- plt.show()
|