mabittar

services:
  db:
    image: pgvector/pgvector:pg17
    restart: always
    container_name: pgvector-db
    environment:
      POSTGRES_USER: postgres
      POSTGRES_PASSWORD: mysecretpassword
      POSTGRES_DB: vectordb
    ports:

mabittar

#################################################
#### Automatic Pythonic Developer Environment ####
#################################################
##
## If you don't really know what to do, run `make help`.
## If you don't have make installed,
## To be used on Linux system
## Make sure you have `pyenv` installed beforehand
##
## https://github.com/pyenv/pyenv

mabittar

1. Install the fonts on your computer:

   git clone https://github.com/powerline/fonts.git
   cd fonts
   ./install.sh
   cd .. && rm -rf fonts

2. Then, open the settings (ctrl+p > "settings json" > enter) in vscode.

mabittar

# Download new VSCode Version
echo "Update and Upgrade"
sudo apt update -y
sudo apt upgrade -y

CODEPATH=~/Downloads/vscode.deb

echo "Download new VSCode version"
curl -L "https://code.visualstudio.com/sha/download?build=stable&os=linux-deb-x64" -o "$CODEPATH"

mabittar

n_entradas = X_train.shape[1]

keras_model = Sequential([
         Dense(n_entradas, input_shape=(n_entradas, ), activation='relu'),
         Dense(32, activation='relu'),
         Dense(2, activation='softmax')                 
])

from keras.utils.vis_utils import plot_model
plot_model(keras_model, show_shapes=True, show_layer_names=True)

mabittar

#1. Escolha do modelo
from sklearn.tree import DecisionTreeClassifier

#2. Instanciar o modelo e escolher os hyperparametros
model_tree = DecisionTreeClassifier(max_depth=4, criterion="entropy")

#3. Fit do modelo
model_tree.fit(X_rand, y_rand)

#Fazer previsões em cima dos novos dados

mabittar

#1. Escolher o modelo
from sklearn.linear_model import LogisticRegression

#2. Instanciar o modelo
np.random.seed(2)
model_lr = LogisticRegression()

#3. Fit do modelo
model_lr.fit(X_rand, y_rand)

mabittar

from imblearn.under_sampling import RandomUnderSampler

rand = RandomUnderSampler()
X_rand, y_rand = rand.fit_sample(X_train, y_train)

mabittar

#remover colunas que contém dados não numéricos
cols_to_remove = [] #criar uma lista vazia onde serão armazenadas as colunas a serem removidas

for col in df.columns:
    try:
        _ = df[col].astype(float) #se os dados na coluna forem do tipo float(numerico) OK
    except ValueError: #se os dados na coluna forem diferentes de numérico, serão acrescidas a lista de colunas removidas
        print('Nao foi possivel converter a coluna %s para float' % col) #mostra quais colunas serão removidas
        cols_to_remove.append(col) #adicionar a coluna para a lista de colunas a remover
        pass

mabittar

# importar os pacotes necessários
import pandas as pd

# importar o arquivo
df = pd.read_csv("train.csv") #aqui você deve especificar o nome do arquivo .csv que deseja importar

# ver 5 as primeiras entradas
df.head()