# Put your MCVE code here
import dask
import dask.bag as db
import river
# Create a Dask bag from your data
df=pd.DataFrame([[0]*2],columns=['VendorID','fare_amount'])
data = db.from_sequence(df, npartitions=4)

# Define a function to process and train on each partition
def process_and_train(partition):
    X_train,X_test,y_train,y_test=get_dask_train_test(partition)
    model = river.linear_model.LinearRegression(optimizer=river.optim.SGD(0.01), l2=0.1)
    # Stream learning from the DataFrame
    for _,row in partition.iterrows():
        y = row['fare_amount']       # Target
        x = row.drop('fare_amount')  # Features
        model = model.learn_one(x, y)
    print("done")
    return model

# Use Dask to process and train in parallel
models = data.map(process_and_train).compute()