Unlock 30% off on Manual Testing Annual Plans this Holiday Season.

Copied to Clipboard!

How to use triage method in avocado

Best Python code snippet using avocado_python

helpers.py

Source:helpers.py

1import pandas as pd2import numpy as np3import re4import os5import math6import matplotlib.pyplot as plt7from datetime import timedelta8from sklearn import metrics9from sklearn.metrics import precision_recall_curve, average_precision_score10from sklearn.preprocessing import LabelEncoder11from tensorflow.keras.utils import Sequence12from tensorflow.keras.preprocessing.sequence import pad_sequences13import collections14def convert_str_to_float(x):15    if isinstance(x, str):16        x_split = re.compile('[^a-zA-Z0-9-]').split(x.strip())17        if '-' in x_split[0]:18            x_split_dash = x_split[0].split('-')19            if len(x_split_dash) == 2 and x_split_dash[0].isnumeric() and x_split_dash[1].isnumeric():20                return (float(x_split_dash[0]) + float(x_split_dash[1])) / 221            else:22                return np.nan23        else:24            if x_split[0].isnumeric():25                return float(x_split[0])26            else:27                return np.nan28    else:29        return x30        31def read_edstays_table(edstays_table_path):32    df_edstays = pd.read_csv(edstays_table_path)33    df_edstays['intime'] = pd.to_datetime(df_edstays['intime'])34    df_edstays['outtime'] = pd.to_datetime(df_edstays['outtime'])35    return df_edstays36def read_patients_table(patients_table_path):37    df_patients = pd.read_csv(patients_table_path)38    df_patients['dod'] = pd.to_datetime(df_patients['dod'])39    return df_patients40def read_admissions_table(admissions_table_path):41    df_admissions = pd.read_csv(admissions_table_path)42    df_admissions = df_admissions.rename(columns={"race": "ethnicity"})43    df_admissions =  df_admissions[['subject_id', 'hadm_id', 'admittime', 'dischtime', 'deathtime','ethnicity', 'edregtime','edouttime', 'insurance']]44    df_admissions['admittime'] = pd.to_datetime(df_admissions['admittime'])45    df_admissions['dischtime'] = pd.to_datetime(df_admissions['dischtime'])46    df_admissions['deathtime'] = pd.to_datetime(df_admissions['deathtime'])47    return df_admissions48def read_icustays_table(icustays_table_path):49    df_icu = pd.read_csv(icustays_table_path)50    df_icu['intime'] = pd.to_datetime(df_icu['intime'])51    df_icu['outtime'] = pd.to_datetime(df_icu['outtime'])52    return df_icu53def read_triage_table(triage_table_path):54    df_triage = pd.read_csv(triage_table_path)55    vital_rename_dict = {vital: '_'.join(['triage', vital]) for vital in ['temperature', 'heartrate', 'resprate', 'o2sat', 'sbp', 'dbp', 'pain', 'acuity']}56    df_triage.rename(vital_rename_dict, axis=1, inplace=True)57    df_triage['triage_pain'] = df_triage['triage_pain'].apply(convert_str_to_float).astype(float)58    return df_triage59def read_diagnoses_table(diagnoses_table_path):60    df_diagnoses = pd.read_csv(diagnoses_table_path)61    return df_diagnoses62def read_vitalsign_table(vitalsign_table_path):63    df_vitalsign = pd.read_csv(vitalsign_table_path)64    vital_rename_dict = {vital: '_'.join(['ed', vital]) for vital in65                         ['temperature', 'heartrate', 'resprate', 'o2sat', 'sbp', 'dbp', 'rhythm', 'pain']}66    df_vitalsign.rename(vital_rename_dict, axis=1, inplace=True)67    df_vitalsign['ed_pain'] = df_vitalsign['ed_pain'].apply(convert_str_to_float).astype(float)68    return df_vitalsign69def read_pyxis_table(pyxis_table_path):70    df_pyxis = pd.read_csv(pyxis_table_path)71    return df_pyxis72def merge_edstays_patients_on_subject(df_edstays,df_patients):73    if 'gender' in df_edstays.columns:74        df_edstays = pd.merge(df_edstays, df_patients[['subject_id', 'anchor_age', 'anchor_year','dod']], on = ['subject_id'], how='left')75    else:76        df_edstays = pd.merge(df_edstays, df_patients[['subject_id', 'anchor_age', 'gender', 'anchor_year','dod']], on = ['subject_id'], how='left')77    return df_edstays78def merge_edstays_admissions_on_subject(df_edstays ,df_admissions):79    df_edstays = pd.merge(df_edstays,df_admissions, on = ['subject_id', 'hadm_id'], how='left')80    return df_edstays81def merge_edstays_triage_on_subject(df_master ,df_triage):82    df_master = pd.merge(df_master,df_triage, on = ['subject_id', 'stay_id'], how='left')83    return df_master84def add_age(df_master):85    df_master['in_year'] = df_master['intime'].dt.year86    df_master['age'] = df_master['in_year'] - df_master['anchor_year'] + df_master['anchor_age']87    #df_master.drop(['anchor_age', 'anchor_year', 'in_year'],axis=1, inplace=True)88    return df_master89def add_inhospital_mortality(df_master):90    inhospital_mortality = df_master['dod'].notnull() & (df_master['dischtime'] >= df_master['dod'])91    df_master['outcome_inhospital_mortality'] = inhospital_mortality92    return df_master93def add_ed_los(df_master):94    ed_los = df_master['outtime'] - df_master['intime']95    df_master['ed_los'] = ed_los96    return df_master97def add_outcome_icu_transfer(df_master, df_icustays, timerange):98    timerange_delta = timedelta(hours = timerange)99    df_icustays_sorted = df_icustays[['subject_id', 'hadm_id', 'intime']].sort_values('intime')100    df_icustays_keep_first = df_icustays_sorted.groupby('hadm_id').first().reset_index()101    df_master_icu = pd.merge(df_master, df_icustays_keep_first, on = ['subject_id', 'hadm_id'], how='left', suffixes=('','_icu'))102    time_diff = (df_master_icu['intime_icu']- df_master_icu['outtime'])103    df_master_icu['time_to_icu_transfer'] = time_diff104    df_master_icu[''.join(['outcome_icu_transfer_', str(timerange), 'h'])] = time_diff <= timerange_delta105    # df_master_icu.drop(['intime_icu', 'time_to_icu_transfer'],axis=1, inplace=True)106    return df_master_icu107def fill_na_ethnicity(df_master): # requires df_master to be sorted 108    N = len(df_master)109    ethnicity_list= [float("NaN") for _ in range(N)]110    ethnicity_dict = {} # dict to store subejct ethnicity111    def get_filled_ethnicity(row):112        i = row.name113        if i % 10000 == 0:114            print('Process: %d/%d' % (i, N), end='\r')115        curr_eth = row['ethnicity']116        curr_subject = row['subject_id']117        prev_subject = df_master['subject_id'][i+1] if i< (N-1) else None118        if curr_subject not in ethnicity_dict.keys(): ## if subject ethnicity not stored yet, look ahead and behind 119            subject_ethnicity_list = []120            next_subject_idx = i+1121            prev_subject_idx = i-1122            next_subject= df_master['subject_id'][next_subject_idx] if next_subject_idx <= (N-1) else None123            prev_subject= df_master['subject_id'][prev_subject_idx] if prev_subject_idx >= 0 else None124            subject_ethnicity_list.append(df_master['ethnicity'][i]) ## add current ethnicity to list125            while prev_subject == curr_subject:126                subject_ethnicity_list.append(df_master['ethnicity'][prev_subject_idx])127                prev_subject_idx -= 1128                prev_subject= df_master['subject_id'][prev_subject_idx] if prev_subject_idx >= 0 else None129            while next_subject == curr_subject:130                subject_ethnicity_list.append(df_master['ethnicity'][next_subject_idx])131                next_subject_idx += 1132                next_subject= df_master['subject_id'][next_subject_idx] if next_subject_idx <= (N-1) else None133        134            eth_counter_list = collections.Counter(subject_ethnicity_list).most_common() #sorts counter and outputs list135            136            if len(eth_counter_list) == 0: ## no previous or next entries 137                subject_eth = curr_eth138            elif len(eth_counter_list) == 1: ## exactly one other ethnicity139                subject_eth = eth_counter_list.pop(0)[0] ## extract ethnicity from count tuple140            else:141                eth_counter_list = [x for x in eth_counter_list if pd.notna(x[0])] # remove any NA142                subject_eth = eth_counter_list.pop(0)[0]143            144            ethnicity_dict[curr_subject] = subject_eth ## store in dict145    146        if pd.isna(curr_eth): ## if curr_eth is na, fill with subject_eth from dict147            ethnicity_list[i]= ethnicity_dict[curr_subject]148        else:149            ethnicity_list[i]= curr_eth150            151    df_master.apply(get_filled_ethnicity, axis=1)152    print('Process: %d/%d' % (N, N), end='\r')153    df_master.loc[:,'ethnicity'] = ethnicity_list154    return df_master155def generate_past_ed_visits(df_master, timerange):156    #df_master = df_master.sort_values(['subject_id', 'intime']).reset_index()157    158    timerange_delta = timedelta(days=timerange)159    N = len(df_master)160    n_ed = [0 for _ in range(N)]161    def get_num_past_ed_visits(df):162        start = df.index[0]163        for i in df.index:164            if i % 10000 == 0:165                print('Process: %d/%d' % (i, N), end='\r')166            while df.loc[i, 'intime'] - df.loc[start, 'intime'] > timerange_delta:167                start += 1168            n_ed[i] = i - start169    grouped = df_master.groupby('subject_id')170    grouped.apply(get_num_past_ed_visits)171    print('Process: %d/%d' % (N, N), end='\r')172    df_master.loc[:, ''.join(['n_ed_', str(timerange), "d"])] = n_ed173    return df_master174def generate_past_admissions(df_master, df_admissions, timerange):175    df_admissions_sorted = df_admissions[df_admissions['subject_id'].isin(df_master['subject_id'].unique().tolist())][['subject_id', 'admittime']].copy()176    177    df_admissions_sorted.loc[:,'admittime'] = pd.to_datetime(df_admissions_sorted['admittime'])178    df_admissions_sorted.sort_values(['subject_id', 'admittime'], inplace=True)179    df_admissions_sorted.reset_index(drop=True, inplace=True)180    timerange_delta = timedelta(days=timerange)181    N = len(df_master)182    n_adm = [0 for _ in range(N)]183    def get_num_past_admissions(df):184        subject_id = df.iloc[0]['subject_id']185        if subject_id in grouped_adm.groups.keys():186            df_adm = grouped_adm.get_group(subject_id)187            start = end = df_adm.index[0]188            for i in df.index:189                if i % 10000 == 0:190                    print('Process: %d/%d' % (i, N), end='\r')191                while start < df_adm.index[-1] and df.loc[i, 'intime'] - df_adm.loc[start, 'admittime'] > timerange_delta:192                    start += 1193                end = start194                while end <= df_adm.index[-1] and \195                        (timerange_delta >= (df.loc[i, 'intime'] - df_adm.loc[end, 'admittime']) > timedelta(days=0)):196                    end += 1197                n_adm[i] = end - start198    grouped = df_master.groupby('subject_id')199    grouped_adm = df_admissions_sorted.groupby('subject_id')200    grouped.apply(get_num_past_admissions)201    print('Process: %d/%d' % (N, N), end='\r')202    df_master.loc[:,''.join(['n_hosp_', str(timerange), "d"])] = n_adm203    return df_master204def generate_past_icu_visits(df_master, df_icustays, timerange):205    df_icustays_sorted = df_icustays[df_icustays['subject_id'].isin(df_master['subject_id'].unique().tolist())][['subject_id', 'intime']].copy()206    df_icustays_sorted.sort_values(['subject_id', 'intime'], inplace=True)207    df_icustays_sorted.reset_index(drop=True, inplace=True)208    timerange_delta = timedelta(days=timerange)209    N = len(df_master)210    n_icu = [0 for _ in range(N)]211    def get_num_past_icu_visits(df):212        subject_id = df.iloc[0]['subject_id']213        if subject_id in grouped_icu.groups.keys():214            df_icu = grouped_icu.get_group(subject_id)215            start = end = df_icu.index[0]216            for i in df.index:217                if i % 10000 == 0:218                    print('Process: %d/%d' % (i, N), end='\r')219                while start < df_icu.index[-1] and df.loc[i, 'intime'] - df_icu.loc[start, 'intime'] > timerange_delta:220                    start += 1221                end = start222                while end <= df_icu.index[-1] and \223                        (timerange_delta >= (df.loc[i, 'intime'] - df_icu.loc[end, 'intime']) > timedelta(days=0)):224                    end += 1225                n_icu[i] = end - start226    grouped = df_master.groupby('subject_id')227    grouped_icu = df_icustays_sorted.groupby('subject_id')228    grouped.apply(get_num_past_icu_visits)229    print('Process: %d/%d' % (N, N), end='\r')230    df_master.loc[:,''.join(['n_icu_', str(timerange), "d"])] = n_icu231    return df_master232def generate_future_ed_visits(df_master, next_ed_visit_timerange):233    N = len(df_master)234    time_of_next_ed_visit = [float("NaN") for _ in range(N)]235    time_to_next_ed_visit = [float("NaN") for _ in range(N)]236    outcome_ed_revisit = [False for _ in range(N)]237    timerange_delta = timedelta(days = next_ed_visit_timerange)238    curr_subject=None239    next_subject=None240    def get_future_ed_visits(row):241        i = row.name242        if i % 10000 == 0:243            print('Process: %d/%d' % (i, N), end='\r')244        curr_subject = row['subject_id']245        next_subject= df_master['subject_id'][i+1] if i< (N-1) else None246        if curr_subject == next_subject:247            curr_outtime = row['outtime']248            next_intime = df_master['intime'][i+1]249            next_intime_diff = next_intime - curr_outtime250            time_of_next_ed_visit[i] = next_intime251            time_to_next_ed_visit[i] = next_intime_diff252            outcome_ed_revisit[i] = next_intime_diff < timerange_delta253    df_master.apply(get_future_ed_visits, axis=1)254    print('Process: %d/%d' % (N, N), end='\r')255    df_master.loc[:,'next_ed_visit_time'] = time_of_next_ed_visit256    df_master.loc[:,'next_ed_visit_time_diff'] = time_to_next_ed_visit257    df_master.loc[:,''.join(['outcome_ed_revisit_', str(next_ed_visit_timerange), "d"])] = outcome_ed_revisit258    return df_master259def generate_numeric_timedelta(df_master):260    N = len(df_master)261    ed_los_hours = [float("NaN") for _ in range(N)]262    time_to_icu_transfer_hours = [float("NaN") for _ in range(N)]263    next_ed_visit_time_diff_days = [float("NaN") for _ in range(N)]264    265    def get_numeric_timedelta(row):266        i = row.name267        if i % 10000 == 0:268            print('Process: %d/%d' % (i, N), end='\r')269        curr_subject = row['subject_id']270        curr_ed_los = row['ed_los']271        curr_time_to_icu_transfer = row['time_to_icu_transfer']272        curr_next_ed_visit_time_diff = row['next_ed_visit_time_diff']273        274        ed_los_hours[i] = round(curr_ed_los.total_seconds() / (60*60),2) if not pd.isna(curr_ed_los) else curr_ed_los275        time_to_icu_transfer_hours[i] = round(curr_time_to_icu_transfer.total_seconds() / (60*60),2) if not pd.isna(curr_time_to_icu_transfer) else curr_time_to_icu_transfer276        next_ed_visit_time_diff_days[i] = round(curr_next_ed_visit_time_diff.total_seconds() / (24*60*60), 2) if not pd.isna(curr_next_ed_visit_time_diff) else curr_next_ed_visit_time_diff277    278    df_master.apply(get_numeric_timedelta, axis=1)279    print('Process: %d/%d' % (N, N), end='\r')280    281    df_master.loc[:,'ed_los_hours'] = ed_los_hours282    df_master.loc[:,'time_to_icu_transfer_hours'] = time_to_icu_transfer_hours283    df_master.loc[:,'next_ed_visit_time_diff_days'] = next_ed_visit_time_diff_days284    return df_master285def encode_chief_complaints(df_master, complaint_dict):286    holder_list = []287    complaint_colnames_list = list(complaint_dict.keys())288    complaint_regex_list = list(complaint_dict.values())289    for i, row in df_master.iterrows():290        curr_patient_complaint = str(row['chiefcomplaint'])291        curr_patient_complaint_list = [False for _ in range(len(complaint_regex_list))]292        complaint_idx = 0293        for complaint in complaint_regex_list:294            if re.search(complaint, curr_patient_complaint, re.IGNORECASE):295                curr_patient_complaint_list[complaint_idx] = True296            complaint_idx += 1297        298        holder_list.append(curr_patient_complaint_list)299    300    df_encoded_complaint = pd.DataFrame(holder_list, columns = complaint_colnames_list)301    df_master = pd.concat([df_master,df_encoded_complaint], axis=1)302    return df_master303def merge_vitalsign_info_on_edstay(df_master, df_vitalsign, options=[]):304    df_vitalsign.sort_values('charttime', inplace=True)305    grouped = df_vitalsign.groupby(['stay_id'])306    for option in options:307        method = getattr(grouped, option, None)308        assert method is not None, "Invalid option. " \309                                   "Should be a list of values from 'max', 'min', 'median', 'mean', 'first', 'last'. " \310                                   "e.g. ['median', 'last']"311        df_vitalsign_option = method(numeric_only=True)312        df_vitalsign_option.rename({name: '_'.join([name, option]) for name in313                                    ['ed_temperature', 'ed_heartrate', 'ed_resprate', 'ed_o2sat', 'ed_sbp', 'ed_dbp', 'ed_pain']},314                                   axis=1,315                                   inplace=True)316        df_master = pd.merge(df_master, df_vitalsign_option, on=['subject_id', 'stay_id'], how='left')317    return df_master318def merge_med_count_on_edstay(df_master, df_pyxis):319    df_pyxis_fillna = df_pyxis.copy()320    df_pyxis_fillna['gsn'].fillna(df_pyxis['name'], inplace=True)321    grouped = df_pyxis_fillna.groupby(['stay_id'])322    df_medcount = grouped['gsn'].nunique().reset_index().rename({'gsn': 'n_med'}, axis=1)323    df_master = pd.merge(df_master, df_medcount, on='stay_id', how='left')324    df_master.fillna({'n_med': 0}, inplace=True)325    return df_master326def merge_medrecon_count_on_edstay(df_master, df_medrecon):327    df_medrecon_fillna = df_medrecon.copy()328    df_medrecon_fillna['gsn'].fillna(df_medrecon['name'])329    grouped = df_medrecon_fillna.groupby(['stay_id'])330    df_medcount = grouped['gsn'].nunique().reset_index().rename({'gsn': 'n_medrecon'}, axis=1)331    df_master = pd.merge(df_master, df_medcount, on='stay_id', how='left')332    df_master.fillna({'n_medrecon': 0}, inplace=True)333    return df_master334def outlier_removal_imputation(column_type, vitals_valid_range):335    column_range = vitals_valid_range[column_type]336    def outlier_removal_imputation_single_value(x):337        if x < column_range['outlier_low'] or x > column_range['outlier_high']:338            # set as missing339            return np.nan340        elif x < column_range['valid_low']:341            # impute with nearest valid value342            return column_range['valid_low']343        elif x > column_range['valid_high']:344            # impute with nearest valid value345            return column_range['valid_high']346        else:347            return x348    return outlier_removal_imputation_single_value349def convert_temp_to_celcius(df_master):350    for column in df_master.columns:351        column_type = column.split('_')[1] if len(column.split('_')) > 1 else None352        if column_type == 'temperature':353            # convert to celcius354            df_master[column] -= 32355            df_master[column] *= 5/9356    return df_master357def remove_outliers(df_master, vitals_valid_range):358    for column in df_master.columns:359        column_type = column.split('_')[1] if len(column.split('_')) > 1 else None360        if column_type in vitals_valid_range:361            df_master[column] = df_master[column].apply(outlier_removal_imputation(column_type, vitals_valid_range))362    return df_master363def display_outliers_count(df_master, vitals_valid_range):364    display_df = pd.DataFrame(columns=['variable', '< outlier_low', '[outlier_low, valid_low)',365                                       '[valid_low, valid_high]', '(valid_high, outlier_high]', '> outlier_high'])366    for column in df_master.columns:367        column_type = column.split('_')[1] if len(column.split('_')) > 1 else None368        if column_type in vitals_valid_range:369            column_range = vitals_valid_range[column_type]370            display_df = display_df.append({'variable': column,371                   '< outlier_low': len(df_master[df_master[column] < column_range['outlier_low']]),372                   '[outlier_low, valid_low)': len(df_master[(column_range['outlier_low'] <= df_master[column])373                                                             & (df_master[column] < column_range['valid_low'])]),374                   '[valid_low, valid_high]': len(df_master[(column_range['valid_low'] <= df_master[column])375                                                            & (df_master[column] <= column_range['valid_high'])]),376                   '(valid_high, outlier_high]': len(df_master[(column_range['valid_high'] < df_master[column])377                                                               & (df_master[column] <= column_range['outlier_high'])]),378                   '> outlier_high': len(df_master[df_master[column] > column_range['outlier_high']])379            }, ignore_index=True)380    return display_df381def add_score_CCI(df):382    conditions = [383        (df['age'] < 50),384        (df['age'] >= 50) & (df['age'] <= 59),385        (df['age'] >= 60) & (df['age'] <= 69),386        (df['age'] >= 70) & (df['age'] <= 79),387        (df['age'] >= 80)388    ]389    values = [0, 1, 2, 3, 4]390    df['score_CCI'] = np.select(conditions, values)    391    df['score_CCI'] = df['score_CCI'] + df['cci_MI'] + df['cci_CHF'] + df['cci_PVD'] + df['cci_Stroke'] + df['cci_Dementia'] + df['cci_Pulmonary'] + df['cci_PUD'] + df['cci_Rheumatic'] +df['cci_Liver1']*1 + df['cci_Liver2']*3 + df['cci_DM1'] + df['cci_DM2']*2 +df['cci_Paralysis']*2 + df['cci_Renal']*2 + df['cci_Cancer1']*2 + df['cci_Cancer2']*6 + df['cci_HIV']*6392    print("Variable 'add_score_CCI' successfully added")393def add_triage_MAP(df):394    df['triage_MAP'] = df['triage_sbp']*1/3 + df['triage_dbp']*2/3395    print("Variable 'add_triage_MAP' successfully added")396def add_score_REMS(df):397    conditions1 = [398        (df['age'] < 45),399        (df['age'] >= 45) & (df['age'] <= 54),400        (df['age'] >= 55) & (df['age'] <= 64),401        (df['age'] >= 65) & (df['age'] <= 74),402        (df['age'] > 74)403    ]404    values1 = [0, 2, 3, 5, 6]405    conditions2 = [406        (df['triage_MAP'] > 159),407        (df['triage_MAP'] >= 130) & (df['triage_MAP'] <= 159),408        (df['triage_MAP'] >= 110) & (df['triage_MAP'] <= 129),409        (df['triage_MAP'] >= 70) & (df['triage_MAP'] <= 109),410        (df['triage_MAP'] >= 50) & (df['triage_MAP'] <= 69),411        (df['triage_MAP'] < 49)412    ]413    values2 = [4, 3, 2, 0, 2, 4]414    conditions3 = [415        (df['triage_heartrate'] >179),416        (df['triage_heartrate'] >= 140) & (df['triage_heartrate'] <= 179),417        (df['triage_heartrate'] >= 110) & (df['triage_heartrate'] <= 139),418        (df['triage_heartrate'] >= 70) & (df['triage_heartrate'] <= 109),419        (df['triage_heartrate'] >= 55) & (df['triage_heartrate'] <= 69),420        (df['triage_heartrate'] >= 40) & (df['triage_heartrate'] <= 54),421        (df['triage_heartrate'] < 40)422    ]423    values3 = [4, 3, 2, 0, 2, 3, 4]424    conditions4 = [425        (df['triage_resprate'] > 49),426        (df['triage_resprate'] >= 35) & (df['triage_resprate'] <= 49),427        (df['triage_resprate'] >= 25) & (df['triage_resprate'] <= 34),428        (df['triage_resprate'] >= 12) & (df['triage_resprate'] <= 24),429        (df['triage_resprate'] >= 10) & (df['triage_resprate'] <= 11),430        (df['triage_resprate'] >= 6) & (df['triage_resprate'] <= 9),431        (df['triage_resprate'] < 6)432    ]433    values4 = [4, 3, 1, 0, 1, 2, 4]434    conditions5 = [435        (df['triage_o2sat'] < 75),436        (df['triage_o2sat'] >= 75) & (df['triage_o2sat'] <= 85),437        (df['triage_o2sat'] >= 86) & (df['triage_o2sat'] <= 89),438        (df['triage_o2sat'] > 89)439    ]440    values5 = [4, 3, 1, 0]441    df['score_REMS'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4) + np.select(conditions5, values5)442    print("Variable 'Score_REMS' successfully added")443    444def add_score_CART(df):445    conditions1 = [446        (df['age'] < 55),447        (df['age'] >= 55) & (df['age'] <= 69),448        (df['age'] >= 70) 449    ]450    values1 = [0, 4, 9]451    conditions2 = [452        (df['triage_resprate'] < 21),453        (df['triage_resprate'] >= 21) & (df['triage_resprate'] <= 23),454        (df['triage_resprate'] >= 24) & (df['triage_resprate'] <= 25),455        (df['triage_resprate'] >= 26) & (df['triage_resprate'] <= 29),456        (df['triage_resprate'] >= 30) 457    ]458    values2 = [0, 8, 12, 15, 22]459    conditions3 = [460        (df['triage_heartrate'] < 110),461        (df['triage_heartrate'] >= 110) & (df['triage_heartrate'] <= 139),462        (df['triage_heartrate'] >= 140) 463    ]464    values3 = [0, 4, 13]465    conditions4 = [466        (df['triage_dbp'] > 49),467        (df['triage_dbp'] >= 40) & (df['triage_dbp'] <= 49),468        (df['triage_dbp'] >= 35) & (df['triage_dbp'] <= 39),469        (df['triage_dbp'] < 35) 470    ]471    values4 = [0, 4, 6, 13]472    df['score_CART'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4)473    print("Variable 'Score_CART' successfully added")474    475def add_score_NEWS(df):476    conditions1 = [477        (df['triage_resprate'] <= 8),478        (df['triage_resprate'] >= 9) & (df['triage_resprate'] <= 11),479        (df['triage_resprate'] >= 12) & (df['triage_resprate'] <= 20),480        (df['triage_resprate'] >= 21) & (df['triage_resprate'] <= 24),481        (df['triage_resprate'] >= 25) 482    ]483    values1 = [3, 1, 0, 2, 3]484    conditions2 = [485        (df['triage_o2sat'] <= 91),486        (df['triage_o2sat'] >= 92) & (df['triage_o2sat'] <= 93),487        (df['triage_o2sat'] >= 94) & (df['triage_o2sat'] <= 95),488        (df['triage_o2sat'] >= 96) 489    ]490    values2 = [3, 2, 1, 0]491    conditions3 = [492        (df['triage_temperature'] <= 35),493        (df['triage_temperature'] > 35) & (df['triage_temperature'] <= 36),494        (df['triage_temperature'] > 36) & (df['triage_temperature'] <= 38),495        (df['triage_temperature'] > 38) & (df['triage_temperature'] <= 39),496        (df['triage_temperature'] > 39) 497    ]498    values3 = [3, 1, 0, 1, 2]499    conditions4 = [500        (df['triage_sbp'] <= 90),501        (df['triage_sbp'] >= 91) & (df['triage_sbp'] <= 100),502        (df['triage_sbp'] >= 101) & (df['triage_sbp'] <= 110),503        (df['triage_sbp'] >= 111) & (df['triage_sbp'] <= 219),504        (df['triage_sbp'] > 219) 505    ]506    values4 = [3, 2, 1, 0, 3]507    conditions5 = [508        (df['triage_heartrate'] <= 40),509        (df['triage_heartrate'] >= 41) & (df['triage_heartrate'] <= 50),510        (df['triage_heartrate'] >= 51) & (df['triage_heartrate'] <= 90),511        (df['triage_heartrate'] >= 91) & (df['triage_heartrate'] <= 110),512        (df['triage_heartrate'] >= 111) & (df['triage_heartrate'] <= 130),513        (df['triage_heartrate'] > 130) 514    ]515    values5 = [3, 1, 0, 1, 2, 3]    516    df['score_NEWS'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4) + np.select(conditions5, values5)517    print("Variable 'Score_NEWS' successfully added")518    519def add_score_NEWS2(df):   520    conditions1 = [521        (df['triage_resprate'] <= 8),522        (df['triage_resprate'] >= 9) & (df['triage_resprate'] <= 11),523        (df['triage_resprate'] >= 12) & (df['triage_resprate'] <= 20),524        (df['triage_resprate'] >= 21) & (df['triage_resprate'] <= 24),525        (df['triage_resprate'] >= 25) 526    ]527    values1 = [3, 1, 0, 2, 3]528    conditions2 = [529        (df['triage_temperature'] <= 35),530        (df['triage_temperature'] > 35) & (df['triage_temperature'] <= 36),531        (df['triage_temperature'] > 36) & (df['triage_temperature'] <= 38),532        (df['triage_temperature'] > 38) & (df['triage_temperature'] <= 39),533        (df['triage_temperature'] > 39) 534    ]535    values2 = [3, 1, 0, 1, 2]536    conditions3 = [537        (df['triage_sbp'] <= 90),538        (df['triage_sbp'] >= 91) & (df['triage_sbp'] <= 100),539        (df['triage_sbp'] >= 101) & (df['triage_sbp'] <= 110),540        (df['triage_sbp'] >= 111) & (df['triage_sbp'] <= 219),541        (df['triage_sbp'] > 219) 542    ]543    values3 = [3, 2, 1, 0, 3]544    conditions4 = [545        (df['triage_heartrate'] <= 40),546        (df['triage_heartrate'] >= 41) & (df['triage_heartrate'] <= 50),547        (df['triage_heartrate'] >= 51) & (df['triage_heartrate'] <= 90),548        (df['triage_heartrate'] >= 91) & (df['triage_heartrate'] <= 110),549        (df['triage_heartrate'] >= 111) & (df['triage_heartrate'] <= 130),550        (df['triage_heartrate'] > 130) 551    ]552    values4 = [3, 1, 0, 1, 2, 3]   553    df['score_NEWS2'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4)554    print("Variable 'Score_NEWS2' successfully added")555    556def add_score_MEWS(df):     557    conditions1 = [558        (df['triage_sbp'] <= 70),559        (df['triage_sbp'] >= 71) & (df['triage_sbp'] <= 80),560        (df['triage_sbp'] >= 81) & (df['triage_sbp'] <= 100),561        (df['triage_sbp'] >= 101) & (df['triage_sbp'] <= 199),562        (df['triage_sbp'] > 199) 563    ]564    values1 = [3, 2, 1, 0, 2]565    conditions2 = [566        (df['triage_heartrate'] <= 40),567        (df['triage_heartrate'] >= 41) & (df['triage_heartrate'] <= 50),568        (df['triage_heartrate'] >= 51) & (df['triage_heartrate'] <= 100),569        (df['triage_heartrate'] >= 101) & (df['triage_heartrate'] <= 110),570        (df['triage_heartrate'] >= 111) & (df['triage_heartrate'] <= 129),571        (df['triage_heartrate'] >= 130) 572    ]573    values2 = [2, 1, 0, 1, 2, 3]574    conditions3 = [575        (df['triage_resprate'] < 9),576        (df['triage_resprate'] >= 9) & (df['triage_resprate'] <= 14),577        (df['triage_resprate'] >= 15) & (df['triage_resprate'] <= 20),578        (df['triage_resprate'] >= 21) & (df['triage_resprate'] <= 29),579        (df['triage_resprate'] >= 30) 580    ]581    values3 = [2, 0, 1, 2, 3]582    conditions4 = [583        (df['triage_temperature'] < 35),584        (df['triage_temperature'] >= 35) & (df['triage_temperature'] < 38.5),585        (df['triage_temperature'] >= 38.5) 586    ]587    values4 = [2, 0, 2]        588    df['score_MEWS'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4) 589    print("Variable 'Score_MEWS' successfully added")590    591def add_score_SERP2d(df): 592    conditions1 = [593        (df['age'] < 30),594        (df['age'] >= 30) & (df['age'] <= 49),595        (df['age'] >= 50) & (df['age'] <= 79),596        (df['age'] >= 80)597    ]598    values1 = [0, 9, 13, 17]599    conditions2 = [600        (df['triage_heartrate'] < 60),601        (df['triage_heartrate'] >= 60) & (df['triage_heartrate'] <= 69),602        (df['triage_heartrate'] >= 70) & (df['triage_heartrate'] <= 94),603        (df['triage_heartrate'] >= 95) & (df['triage_heartrate'] <= 109),604        (df['triage_heartrate'] >= 110) 605    ]606    values2 = [3, 0, 3, 6, 10]607    conditions3 = [608        (df['triage_resprate'] < 16),609        (df['triage_resprate'] >= 16) & (df['triage_resprate'] <= 19),610        (df['triage_resprate'] >= 20) 611    ]612    values3 = [11, 0, 7]613    conditions4 = [614        (df['triage_sbp'] < 100),615        (df['triage_sbp'] >= 100) & (df['triage_sbp'] <= 114),616        (df['triage_sbp'] >= 115) & (df['triage_sbp'] <= 149),617        (df['triage_sbp'] >= 150) 618    ]619    values4 = [10, 4, 1, 0]620    conditions5 = [621        (df['triage_dbp'] < 50),622        (df['triage_dbp'] >= 50) & (df['triage_dbp'] <= 94),623        (df['triage_dbp'] >= 95) 624    ]625    values5 = [5, 0, 1]626    conditions6 = [627        (df['triage_o2sat'] < 90),628        (df['triage_o2sat'] >= 90) & (df['triage_o2sat'] <= 94),629        (df['triage_o2sat'] >= 95) 630    ]631    values6 = [7, 5, 0]632    df['score_SERP2d'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4) + np.select(conditions5, values5) + np.select(conditions6, values6)633    print("Variable 'Score_SERP2d' successfully added")634def add_score_SERP7d(df): 635    conditions1 = [636        (df['age'] < 30),637        (df['age'] >= 30) & (df['age'] <= 49),638        (df['age'] >= 50) & (df['age'] <= 79),639        (df['age'] >= 80)640    ]641    values1 = [0, 10, 17, 21]642    conditions2 = [643        (df['triage_heartrate'] < 60),644        (df['triage_heartrate'] >= 60) & (df['triage_heartrate'] <= 69),645        (df['triage_heartrate'] >= 70) & (df['triage_heartrate'] <= 94),646        (df['triage_heartrate'] >= 95) & (df['triage_heartrate'] <= 109),647        (df['triage_heartrate'] >= 110) 648    ]649    values2 = [2, 0, 4, 8, 12]650    conditions3 = [651        (df['triage_resprate'] < 16),652        (df['triage_resprate'] >= 16) & (df['triage_resprate'] <= 19),653        (df['triage_resprate'] >= 20) 654    ]655    values3 = [10, 0, 6]656    conditions4 = [657        (df['triage_sbp'] < 100),658        (df['triage_sbp'] >= 100) & (df['triage_sbp'] <= 114),659        (df['triage_sbp'] >= 115) & (df['triage_sbp'] <= 149),660        (df['triage_sbp'] >= 150) 661    ]662    values4 = [12, 6, 1, 0]663    conditions5 = [664        (df['triage_dbp'] < 50),665        (df['triage_dbp'] >= 50) & (df['triage_dbp'] <= 94),666        (df['triage_dbp'] >= 95) 667    ]668    values5 = [4, 0, 2]669    df['score_SERP7d'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4) + np.select(conditions5, values5)670    print("Variable 'Score_SERP7d' successfully added")671    672def add_score_SERP30d(df): 673    conditions1 = [674        (df['age'] < 30),675        (df['age'] >= 30) & (df['age'] <= 49),676        (df['age'] >= 50) & (df['age'] <= 79),677        (df['age'] >= 80)678    ]679    values1 = [0, 8, 14, 19]680    conditions2 = [681        (df['triage_heartrate'] < 60),682        (df['triage_heartrate'] >= 60) & (df['triage_heartrate'] <= 69),683        (df['triage_heartrate'] >= 70) & (df['triage_heartrate'] <= 94),684        (df['triage_heartrate'] >= 95) & (df['triage_heartrate'] <= 109),685        (df['triage_heartrate'] >= 110) 686    ]687    values2 = [1, 0, 2, 6, 9]688    conditions3 = [689        (df['triage_resprate'] < 16),690        (df['triage_resprate'] >= 16) & (df['triage_resprate'] <= 19),691        (df['triage_resprate'] >= 20) 692    ]693    values3 = [8, 0, 6]694    conditions4 = [695        (df['triage_sbp'] < 100),696        (df['triage_sbp'] >= 100) & (df['triage_sbp'] <= 114),697        (df['triage_sbp'] >= 115) & (df['triage_sbp'] <= 149),698        (df['triage_sbp'] >= 150) 699    ]700    values4 = [8, 5, 2, 0]701    conditions5 = [702        (df['triage_dbp'] < 50),703        (df['triage_dbp'] >= 50) & (df['triage_dbp'] <= 94),704        (df['triage_dbp'] >= 95) 705    ]706    values5 = [3, 0, 2]707    df['score_SERP30d'] = np.select(conditions1, values1) + np.select(conditions2, values2) + np.select(conditions3, values3) +                              np.select(conditions4, values4) + np.select(conditions5, values5) + df['cci_Cancer1']*6 + df['cci_Cancer2']*12708    print("Variable 'Score_SERP30d' successfully added")709    710    711def PlotROCCurve(probs,y_test_roc, ci= 95, random_seed=0):712    713    fpr, tpr, threshold = metrics.roc_curve(y_test_roc,probs)714    roc_auc = metrics.auc(fpr, tpr)715    average_precision = average_precision_score(y_test_roc, probs)716    a=np.sqrt(np.square(fpr-0)+np.square(tpr-1)).argmin()717    sensitivity = tpr[a]718    specificity = 1-fpr[a]719    threshold = threshold[a]720    print("AUC:",roc_auc)721    print("AUPRC:", average_precision)722    print("Sensitivity:",sensitivity)723    print("Specificity:",specificity)724    print("Score thresold:",threshold)725    lower_auroc, upper_auroc, std_auroc, lower_ap, upper_ap, std_ap, lower_sensitivity, upper_sensitivity, std_sensitivity, lower_specificity, upper_specificity, std_specificity = auc_with_ci(probs,y_test_roc, lower = (100-ci)/2, upper = 100-(100-ci)/2, n_bootstraps=20, rng_seed=random_seed)726    plt.title('Receiver Operating Characteristic: AUC={0:0.4f}'.format(727          roc_auc))728    plt.plot(fpr, tpr, 'b')729    plt.plot([0, 1], [0, 1],'r--')730    plt.xlim([0, 1])731    plt.ylim([0, 1])732    plt.ylabel('True Positive Rate')733    plt.xlabel('False Positive Rate')734    plt.show()735    precision, recall, threshold2 = precision_recall_curve(y_test_roc, probs)736    plt.step(recall, precision, color='b', alpha=0.2,737         where='post')738    plt.fill_between(recall, precision, step='post', alpha=0.2,739                 color='b')740    plt.xlabel('Recall')741    plt.ylabel('Precision')742    plt.ylim([0.0, 1.05])743    plt.xlim([0.0, 1.0])744    plt.title('Precision-Recall Curve: AUPRC={0:0.4f}'.format(745          average_precision))746    plt.show()747    return [roc_auc, average_precision, sensitivity, specificity, threshold, lower_auroc, upper_auroc, std_auroc, lower_ap, upper_ap, std_ap, lower_sensitivity, upper_sensitivity, std_sensitivity, lower_specificity, upper_specificity, std_specificity]748def auc_with_ci(probs,y_test_roc, lower = 2.5, upper = 97.5, n_bootstraps=200, rng_seed=10):749    print(lower, upper)750    y_test_roc = np.asarray(y_test_roc)751    bootstrapped_auroc = []752    bootstrapped_ap = []753    bootstrapped_sensitivity = []754    bootstrapped_specificity = []755    rng = np.random.default_rng(rng_seed)756    for i in range(n_bootstraps):757        # bootstrap by sampling with replacement on the prediction indices758        indices = rng.integers(0, len(y_test_roc)-1, len(y_test_roc))759        if len(np.unique(y_test_roc[indices])) < 2:760            # We need at least one positive and one negative sample for ROC AUC761            # to be defined: reject the sample762            continue763        fpr, tpr, threshold = metrics.roc_curve(y_test_roc[indices],probs[indices])764        auroc = metrics.auc(fpr, tpr)765        ap = metrics.average_precision_score(y_test_roc[indices], probs[indices])766        a=np.sqrt(np.square(fpr-0)+np.square(tpr-1)).argmin()767        sensitivity = tpr[a]768        specificity = 1-fpr[a]769        bootstrapped_auroc.append(auroc)770        bootstrapped_ap.append(ap)771        bootstrapped_sensitivity.append(sensitivity)772        bootstrapped_specificity.append(specificity)773    lower_auroc,upper_auroc = np.percentile(bootstrapped_auroc, [lower, upper])774    lower_ap,upper_ap = np.percentile(bootstrapped_ap, [lower, upper])775    lower_sensitivity,upper_sensitivity = np.percentile(bootstrapped_sensitivity, [lower, upper])776    lower_specificity,upper_specificity = np.percentile(bootstrapped_specificity, [lower, upper])777    std_auroc = np.std(bootstrapped_auroc)778    std_ap = np.std(bootstrapped_ap)779    std_sensitivity = np.std(bootstrapped_sensitivity)780    std_specificity = np.std(bootstrapped_specificity)781    return lower_auroc, upper_auroc, std_auroc, lower_ap, upper_ap, std_ap, lower_sensitivity, upper_sensitivity, std_sensitivity, lower_specificity, upper_specificity, std_specificity782def plot_confidence_interval(dataset, metric= 'auroc', ci=95, name = 'AUROC', my_file = 'AUROC_hosp.eps', my_path = 'my_path', dpi=300):783    ci_list = [dataset['lower_'+metric].values.tolist(),dataset['upper_'+metric].values.tolist()]784    std = [(dataset[metric]-dataset['std_'+metric]).values.tolist(), (dataset[metric]+dataset['std_'+metric]).values.tolist()]785    auc = dataset[metric].values.tolist()786    y = [range(len(dataset)), range(len(dataset))]787    plt.plot(ci_list,y, '-', color='gray',linewidth=1.5)788    plt.plot(std,y,'-', color='black', linewidth=2)789    plt.plot(auc,y[0],'|k', markersize=4)790    plt.xlabel(name)791    plt.yticks(range(len(dataset)),list(dataset['Model']))792    plt.savefig(os.path.join(my_path, my_file), format='eps', dpi=dpi)793    794    plt.show()795class LSTMDataGenerator(Sequence):796    def __init__(self, main_df, vitalsign_df, y, batch_size, x1_cols, x2_cols):797        self.main_df = main_df798        self.vitalsign_df = vitalsign_df799        self.batch_size = batch_size800        self.x1_cols = x1_cols801        self.x2_cols = x2_cols802        self.y_df = y803    def __len__(self):804        return math.ceil(len(self.main_df) / self.batch_size)805    def __getitem__(self, index):806        df_batch = self.main_df.iloc[index * self.batch_size:(index + 1) * self.batch_size]807        x1 = df_batch[self.x1_cols].to_numpy().astype(np.float64)808        y = self.y_df.iloc[index * self.batch_size:(index + 1) * self.batch_size].to_numpy()809        stay_ids = df_batch['stay_id'].to_numpy().astype(np.int64)810        batch_size = len(df_batch)811        df_batch = df_batch.merge(self.vitalsign_df, on='stay_id', how='left')812        x2 = []813        for i in range(batch_size):814            x2.append(df_batch[df_batch['stay_id'] == stay_ids[i]][self.x2_cols].to_numpy())815        padded_x2 = pad_sequences(x2, padding='post')816        return [x1, padded_x2.astype(np.float64)], y817def get_lstm_data_gen(df_train, df_test, df_vitalsign, variable, outcome, batch_size=200):818    variable_with_id = ["stay_id"]819    variable_with_id.extend(variable)820    X_train = df_train[variable_with_id].copy()821    y_train = df_train[outcome].copy()822    X_test = df_test[variable_with_id].copy()823    y_test = df_test[outcome].copy()824    if 'gender' in variable:825        encoder = LabelEncoder()826        X_train['gender'] = encoder.fit_transform(X_train['gender'])827        X_test['gender'] = encoder.transform(X_test['gender'])828    if 'ed_los' in variable:829        X_train['ed_los'] = pd.to_timedelta(X_train['ed_los']).dt.seconds / 60830        X_test['ed_los'] = pd.to_timedelta(X_test['ed_los']).dt.seconds / 60831    x1_cols = [x for x in variable_with_id[1:] if not ('ed' in x and 'last' in x)]832    x2_cols = [x for x in df_vitalsign.columns if 'ed' in x]833    train_data_gen = LSTMDataGenerator(X_train, df_vitalsign, y_train, batch_size, x1_cols, x2_cols)834    test_data_gen = LSTMDataGenerator(X_test, df_vitalsign, y_test, batch_size, x1_cols, x2_cols)...

forms.py

Source:forms.py

1from django import forms2from django.conf import settings3from directory_validators.common import not_contains_url_or_email4from directory_validators.company import no_html5from .generics.forms import DITHelpForm, DITHelpModelForm6from .meta import choices, label, help_text, placeholder, validation7from .models import FeedbackModel, TriageModel8from . import fields9class FeedbackForm(DITHelpModelForm):10    title = "Help us improve great.gov.uk"11    subtitle = (12        "Give your feedback on the guidance and services on great.gov.uk. "13        "If something is wrong, give as much detail as you can."14    )15    class Meta:16        model = FeedbackModel17        exclude = []18    content = fields.CharField(19        label="Feedback",20        required=True,21        widget=forms.Textarea,22        attrs={23            'data-message': validation.FEEDBACK,24            'data-validate': 'feedback'25        },26        validators=[not_contains_url_or_email, no_html]27    )28class TriageForm(DITHelpModelForm):29    subtitle = "Application via Department for International Trade"30    submit_text = "Apply to join"31    def get_title(self):32        return self.request.GET.get('market', None)33    class Meta:34        model = TriageModel35        exclude = []36    def save(self, *args, **kwargs):37        # Saving the form submission to the database cannot be justified from38        # a GDPR point of view. It is sent to zendesk.39        return40    company_name = fields.CompanyField(required=True,41                                       label=label.COMPANY_NAME,42                                       help_text=help_text.COMPANY_NAME,43                                       error_messages=validation.TRIAGE_COMPANY_NAME,44                                       attrs={45                                           'data-validate': 'company',46                                           'data-action': 'get-companies',47                                           'data-message': validation.TRIAGE_COMPANY_NAME['required'],48                                           'autocomplete': 'off',49                                           'class': 'form-dropdown-input'50                                       },51                                       button_label="Search Companies House",52                                       button_attrs={53                                           'class': 'button button-border button-border--blue button-medium\54                                           push--ends search-companies',55                                           'data-action': 'get-companies'56                                       })57    soletrader = fields.BooleanField(required=False,58                                     label=label.TRIAGE_UNREGISTERED_COMPANY,59                                     attrs={'data-validate': 'soletrader'})60    company_number = fields.CharField(required=False,61                                      label=label.COMPANY_NUMBER,62                                      attrs={63                                          'class': 'form-control--medium',64                                          'data-validate': 'company-number',65                                          'data-message': validation.TRIAGE_COMPANY_NUMBER['required']66                                      })67    company_postcode = fields.CharField(required=True,68                                        label=label.TRIAGE_POSTCODE,69                                        attrs={70                                            'class': 'form-control--medium',71                                            'data-validate': 'postcode',72                                            'placeholder': placeholder.POSTCODE,73                                            'data-message': validation.TRIAGE_COMPANY_POSTCODE['required']74                                        })75    email_pref = fields.BooleanField(required=False, label=label.EMAIL_PREFERENCE)76    contact_name = fields.CharField(required=True, label=label.CONTACT_NAME,77                                    error_messages=validation.TRIAGE_CONTACT_NAME,78                                    attrs={79                                        'data-validate': 'name',80                                        'data-message': validation.TRIAGE_CONTACT_NAME['required']81                                    })82    contact_email = fields.EmailField(required=True, label=label.CONTACT_EMAIL,83                                      error_messages=validation.TRIAGE_CONTACT_EMAIL,84                                      attrs={85                                          'data-validate': 'email',86                                          'data-message': validation.TRIAGE_CONTACT_EMAIL['required']87                                      })88    website_address = fields.URLField(required=True, label=label.COMPANY_WEBSITE, help_text=help_text.COMPANY_WEBSITE,89                                      error_messages=validation.TRIAGE_COMPANY_WEBSITE,90                                      attrs={91                                          'placeholder': placeholder.URL,92                                          'data-validate': 'url',93                                          'data-message': validation.TRIAGE_COMPANY_WEBSITE['required']94                                      })95    turnover = fields.ChoiceField(required=True, label=label.TRIAGE_SALES, choices=choices.TRIAGE_SALES_THRESHOLDS,96                                  widget=forms.RadioSelect(),97                                  error_messages=validation.TRIAGE_BUSINESS_TURNOVER,98                                  attrs={99                                      'data-validate': 'turnover',100                                      'data-message': validation.TRIAGE_BUSINESS_TURNOVER['required']101                                  },102                                  help_text=help_text.TRIAGE_TURNOVER)103    trademarked = fields.ChoiceField(required=True, label=label.TRIAGE_TRADEMARKED, choices=choices.BOOLEAN_YES_NO,104                                     widget=forms.RadioSelect(),105                                     error_messages=validation.TRIAGE_BUSINESS_TRADEMARK,106                                     attrs={107                                         'data-validate': 'trademark',108                                         'data-message': validation.TRIAGE_BUSINESS_TRADEMARK['required']109                                     },110                                     help_text=help_text.TRIAGE_TRADEMARKED)111    experience = fields.ChoiceField(required=True, label=label.TRIAGE_EXPERIENCE, choices=choices.TRIAGE_EXPERIENCE,112                                    widget=forms.RadioSelect(),113                                    error_messages=validation.TRIAGE_EXPERIENCE_EXPORT,114                                    attrs={115                                        'data-validate': 'export',116                                        'data-message': validation.TRIAGE_EXPERIENCE_EXPORT['required']117                                    })118    description = fields.CharField(required=True, widget=forms.Textarea, help_text=help_text.TRIAGE_DESCRIPTION,119                                   label=label.TRIAGE_DESCRIPTION,120                                   error_messages=validation.TRIAGE_EXPERIENCE_INTRODUCTION,121                                   attrs={122                                       'data-validate': 'description',123                                       'class': 'form-textarea--wide',124                                       'data-message': validation.TRIAGE_EXPERIENCE_INTRODUCTION['required']125                                   })126    contact_phone = fields.IntegerField(required=True, label=label.CONTACT_PHONE, prefix='+44',127                                        error_messages=validation.TRIAGE_CONTACT_PHONE,128                                        attrs={129                                          'data-validate': 'contact-number',130                                          'data-message': validation.TRIAGE_CONTACT_PHONE['required']131                                        })132    sku_count = fields.IntegerField(required=True, label=label.TRIAGE_SKU_NUMBER,133                                    help_text=help_text.TRIAGE_SKU_NUMBER,134                                    error_messages=validation.TRIAGE_BUSINESS_SKU,135                                    attrs={136                                        'class': 'form-control--medium',137                                        'placeholder': placeholder.SKU,138                                        'data-validate': 'sku',139                                        'data-message': validation.TRIAGE_BUSINESS_SKU['required']140                                    })141    @property142    def fieldsets(self):143        contact_details_fields = [144            'contact_name',145            'contact_email',146            'contact_phone',147            'email_pref',148        ]149        if settings.USE_CAPTCHA:150            contact_details_fields.append('captcha')151        return (152            (153                'Your business',154                {155                    'fields': (156                        'company_name',157                        'soletrader',158                        'company_number',159                        'company_postcode',160                        'website_address',161                    )162                }163            ),164            (165                'Business details',166                {167                    'fields': (168                        'turnover',169                        'sku_count',170                        'trademarked',171                    )172                }173            ),174            (175                'Your experience',176                {177                    'fields': (178                        'experience',179                        'description',180                    )181                }182            ),183            (184                'Contact details',185                {186                    'fields': contact_details_fields,187                }188            ),...

main_inference.py

Source:main_inference.py

...23    # ##################24    # # build the model from a config file and a checkpoint file25    model_person = init_detector(config_file_person, checkpoint_file_person, device="cuda:0")26    model_triage = init_detector(config_file_triage, checkpoint_file_triage, device="cuda:0")27    # set(ë²í¸)_drone(ë²í¸)_triage(ë²í¸).jpg28    # set(1~5)_drone(1~3)_triage(1~3).jpg29    # im_folder = "dataset_path/set0"30    im_folder = path31    # ì ì¶ ë ì´ë¸ íì : set_num(1~5), drone_num(1~3), frame_name[ì¬ë§, ê¸´ê¸, ìê¸, ë¹ìê¸]32    #import pdb; pdb.set_trace()33    person_results = []34    set_keys = ["set_1", "set_2", "set_3", "set_4", "set_5"]35    task4_answer = dict.fromkeys(set_keys)36    # #### Extract Region of Person ####37    for set_n in range(1,6):38        set_dict = dict()39        set_name = "set_"+ str(set_n)40        set_dir= im_folder + "set_0" + str(set_n) + "/"41        # print("set_dir : ",set_dir)42        for filename in glob.glob(set_dir): # filename : dataset_path/set_01/...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.