# -*- coding: utf-8 -*-
"""Copy of traffic_signs_detection1.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1727PZD_LY3hPDybzLfsd5n3rmoebMvGW

# CNN Keras

## Data importation
"""

! pip install -q kaggle
!pip install --upgrade --force-reinstall --no-deps kaggle
!mkdir -p ~/.kaggle
!cp kaggle.json ~/.kaggle/
!ls ~/.kaggle
!chmod 600 ~/.kaggle/kaggle.json
!kaggle competitions -h
!kaggle competitions download -c DATASET
from IPython.display import clear_output
clear_output(wait=True)
print("All Good")

!kaggle datasets download -d valentynsichkar/traffic-signs-preprocessed
from IPython.display import clear_output
# clear_output(wait=True)
print("DataSet Downloaded Seccessfully")

!ls

!mkdir trafficSigns
!unzip traffic-signs-preprocessed.zip -d trafficSigns
clear_output(wait=True)
print("file unziped seccessfully")

import pandas as pd
Data = pd.read_pickle("/content/trafficSigns/data0.pickle")

len(Data)

Data.keys()

X_train = Data["x_train"]
Y_train = Data["y_train"]
X_test = Data["x_test"]
Y_test = Data["y_test"]
X_validation = Data["x_validation"]
Y_validation = Data["y_validation"]
labels = Data["labels"]

import matplotlib.pyplot as plt

s = Data["x_train"][2].swapaxes(0,1)
s = s.swapaxes(1,2)
plt.imshow(s)
plt.show()

X_train.shape

X_train = X_train.swapaxes(1,2)
X_train = X_train.swapaxes(2,3)

X_test = X_test.swapaxes(1,2)
X_test = X_test.swapaxes(2,3)

X_validation = X_validation.swapaxes(1,2)
X_validation = X_validation.swapaxes(2,3)

from tabulate import tabulate
data = [["X_train", X_train.shape],
        ["X_test", X_test.shape],
        ["X_validation", X_validation.shape]
      ]
print(tabulate(data, headers=["Data Array", "Shape"], tablefmt="grid"))

len(labels)

plt.figure(figsize=(15,15))

for i in range(len(labels)):
    plt.subplot(11,4,i+1)
    plt.title(labels[Y_train[i]])
    plt.imshow(X_train[i])
    plt.axis("off")

plt.show()

from tensorflow.keras.utils import to_categorical
Y_train = to_categorical(Y_train,43)
Y_test = to_categorical(Y_test,43)

from keras.layers import Dense,Flatten,Conv2D,MaxPool2D
from keras import Sequential

from numpy import matrix
Model = Sequential(
    [
     Conv2D(64, kernel_size = 3, padding = "Same", activation = "relu", input_shape = (32,32,3)),
     MaxPool2D(2),
     Flatten(),
     Dense(500, activation='relu'),
     Dense(43,activation="softmax")
    ]
)

Model.compile(
    optimizer = 'adamax',
    loss = 'categorical_crossentropy',
    metrics = ['accuracy']
)

Model.fit(X_train , Y_train , epochs=2)

Model.evaluate(X_test,Y_test)

Model.save('Traffic-Model1.h5')

Model.fit(X_train , Y_train , epochs=10)

Model.save('Traffic-Model2.h5')

Model.evaluate(X_test,Y_test)

predict = Model.predict(X_validation)

import numpy as np

pred = labels[np.argmax(predict[1])]

print(pred)

plt.imshow(X_validation[1]/255)
plt.show()

"""# YOLOv8

## Importing necessary libraries.
"""

from IPython.display import Image, display
from IPython import display
import os
import random

"""## Installing ultralytics"""

!pip install ultralytics
display.clear_output()
print("Installed successfully!")

"""## Importing YOLO and checking requirements"""

from ultralytics import YOLO
display.clear_output()
!yolo checks

"""## LOADING data from drive"""

!pip install gdown

import gdown

file_id = '1b0fct37QWyRHWpcw31zpeV_nd8KF20ux'

url = f'https://drive.google.com/uc?id={file_id}'

# Specify the destination file path
output = 'data.zip'  # You can change the name if needed

# Download the file
gdown.download(url, output, quiet=False)

# Specify the path to the zip file
zip_file_path = '/content/data.zip'

# Specify the extraction path
extracted_path = '/content/'

# Unzip the file
!unzip {zip_file_path} -d {extracted_path}

"""## Distribution of annotations (classes)"""

Image(filename=f"/content/runs/detect/train/labels.jpg", width=1300)

Image(filename=f"/content/runs/detect/train2/train_batch0.jpg", width=1000)

"""## Start the training with yolo"""

!yolo task=detect mode=train model=yolov8m.pt data=/content/data/data.yaml epochs=40 imgsz=640

"""## Evolution of metrics during the training"""

Image(filename=f"/content/runs/detect/train/results.png", width=1000)

"""## Using the best weights generated from the training to check the validation"""

!yolo task=detect mode=val model=/content/runs/detect/train/weights/best.pt data=/content/data/data.yaml

!yolo task=detect mode=val model=/content/runs/detect/train/weights/last.pt data=/content/data/data.yaml

"""## Samples

### Real labels
"""

Image(filename=f"/content/runs/detect/val/val_batch2_labels.jpg", width=1400)

"""### predicted labels"""

Image(filename=f"/content/runs/detect/val/val_batch2_pred.jpg", width=1400)

"""## Prediction on the test sets"""

!yolo task=detect mode=predict model=/content/runs/detect/train/weights/best.pt conf=0.25 source=/content/data/test/images

!yolo task=detect mode=predict model=/content/runs/detect/train/weights/last.pt conf=0.25 source=/content/data/test/images

"""### samples of the predicted labels

### From best.pt
"""

from IPython.display import Image, display
import os
import random

# Path to the directory containing your images
directory_path = '/content/runs/detect/predict'

# Get a list of all image files in the directory
image_files = [f for f in os.listdir(directory_path) if f.endswith(('.jpg', '.jpeg', '.png', '.gif'))]

# Choose 10 random images
random_images = random.sample(image_files, 10)

# Display the selected images
for image_file in random_images:
    image_path = os.path.join(directory_path, image_file)
    display(Image(filename=image_path))

# Path to the directory containing your images
directory_path = '/content/runs/detect/predict'

# Get a list of all image files in the directory
image_files = [f for f in os.listdir(directory_path) if f.endswith(('.jpg', '.jpeg', '.png', '.gif'))]

# Choose 10 random images
random_images = random.sample(image_files, 10)

# Display the selected images
for image_file in random_images:
    image_path = os.path.join(directory_path, image_file)
    display(Image(filename=image_path))

"""### From last.pt"""

# Path to the directory containing your images
directory_path = '/content/runs/detect/predict3'

# Get a list of all image files in the directory
image_files = [f for f in os.listdir(directory_path) if f.endswith(('.jpg', '.jpeg', '.png', '.gif'))]

# Choose 10 random images
random_images = random.sample(image_files, 10)

# Display the selected images
for image_file in random_images:
    image_path = os.path.join(directory_path, image_file)
    display(Image(filename=image_path))

# Path to the directory containing your images
directory_path = '/content/runs/detect/predict3'

# Get a list of all image files in the directory
image_files = [f for f in os.listdir(directory_path) if f.endswith(('.jpg', '.jpeg', '.png', '.gif'))]

# Choose 10 random images
random_images = random.sample(image_files, 10)

# Display the selected images
for image_file in random_images:
    image_path = os.path.join(directory_path, image_file)
    display(Image(filename=image_path))

"""## Doing further training hoping for better results"""

!yolo task=detect mode=train model=/content/runs/detect/train/weights/last.pt data=/content/data/data.yaml epochs=20 imgsz=640

!yolo task=detect mode=val model=/content/runs/detect/train2/weights/best.pt data=/content/data/data.yaml

!yolo task=detect mode=val model=/content/runs/detect/train2/weights/last.pt data=/content/data/data.yaml