path: /home/username/venv_folder/venv_name/yolov8-project/ # absolute path to dataset
test: test/images # relative path to test images
train: train/images # relative path to training images
val: val/images # relative path to validation images

# classes
names:
  0: postage-stamp

import torch, ultralytics


# Check library versions
print("PyTorch version:", torch.__version__)
print("Ultralytics version:", ultralytics.__version__)


# Check if GPU is available
if torch.cuda.is_available():
    print(f"GPU is available: {torch.cuda.get_device_name(0)}")
else:
    print("GPU is not available")

from ultralytics import YOLO


# Load model with pretrained weights (recommended)
model = YOLO("yolov8n.pt") 


# Optionally load a model without pretrained weights
# model = YOLO("yolov8n.yaml")


# You can also try yolo11n
# model = YOLO("yolo11n.pt")


# Train model
model.train(data="config.yaml", epochs=100, patience=10)

# Load the trained YOLO model
model = YOLO("/home/username/venv_folder/venv_name/runs/detect/train/weights/best.pt")


# Specify folder containing images for object detection
image_folder = "/home/username/venv_folder/venv_name/images"


# Perform object detection on every image in specified folder
results = model(
    source=image_folder,
    show=False,  # Set to True if you want to display the image
    cnotallow=0.60,  # Confidence threshold
    save=True,  # Save results
    save_txt=True,  # Save results as text files
    save_cnotallow=True,  # Save confidence scores
    line_width=3,  # Adjust line width for bounding boxes
    save_crop=True  # Save cropped detections as image files
    )

import os
import pandas as pd


# Path to folder containing result text files
folder_path = "/home/username/venv_folder/venv_name/runs/detect/predict/labels"


# Empty list to store data from each file
dataframes = []


# Define column names
column_names = ["class", "x_center", "y_center", "width", "height", "confidence"]


# Iterate over all result files in the folder
for filename in os.listdir(folder_path):
    file_path = os.path.join(folder_path, filename)
    df = pd.read_csv(file_path, delimiter=' ', header=None, names=column_names,)
    df['filename'] = filename  # Add column to record filename of each result
    dataframes.append(df) # Append results to dataframes list


# Combine all into a single DataFrame
combined_df = pd.concat(dataframes, ignore_index=True)


# Check the top five rows of the DataFrame
display(combined_df.head(5))


# Count number of rows in DataFrame
num_rows = combined_df.shape[0]
print(f"Number of detected stamps: {num_rows}")