# Instantiating Main Python Script File
# Collects stuff from the rest of the scripts
import pandas as pd
import numpy as np

# containment breach
import scipy as scp
from sklearn.model_selection import train_test_split
from sklearn.cluster import KMeans
from sklearn import metrics
import gunner
import digger
import scout

# Instantiating globals to be used in other files
global games_merged_dat
global games_sales_split_pre
global games_sales_split_dur
global games_sales_split_pos

games_review = pd.read_csv("datasets/videogames/Games.xls")
games_sales = scout.cure_depression(
    pd.read_csv("datasets/videogames/vgsales-12-4-2019-short.csv")
)

print(games_review.count())
print(games_sales.count())

games_review_phase1 = digger.slice_column(games_review, "GameName", "Review")
games_review_final = digger.slice_column(games_review, "GameName", "(Import)")

games_merged_dat = digger.write_joined_df(games_sales, games_review_final)


# Acquisition of Merged dataset
print(games_merged_dat.count())

# Loading Crime Datasets
crime_CA = pd.read_excel("datasets/crime/clean_crime_canada_dataset.xlsx")

crime_US = pd.read_csv("datasets/crime/report.csv")

print(crime_US.isnull().count())
print(crime_CA.isnull().count())

year_interval = gunner.year_interval(crime_US, crime_CA, "report_year", "year")

year_max = year_interval[0]
year_min = year_interval[1]

crime_intersect = gunner.intersect_by_year(crime_US, crime_CA, "report_year", "year")

crime_US_intersect = crime_intersect[0]
crime_CA_intersect = crime_intersect[1]

NA_col_list = [
    "JP_Sales",
    "Other_Sales",
    "Global_Sales",
    "PAL_Sales",
    "GameName",
    "Review",
    "Console",
    "Score",
]
GLO_col_list = [
    "JP_Sales",
    "Other_Sales",
    "NA_Sales",
    "PAL_Sales",
    "GameName",
    "Review",
    "Console",
    "Score",
]

# Splitting crime datasets
# Collecting Split-Up Datasets
games_merged_dat = gunner.drop_kick(NA_col_list, games_merged_dat)
games_merged_dat.to_csv("datasets/videogames/games_merged.csv", index=False)

sale_tri_split = gunner.trisect_by_year(games_merged_dat, "Year", year_interval)

games_sales_split_pre = sale_tri_split[0]
games_sales_split_dur = sale_tri_split[1]
games_sales_split_pos = sale_tri_split[2]

# Displaying Acquired Data
print("Dataset Info:\n")
games_sales_split_pre.info()
games_sales_split_dur.info()
games_sales_split_pos.info()

print("Yer forsaken Statistical Description:\n", games_sales_split_dur.describe())

print(
    games_sales_split_pre.head(5),
    games_sales_split_dur.head(5),
    games_sales_split_pos.head(5),
)

# Load merged gammas

# Required to use binning for cleaning, idk (DONE LESGOOOOOOOOOOOOOOOOOOOO)
# https://towardsdatascience.com/data-preprocessing-with-python-pandas-part-5-binning-c5bd5fd1b950
gammas = pd.read_csv("datasets/videogames/games_merged.csv")
labels = ["smol", "epik", "larg"]
gammas = digger.slam_dunk(gammas, "Critic_Score", labels=labels)
# gammas = gammas[gammas["Genre"].isna() == False]
# gammas = scout.cure_depression(gammas)

# Also need to transform using Z-score (normal distr go brrrr lmao), or min-max
# ah, scheiße
# nvm, done, kekW
gammas["Critic_Score_Norm"] = scout.scaling_zscore(gammas, "Critic_Score")
print(gammas["Critic_Score_Norm"].head(10))

# Saving all into a file
gammas = gammas.dropna(how="any", axis=0)  # nuke them empty poopers
gammas.to_csv("datasets/videogames/games_cleanish.csv", index=False)
# Need similarity and dissimialrity, scipy time
# Selecting 5 random rows
chosen_idx = np.random.choice(len(gammas), replace=False, size=5)
sample_rows = gammas.iloc[chosen_idx]
print(sample_rows.head())

scout.dissimilarity(sample_rows)
scout.similarity(sample_rows)
# split the data set
gammas_train, gammas_test = train_test_split(gammas, test_size=0.20, random_state=69)
gammas_train.to_csv("datasets/videogames/games_train.csv", index=False)
gammas_test.to_csv("datasets/videogames/games_test.csv", index=False)

# kmeans pls
gammas_train_kmeans = KMeans(n_clusters=10, random_state=420, n_init="auto").fit(
    gammas_train[["Critic_Score", "User_Score", "Total_Shipped"]]
)
gammas_labels = gammas_train_kmeans.labels_

silh_score = metrics.silhouette_score(
    gammas_train[["Critic_Score", "User_Score", "Total_Shipped"]],
    gammas_labels,
    metric="euclidean",
)
print(silh_score)
gammas_train["Kmean Labels"] = gammas_labels
print(gammas_train.head())