csvgen.py

import csv
from datetime import datetime, timedelta
import random
import numpy as np

# Configuración
COCO_LABELS_PATH = './coco_80_labels_list.txt'
OUTPUT_FILE = 'data.csv'
START_DATE = datetime(2025, 3, 1, 0, 0, 0)
END_DATE = datetime(2025, 4, 1, 0, 0, 0)

def cargar_labels():
    with open(COCO_LABELS_PATH, 'r') as f:
        return [line.strip() for line in f if line.strip()]

def generar_datos_objeto(hora, dia_semana, label):
    # Frecuencias base por tipo de objeto (ejemplo)
    frecuencias = {
        'person': 0.8, 'car': 0.7, 'chair': 0.5, 'book': 0.3,
        'truck': 0.4, 'dog': 0.2, 'cell phone': 0.6, 'bottle': 0.5
    }
    base = frecuencias.get(label, 0.2)  # Valor por defecto para objetos menos comunes
    
    # Variación por hora y día
    if dia_semana < 5:  # Días laborables
        if 7 <= hora <= 9 or 17 <= hora <= 19:
            multiplicador = random.uniform(1.5, 3.0)  # Horas pico
        else:
            multiplicador = random.uniform(0.5, 1.2)
    else:  # Fin de semana
        if 10 <= hora <= 20:
            multiplicador = random.uniform(1.2, 2.0)
        else:
            multiplicador = random.uniform(0.3, 0.8)
    
    return max(0, int(np.random.poisson(base * multiplicador * 10)))

# Generar el archivo CSV
labels = cargar_labels()
with open(OUTPUT_FILE, 'w', newline='') as csvfile:
    writer = csv.DictWriter(csvfile, fieldnames=['Timestamp'] + labels)
    writer.writeheader()
    
    current = START_DATE
    while current < END_DATE:
        fila = {'Timestamp': current.strftime('%d/%m/%Y %H:%M:%S')}
        for label in labels:
            fila[label] = generar_datos_objeto(
                hora=current.hour,
                dia_semana=current.weekday(),
                label=label
            )
        
        writer.writerow(fila)
        current += timedelta(hours=1)

print(f'Archivo {OUTPUT_FILE} generado con {len(labels)} objetos COCO')