import pickle
import os
import pandas as pd
import numpy as np
import argparse
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, mean_squared_error
from sklearn.preprocessing import label_binarize

workspace_path = os.getenv('WORKSPACE')
pickle_path = '/app/model_with_data.pickle'

if os.path.exists(pickle_path):
    with open(pickle_path, 'rb') as file:
        loaded_data = pickle.load(file)
else:
    ## else if aby było mozna uruchomić lokalnie
    with open('model_with_data.pickle', 'rb') as file:
        loaded_data = pickle.load(file)

# Wczytanie modelu
model = loaded_data[0]

#Wczytanie danych testowych
X_test_scaled = loaded_data[3]
y_test_encoded = loaded_data[4]

# Predykcja
y_pred = model.predict(X_test_scaled)
y_pred_classes = np.argmax(y_pred, axis=1)

# Zapisanie wyników predykcji do pliku
np.savetxt('results_prediction.csv', y_pred, delimiter=',')

# Wyliczenie zbiorczych metryk i zapis do pliku - zadanie 2 - LAB 06
accuracy = accuracy_score(np.argmax(y_test_encoded, axis=1), y_pred_classes)
precision = precision_score(np.argmax(y_test_encoded, axis=1), y_pred_classes, average='micro')
recall = recall_score(np.argmax(y_test_encoded, axis=1), y_pred_classes, average='micro')
f1 = f1_score(np.argmax(y_test_encoded, axis=1), y_pred_classes, average='micro')
rmse = np.sqrt(mean_squared_error(np.argmax(y_test_encoded, axis=1), y_pred_classes))

print("Metrics results")
print("Accuracy:", accuracy)
print("Micro-avg Precision:", precision)
print("Micro-avg Recall:", recall)
print("F1 Score:", f1)
print("RMSE:", rmse)

# Zapisanie metryk do pliku CSV
parser = argparse.ArgumentParser()
parser.add_argument('--build-number', type=int, help='Build number')
args = parser.parse_args()

build_number = args.build_number
print(f"Numer builda: {build_number}")

metrics_file = os.path.join('metrics.csv')

metrics_data = {
    'Build Number': [build_number],
    'Accuracy': [accuracy],
    'Micro-avg Precision': [precision],
    'Micro-avg Recall': [recall],
    'F1 Score': [f1],
    'RMSE': [rmse]
}

df = pd.DataFrame(metrics_data)

if os.path.exists(metrics_file):
    # Odczytanie istniejącego pliku CSV
    existing_df = pd.read_csv(metrics_file)
    df = pd.concat([existing_df, df], ignore_index=True)

# Zapisanie metryk do pliku CSV
df.to_csv(metrics_file, index=False)

# Wyświetlenie informacji o zapisanym pliku
print("Metryki zostały zapisane do pliku:", metrics_file)