# ------------------------------------------------------------------------------------------------------------------
## This code is for checking Overloading transformers conditions
## Input Files : >> CSV of Timestamp and Simulated measured Power (VA)
## Output >> Overloaded Transformers depend on % of Overloaded condition provided with the duration of Overloading
## To be able to check overloading conditions, the user must specify the rated value of the distribution transformers
# ------------------------------------------------------------------------------------------------------------------
import pandas as pd
import argparse
import time
import numpy as np
import os
def get_power_rates(cols):
    power_rates = []
    # compare against the given Y value
    for i in cols:
        power_rates.append(float(input(f"enter {i} Power Rate: ")))
    return power_rates
def get_file_names():
    files = []
    for c in cols:
        files.append(input(f"enter name of output file for {c}: "))
    return files
def query_csv():
    vals = []
    for y,c in zip(power_rates,cols):
        vals.append(csv.query(f"`{c}` >= {y}"))
    return vals
def first_version():
    '''
    first verison just writes the query results to their respective files
    '''
    for v,f,c in zip(query_results,output_files,cols):
        v[["# timestamp",f"{c}"]].to_csv(f"{f}",index=False)
    return
def query(df, col, t_delta, ep, pr,step):
    sz = csv.shape[0]
    res = []
    result = []
    df = csv[['# timestamp',col,'timestamp']] # change from csv to df
    i = 0
    steps = pd.Timedelta(step,unit='m')
    if i<sz:
        prev = df.loc[0]
    while i < sz:
        cur = df.loc[i]
        if cur['timestamp'] == (prev['timestamp'] + steps) and (abs(cur[col] - pr) <= ep) and (abs(prev[col] - pr) <= ep):
            res.append(prev)
            if t_delta/step == len(res):
                res.append(cur)
                result.append(res)
                res = []
        else:
            res = []
        prev = cur
        i += 1
    return result
def open_file(FILE):
    csv = pd.read_csv(FILE)
    cols = csv.columns[1:]
    csv['timestamp'] = pd.to_datetime(csv['# timestamp']) # convert normal timestamp format to datetime timestamp
    csv['timestamp'].apply(lambda x: time.mktime(x.timetuple())) # converts datetime timestamp to unix timestamp
    return (csv,cols)
def second_version(csv):
    '''
    200%   ----------------- 30 minutes
    150%   ----------------- 1 hour
    125%   ----------------- 4 hours
    others ----------------- remaining
    second version checks for time constraints before writing the query results to their respective files
    '''
    specs = {} # these specs define the previously stated rules
    # query for the 1.25  === %125
    # there will be only 1 column (besides timestamps) & 1 power_rate for the column
    pr = power_rates[0]
    col = cols[0]
    # ------------- 200% -------------
    x = 2 * pr
    # 10% of x
    ep = 0.1 * x
    print(f"{50*'-'} calculating 200% {50*'-'}")
    res = query(csv,col,30 ,ep,x, 10) # 30 minutes and 10 minutes step ( in case that changes in the future)
    if len(res) > 0:
        res = sum(res,[])
        df = pd.DataFrame(res)
        print(df)
        specs[f"#{50*'-'} {FILE}--(200%) {50*'-'}"] = df
    # ------------- 150% -------------
    x = 1.5 * pr
    # 10% of x
    ep = 0.1 * x
    print(f"{50*'-'} calculating 150% {50*'-'}")
    res = query(csv,col,60 ,ep,x, 10) # 60 minutes and 10 minutes step ( in case that changes in the future)
    if len(res) > 0:
        res = sum(res,[])
        df = pd.DataFrame(res)
        print(df)
        specs[f"#{50*'-'} {FILE}--(150%) {50*'-'}"] = df
    # ------------- 125% -------------
    x = 1.25 * pr
    # 10% of x
    ep = 0.1 * x
    print(f"{50*'-'} calculating 125% {50*'-'}")
    res = query(csv,col,240 ,ep,x, 10) # 240 minutes and 10 minutes step ( in case that changes in the future)
    if len(res) > 0:
        res = sum(res,[])
        df = pd.DataFrame(res)
        print(df)
        specs[f"#{50*'-'} {FILE}--(125%) {50*'-'}"] = df
    return specs
def write_output(fname,output):
    if os.path.exists(fname):
        os.remove(fname)
    for DICT in output:
        if DICT:
            for k,v in DICT.items():
                with open(fname,'a+') as f:
                    f.write(k+'\n')
                del v['timestamp']
                v.to_csv(fname,mode='a',index=False)
    return

parser = argparse.ArgumentParser(description='.')
parser.add_argument('-f', type=str, help='input file name',default='')
parser.add_argument('-d', type=str, help='input dir name',default='')
parser.add_argument('-m', type=int, help='mode of operation (0: old version; 1: percentage mode)', default=1)
args = parser.parse_args()
FILE = args.f # grab file name
mode = args.m # grab the mode (version of file)
DIR = args.d # grab dir
if FILE == '' and DIR=='':
    parser.print_help()
    exit(0)

if mode == 0:
    if FILE == '':
        parser.print_help()
        exit(0)
    (csv,cols) = open_file(f'{FILE}')
    power_rates = get_power_rates(cols) # get power rates
    query_results = query_csv()
    output_files = get_file_names()
    # change columns to floats
    for i in cols:
        csv[f"{i}"] = csv[f"{i}"].astype(float)
    first_version()
elif mode == 1:
    output = []
    if DIR == '':
        parser.print_help()
        exit(0)
    # read all files in the given directory
    if DIR[-1] != '/':
        DIR += '/'
    for FILE in os.listdir(DIR):
        print(f"{'*'*50} {FILE} {'*'*50}")
        (csv,cols) = open_file(f'{DIR}{FILE}')
        power_rates = get_power_rates(cols) # get power rates
        res = second_version(csv)
        output.append(res)
    write_output('output.csv',output)
    # second_version()
#--------------------------------------------------------------------------------------------------------