import pandas as pd
import time
import csv
import os
from time import gmtime, strftime
import numpy as np
import matplotlib.pyplot as plt

PATH = "/src/experiment1/output/gmail/greenifier-demo-scaling/raw-output/0/seed=0/"
CONSTANT = 3_6000_000 / 24


def create_dataframes() -> tuple[pd.DataFrame, pd.DataFrame]:
    cwd = os.getcwd()
    cwd = cwd + PATH
    print(cwd)
    hosts: str = "%shost.parquet" % cwd
    tasks: str = "%stask.parquet" % cwd

    try:
        df_hosts = pd.read_parquet(hosts)
        df_tasks = pd.read_parquet(tasks)
        return (df_hosts, df_tasks)

    except Exception:
        print("Exception: error opening files.")
        exit(1)


def get_name() -> str:
    curr = time.time()
    s = strftime("%d_%b_%Y_%H%M%S", gmtime(curr))
    return s


def iterate(frame):
    dictionary = {}
    for i in range(len(frame)):
        if frame["downtime"].iloc[i] > 0.0:
            ts = frame["timestamp"].iloc[i]
            dictionary[(ts)] = dictionary.get((ts), 0) + 1
    return dictionary


# This is a running function frequently changed
def plot_hosts(frame: pd.DataFrame):
    dictionary = {
        216000000: 277,
        219600000: 277,
        280800000: 139,
        727200000: 139,
        730800000: 139,
        802800000: 162,
        1407600000: 139,
        1850400000: 139,
        1857600000: 277,
        1861200000: 277,
        1872000000: 139,
        1926000000: 62,
        2023200000: 162,
        2037600000: 277,
        2041200000: 208,
        2070000000: 139,
    }

    dictionary2 = {
        3600000: 11,
        180000000: 11,
        212400000: 103,
        216000000: 103,
        219600000: 103,
        223200000: 103,
        374400000: 57,
        712800000: 103,
        795600000: 57,
        838800000: 11,
        882000000: 103,
        975600000: 57,
        979200000: 11,
        982800000: 11,
        1087200000: 11,
        1234800000: 11,
        1404000000: 11,
        1854000000: 103,
        1857600000: 103,
        1861200000: 103,
        1926000000: 36,
        2034000000: 103,
        2037600000: 103,
        2041200000: 34,
        2080800000: 11,
        2102400000: 11,
        2163600000: 11,
        2185200000: 57,
        2383200000: 57,
    }

    for key in dictionary2.keys():
        dictionary2[key] += 150

    df = pd.DataFrame(list(dictionary2.items()), columns=["timestamp", "count"])
    df3 = iterate(frame)
    df2 = pd.DataFrame(list(df3.items()), columns=["timestamp", "count"])

    plt.plot(
        df2["timestamp"] / CONSTANT,
        df2["count"],
        label="Actual failures",
        linewidth=1,
        zorder=2,
    )
    plt.plot(
        df["timestamp"] / CONSTANT,
        df["count"],
        label="Detected failures",
        linewidth=3,
        zorder=1,
    )
    plt.xlabel("Time [days]")
    plt.ylabel("Failures per timestamp")
    plt.title("Failure detection results")
    plt.legend()

    path = os.getcwd() + "/src/experiment1/"
    location: str = path + "figures/%s.pdf" % get_name()
    plt.savefig(location, dpi=300)
    total = sum(dictionary.values())
    total2 = sum(df3.values())
    print(f"Percentage detected: {total/total2}")


def export_results(path: str):
    assert path
    # 0 is hosts, 1 is tasks
    frames: tuple = create_dataframes()
    plot_hosts(frames[0])


def main():
    export_results(PATH)


if __name__ == "__main__":
    main()