import json import urllib.parse import boto3 from csv import reader import pandas as pd import pymysql import numpy as np import awswrangler as wr rds_host_write = "database-ambt.cluster-cvnwxgmngsms.ap-northeast-2.rds.amazonaws.com" rds_host_read = "database-ambt.cluster-ro-cvnwxgmngsms.ap-northeast-2.rds.amazonaws.com" name = "admin" password = "hdci12#$" db_name = "ambt" PORT=3306 def lambda_handler(event, context): conn = pymysql.connect(host=rds_host_read, user=name, passwd=password, port=PORT, database=db_name, charset='utf8') #conn = pymysql.connect(host=rds_host, user=name, passwd=password, db=db_name, use_unicode=True, charset='utf8', connect_timeout=5) sql_statement = "SELECT * FROM `ambt.icos`.facility_type;" facility_type_id = pd.read_sql(sql=sql_statement, con=conn) sql_statement = "SELECT * FROM `ambt.icos`.facility_code;" material_code = pd.read_sql(sql=sql_statement, con=conn) sql_statement = "SELECT * FROM `ambt.icos`.control_value;" control_value = pd.read_sql(sql=sql_statement, con=conn) sql_statement = "SELECT * FROM `ambt.icos`.value_type;" value_type = pd.read_sql(sql=sql_statement, con=conn) sql_statement = "SELECT * FROM `ambt.icos`.ambt_anoicos_code;" mappingTable = pd.read_sql(sql=sql_statement, con=conn) print(mappingTable.head()) for record in event['Records']: #print("Received event: " + json.dumps(event, indent=2)) bucket = record['s3']['bucket']['name'] key = urllib.parse.unquote_plus(record['s3']['object']['key']) print('bucket : ', bucket) print('key : ', key) s3 = boto3.client('s3') obj = s3.get_object(Bucket=bucket, Key=key) raw_data = pd.read_csv(obj['Body'], sep=',') raw_data = raw_data.drop(['cnt'], axis=1) # raw_data = raw_data.drop(['Unnamed: 0'], axis=1) raw_data = raw_data.dropna(axis=0) # drop the row including null raw_data = raw_data.reset_index(drop=True) print('raw', raw_data.columns) print (raw_data[130:140]) print('length of raw data', len(raw_data)) # --------------- Data pre-processing (considering data minimum properties) # temperature, humidity (set point, ) - -15 ~ 100 # operating status - 0/1 # open ratio (SP) - 0 ~ 100 # operating mode - 0 ~ 10 processed_data = raw_data.copy() # control_value = 3(온도), 4(습도), 5(온도 설정값), 6(습도 설정값) target_row = mappingTable.loc[(mappingTable['site_id']==1) & (mappingTable['collect_status'] == 1) & ((mappingTable['control_value'] == 3) | (mappingTable['control_value'] == 4) | (mappingTable['control_value'] == 5) | (mappingTable['control_value'] == 6))] for c_idx in range(raw_data[list(set(target_row['raw_tag']))].shape[1]): processed_data[list(set(target_row['raw_tag']))[c_idx]] \ = [r_value if r_value > -15 and r_value <= 100 else np.nan \ for r_value in raw_data[list(set(target_row['raw_tag']))[c_idx]]] # control_value = 1(운전상태) target_row = mappingTable.loc[(mappingTable['site_id']==1) & (mappingTable['collect_status'] == 1) & (mappingTable['control_value'] == 1) ] for c_idx in range(raw_data[list(set(target_row['raw_tag']))].shape[1]): processed_data[list(set(target_row['raw_tag']))[c_idx]] \ = [r_value if r_value == 0 or r_value == 1 else np.nan \ for r_value in raw_data[list(set(target_row['raw_tag']))[c_idx]]] # control_value = 7(개도율), 8(개도율 설정값) target_row = mappingTable.loc[(mappingTable['site_id']==1) & (mappingTable['collect_status'] == 1) & ((mappingTable['control_value'] == 7) | (mappingTable['control_value'] == 8)) ] for c_idx in range(raw_data[list(set(target_row['raw_tag']))].shape[1]): processed_data[list(set(target_row['raw_tag']))[c_idx]] \ = [r_value if r_value >= 0 or r_value <= 100 else np.nan \ for r_value in raw_data[list(set(target_row['raw_tag']))[c_idx]]] # control_value = 30(운전모드) target_row = mappingTable.loc[(mappingTable['site_id']==1) & (mappingTable['collect_status'] == 1) & (mappingTable['control_value'] == 30) ] for c_idx in range(raw_data[list(set(target_row['raw_tag']))].shape[1]): processed_data[list(set(target_row['raw_tag']))[c_idx]] \ = [r_value if r_value >= 0 and r_value <= 10 else np.nan \ for r_value in raw_data[list(set(target_row['raw_tag']))[c_idx]]] print('sum of null point:', processed_data.isnull().sum().sum()) # --------------- Fill the missing data and round off time index time_resolution = 5 #data = raw_data round_m_comp = 0 missing_idx = [] missing_date = [] print('len(processed_data))',len(processed_data)) for time_idx in range(len(processed_data)): str_date = processed_data['time'][time_idx] date = str_date.split()[0] round_m = int(int(str_date.split()[1].split(':')[1])/time_resolution)*time_resolution h, m, s = [str_date.split()[1].split(':')[0], str(round_m) if round_m >= 10 else '0'+str(round_m), '00'] round_m_comp += time_resolution # for compare time index # ----- find the missing row ----- # while round_m_comp <= int(h) * 60 + round_m: round_m_comp_tmp = round_m_comp - time_resolution round_h_tmp = int(round_m_comp_tmp/60) round_m_tmp = int(round_m_comp_tmp%60) h_tmp, m_tmp, s_tmp = [str(round_h_tmp) if round_h_tmp >= 10 else '0'+str(round_h_tmp), str(round_m_tmp) if round_m_tmp >= 10 else '0'+str(round_m_tmp), '00'] missing_date.append(date + ' ' + h_tmp + ':' + m_tmp + ':' + s_tmp) missing_idx.append(time_idx + len(missing_idx)) # save missing index considering append index round_m_comp += time_resolution # ----- find the missing row ----- # processed_data.at[time_idx,'time'] = date + ' ' + h + ':' + m + ':' + s # round off time index for rows of normal data print('len', len(processed_data)) print('sum of missing row data:', len(missing_idx)) # ----- fill nan on the missing row ----- # idx = 0 for miss_idx in missing_idx: tmp_data = np.zeros((1,len(raw_data.columns))) tmp_data[:] = np.nan reconstructed_data = pd.DataFrame(tmp_data, columns = raw_data.columns) idx_temp = miss_idx temp1 = processed_data[processed_data.index < idx_temp].copy() temp2 = processed_data[processed_data.index >= idx_temp].copy() temp1 = temp1.append(reconstructed_data, ignore_index=True) processed_data = temp1.append(temp2, ignore_index=True) processed_data.at[idx_temp, 'time'] = missing_date[idx] idx += 1 # ----- fill nan on the missing row ----- # # ----- interpolation ----- # # ControlValue = 3(온도), 4(습도) TemHum_row = mappingTable.loc[(mappingTable['site_id']==1) & (mappingTable['collect_status'] == 1) & ((mappingTable['control_value'] == 3) | (mappingTable['control_value'] == 4)) ] temp1 = processed_data[list(processed_data[list(set(TemHum_row['raw_tag']))])].fillna(method='backfill').copy() # ControlValue = 1(운전상태(On/off)), 7(개도율), 8(개도율 설정값), 19(차압), 30(운전모드) StatusOpenRatio_row = mappingTable.loc[(mappingTable['site_id']==1) & (mappingTable['collect_status'] == 1) & ((mappingTable['control_value'] == 1) | (mappingTable['control_value'] == 7) | (mappingTable['control_value'] == 8) | (mappingTable['control_value'] == 19) | (mappingTable['control_value'] == 30)) ] temp2 = processed_data[list(processed_data[list(set(StatusOpenRatio_row['raw_tag']))])].fillna(method='backfill').copy() others_tag = [x for x in list(set(raw_data.columns)) if x not in list(set(TemHum_row['raw_tag']))] others_tag = [x for x in others_tag if x not in list(set(StatusOpenRatio_row['raw_tag']))] #print('iik',list(set(raw_data.columns)).remove(list(set(TemHum_row['raw_tag'])) + list(set(StatusOpenRatio_row['raw_tag'])))) others = processed_data[others_tag].fillna(method='backfill').copy() processed_data = pd.concat([others, temp1, temp2], axis=1) processed_data = processed_data.round(1) processed_data = processed_data.fillna(method='pad') # ----- interpolation ----- # print('processed', processed_data) # ----- load to S3 prep bucket ----- # curated_file_key = key.replace(key.split('/')[-1], '') target_path = 's3://hdci-ambt-anoicos-prep/{}'.format(curated_file_key) print('target_path', target_path) wr.s3.to_csv( df=processed_data, path=target_path, mode='overwrite', dataset=True ) # ----- load to S3 prep bucket ----- #