import directory from EEG import EEG from PPG import PPG import PNG_REPORT import PNG_EBSI import PDF_ALL import PNG_MDD_REPORT import MDD_classifier import traceback import os import sys import json import time import shutil import datetime import pandas as pd import PNG_classifier as PNG_classifier import classifier as classifier # filename # device # sex # age if __name__ == "__main__": start_time = time.perf_counter() # sys.argv로 들어오는 값은 모두 str type 으로 들어옴. # 웹에서 넘겨주는 값들 report_ok = {'result': 1, 'ebsi': 0} #넘겨주는 값 null 일 경우 오류 처리 try: file = sys.argv[1] #뇌파 파일 device = sys.argv[2] #측정기기 laxtha, soso number = sys.argv[3] #넘버 date = sys.argv[4] name = sys.argv[5] #이름 sex = sys.argv[6] # 성별, M - 남자 , F - 여자 age = sys.argv[7] # 나이(만) depressed = float(sys.argv[8]) # K-DEP, 우울장애 unrest = float(sys.argv[9]) # K-ANX, 불안장애 reportLanguage = sys.argv[10] # REPORT 언어 RESULT_DIR = sys.argv[11] # 저장 PATH password = sys.argv[12] # PDF password 여부 birth = sys.argv[13] # PDF password (사용자 생년월일) eeg_only = bool(int(sys.argv[14])) # eeg레포트만 ppg_only = bool(int(sys.argv[15])) # ppg레포트만 ebsi = bool(int(sys.argv[16])) # EBSI 레포트 생성 여부 msn = sys.argv[17] # CECLAB 구분하기 위해 받음. ai_use = sys.argv[18] mdd_use = sys.argv[19] # 1: 우울증 진단 페이지 추가, 2(그 외): 기존과 동일 # MDD 페이지 사용 시 PDF 비밀번호 설정 안 함 if mdd_use == '1': password = 'N' # print("file:", file) # print("device:", device) # print("number:", number) # print("date:", date) # print("name:", name) # print("sex:", sex) # print("age:", age) # print("depressed:", depressed) # print("unrest:", unrest) # print("reportLanguage:", reportLanguage) # print("RESULT_DIR:", RESULT_DIR) # print("password:", password) # print("birth:", birth) # print("eeg_only:", eeg_only) # print("ppg_only:", ppg_only) # print("ebsi:", ebsi) # print("msn:", msn) # print("ai_use:", ai_use) except Exception as e: print(traceback.format_exc()) report_ok['result'] = -2999 report_ok = json.dumps(report_ok) current_time = str(round(datetime.datetime.utcnow().timestamp() * 1000)) # 일괄생성시 설문지는 같고 파일이 다를 경우를 대비하여 폴더이름에 시간 추가 RESULT_DIR = os.path.join(RESULT_DIR, current_time + name) directory.createFolder(RESULT_DIR) print("RESULT_DIR joined = {}".format(RESULT_DIR)) s_time = time.perf_counter() """ EEG & PPG 계산 part """ if (eeg_only==False) and (ppg_only==False): eeg = EEG(file, RESULT_DIR, age=age, device=device, reportLanguage=reportLanguage) #eeg.py 에 있는 클래스 호출 try: eeg.run_function_eeg() eeg.report['eeg_code'] = 1 # 오류 없으면 eeg_code = 1 (성공했다는 뜻) default 값은 -2005 print("EEG OK") except Exception as e: print(traceback.format_exc()) # 오류 뱉고, ppg는 괜찮을 수도 있으니 ppg_only=True 로 ppg_only=True eeg_json_val = json.dumps(eeg.report) # 웹에 던져주기 위해 json 으로 묶음 print(eeg_json_val) ppg = PPG(file, RESULT_DIR, age=age, device=device, depressed=depressed, unrest=unrest, reportLanguage=reportLanguage) # ppg.py에 있는 클래스 호출 try: ppg.run_function_ppg() ppg.report['ppg_code'] = 1 # eeg 와 마찬가지, ppg_code 의 default 값은 -2006 print("PPG_OK") except Exception as e: print(traceback.format_exc()) eeg_only=True ppg_json_val = json.dumps(ppg.report) print(ppg_json_val) elif eeg_only == True: eeg = EEG(file, RESULT_DIR, age=age, device=device, reportLanguage=reportLanguage) eeg.report['ppg_code'] = 1 # ppg_code 가 -2006일 경우 오류 코드이기 떄문에 임시로 1로 지정 후 나머지 값들은 0으로 세팅 eeg.report['ppgScore'] = 0 eeg.report['ppgOnlyScore'] = 0 eeg.report['dataQuality_ppg'] = None eeg.report['ppg_subtype'] = None try: eeg.run_function_eeg() eeg.report['eeg_code'] = 1 print("EEG OK") except Exception as e: eeg.report = None print(traceback.format_exc()) eeg_json_val = json.dumps(eeg.report) print(eeg_json_val) elif ppg_only == True: ppg = PPG(file, RESULT_DIR, age=age, device=device, depressed=depressed, unrest=unrest, reportLanguage=reportLanguage) ppg.report['eeg_code'] = 1 # eeg_code 가 -2005일 경우 오류 코드이기 떄문에 임시로 1로 지정 후 나머지 값들은 0으로 세팅 ppg.report['eegScore'] = 0 ppg.report['dataQuality_eeg'] = None ppg.report['eeg_subtype'] = None try: ppg.run_function_ppg() ppg.report['ppg_code'] = 1 except Exception as e: ppg.report = None print(traceback.format_exc()) ppg_json_val = json.dumps(ppg.report) print(ppg_json_val) report_ok = {'result': 1, 'ebsi': 0} if (eeg_only==False) and (ppg_only==False)and (ai_use=='1'): feature_4class= pd.DataFrame([{"eeg":eeg.eeg_score, "ppg_hrv_only":ppg.hrv_only_score,"theta": eeg.rel[0],"l_alpha": eeg.rel[2],"h_alpha": eeg.rel[3],"l_beta": eeg.rel[4],"m_beta": eeg.rel[5],"h_beta": eeg.rel[6],"gamma": eeg.rel[7], "com": eeg.com,"faa": eeg.faa,"apf_freq": eeg.peak_x,"apf_psd": eeg.peak_y,"tp": ppg.fd['TP'].values[0][0],"vlf": ppg.fd['VLF'].values[0][0],"lf": ppg.fd['LF'].values[0][0],"hf": ppg.fd['HF'].values[0][0], "sns": ppg.sns,"pns": 100-int(ppg.sns),"si": ppg.si,"sdnn": ppg.sdnn,"rmssd": ppg.rmssd,"age": age,"sex": directory.sex_modify_int(sex)}]) predict_result = { k: { 'mean': round(v['mean'] * 100, 1), 'lo': round(v['lo'] * 100, 1), 'hi': round(v['hi'] * 100, 1), } for k, v in classifier.predict(feature_4class).items() } PNG_classifier.make_classifier(file, RESULT_DIR, number=number, name=name, sex=sex, age=age, date=date,eeg_only=eeg_only, ppg_only=ppg_only, reportLanguage=reportLanguage, predict_result=predict_result) """ MDD(우울증) AI 진단 페이지 - mdd_use=='1' 일 때만 EBSI 리포트 앞에 한 장 추가. EEG+PPG 신호가 둘 다 있어야 예측 가능하므로 eeg_only/ppg_only 모드에서는 스킵. """ report_ok['mdd_code'] = 0 if (eeg_only == False) and (ppg_only == False) and (mdd_use == '1'): try: eeg_ch1, eeg_ch2 = eeg.raw.get_data() ppg_raw = ppg.raw.values dep_prob = MDD_classifier.predict(eeg_ch1, eeg_ch2, ppg_raw, fs_eeg=eeg.samplefreq, fs_ppg=ppg.samplefreq) if dep_prob is not None: PNG_MDD_REPORT.make_mdd_report(dep_prob, RESULT_DIR, number=number, name=name, sex=directory.sex_modify(sex, reportLanguage=reportLanguage), age=age, date=date, reportLanguage=reportLanguage) report_ok['mdd_code'] = 1 report_ok['mdd_probability'] = round(dep_prob, 4) else: print("MDD: 잡음이 많아 분석 가능한 구간이 없습니다") except Exception as e: print(traceback.format_exc()) """ REPORT 생성 part """ try: PNG_REPORT.make_report(file, RESULT_DIR, depressed, unrest, number, name, directory.sex_modify(sex, reportLanguage=reportLanguage), age, date, eeg_only=eeg_only, ppg_only=ppg_only, reportLanguage=reportLanguage, msn=msn) PNG_EBSI.make_ebsi(file, RESULT_DIR, age=age, sex=directory.sex_modify(sex, reportLanguage=reportLanguage), number=number, name=name, date=date, eeg_only=eeg_only, ppg_only=ppg_only, reportLanguage=reportLanguage, msn=msn) except: print(traceback.format_exc()) report_ok['result'] = -2009 try: PDF_ALL.make_pdf(RESULT_DIR, number+"_"+name, password=password, birth=birth, eeg_only=eeg_only, ppg_only=ppg_only, msn=msn, eeg_subtype = eeg.report['eeg_subtype'], ppg_subtype = ppg.report['ppg_subtype'], reportLanguage=reportLanguage) except: print(traceback.format_exc()) report_ok['result'] = -2009 try: os.remove(os.path.join(RESULT_DIR, os.path.join(RESULT_DIR, file.split('/')[-1][:-4] + '_eeg.csv'))) os.remove(os.path.join(RESULT_DIR, os.path.join(RESULT_DIR, file.split('/')[-1][:-4] + '_ppg.csv'))) except FileNotFoundError: pass report_ok['date'] = sys.argv[4] report_ok['number'] = sys.argv[3] report_ok['name'] = sys.argv[5] report_ok['sex'] = sys.argv[6] report_ok['age'] = sys.argv[7] #result_data에 성별 평균, 연령대 평균, 전체평균 추가 25.07.30 info = pd.read_csv(os.path.join(directory.ROOT_DIR, 'all_info.csv')) info.index = ['eeg_mean', 'eeg_std', 'ppg_mean', 'ppg_std'] sex_api = directory.sex_modify(sex, reportLanguage= "kr")#all_info.csv에 성별 구분 남/여 로 돼있기 때문에 성별데이터 남/여로 변환 25.07.30 age_api = directory.age_modify(age)#all_info.csv에 연령대별로 데이터 구분되어 있기 때문에 10, 20, 30 등 연령대로 변환 report_ok['sex_eeg_mean'] = round(info[sex_api]['eeg_mean'],2) report_ok['sex_ppg_mean'] = round(info[sex_api]['ppg_mean'],2) report_ok['age_eeg_mean'] = round(info[age_api]['eeg_mean'],2) report_ok['age_ppg_mean'] = round(info[age_api]['ppg_mean'],2) report_ok['all_eeg_mean'] = round(info['all']['eeg_mean'],2) report_ok['all_ppg_mean'] = round(info['all']['ppg_mean'],2) report_ok = json.dumps(report_ok) print(report_ok) shutil.rmtree(RESULT_DIR) end_time = time.perf_counter() #print("eeg_only : {}".format(eeg_only)) #print("ppg_only : {}".format(ppg_only)) #print(f"python time elapsed : {int(round((end_time - start_time) * 1000))}ms")