# Import necessary libraries
import pandas as pd
from datasets import load_dataset, Dataset
from transformers import AutoTokenizer, AutoModelForSequenceClassification, Trainer, TrainingArguments
from sklearn.model_selection import train_test_split

# Load the dataset
# Make sure you have the correct path to the CSV file
df = pd.read_csv('diabetes_data.csv')

# Define target column and preprocess
threshold_value = 0  # Set threshold if needed
df['label'] = (df['hypertension'] > threshold_value).astype(int)  # Binary classification based on hypertension

# Split the dataset into train and test sets
train_df, test_df = train_test_split(df, test_size=0.2, random_state=42)
train_dataset = Dataset.from_pandas(train_df)
test_dataset = Dataset.from_pandas(test_df)

# Load tokenizer and model
model_name = "bert-base-uncased"  # Replace with any compatible model from Hugging Face Model Hub
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name, num_labels=2)

# Tokenization function
def preprocess_function(examples):
    # Concatenate relevant columns to form the input text if needed
    inputs = examples["age"].astype(str) + " " + examples["bmi"].astype(str) + " " + examples["HbA1c_level"].astype(str)
    return tokenizer(inputs, padding="max_length", truncation=True, max_length=32)

# Apply tokenization to the datasets
tokenized_train = train_dataset.map(preprocess_function, batched=True)
tokenized_test = test_dataset.map(preprocess_function, batched=True)

# Set up Trainer with training arguments
training_args = TrainingArguments(
    output_dir="./results",
    evaluation_strategy="epoch",
    per_device_train_batch_size=16,
    per_device_eval_batch_size=16,
    num_train_epochs=3,
    weight_decay=0.01,
)

# Initialize Trainer
trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_train,
    eval_dataset=tokenized_test,
)

# Train and evaluate
trainer.train()
trainer.evaluate()