`src.inference`¶

Inference module. CLI tool that receives a CSV file as input, runs predictions using the saved best model, and outputs a CSV with PassengerId and Survived columns.

Usage¶

python src/inference.py --input <input_csv> --output <output_csv>

Example:

python src/inference.py --input data/raw/test.csv --output reports/predictions.csv

CLI Arguments¶

Argument	Short	Required	Description
`--input`	`-i`	Yes	Path to input CSV file (must contain `PassengerId` column)
`--output`	`-o`	Yes	Path to output CSV file

Functions¶

`prepare_catboost_frame`¶

def prepare_catboost_frame(
    frame: pd.DataFrame,
    categorical_features: list[str]
) -> pd.DataFrame

Prepares a DataFrame for CatBoost by filling NaN values in categorical columns with "missing" and casting them to string type.

Parameters:

Parameter	Type	Description
`frame`	`pd.DataFrame`	Input DataFrame
`categorical_features`	`list[str]`	Names of categorical columns to prepare

Returns: A copy of the DataFrame with categorical columns cleaned.

`run_inference`¶

def run_inference(input_csv: str, output_csv: str) -> None

Runs inference on a given CSV file using the saved best model.

Parameters:

Parameter	Type	Description
`input_csv`	`str`	Path to the input CSV file
`output_csv`	`str`	Path to save the predictions CSV

Steps:

Load the model bundle from models/best_model.pkl
Load the preprocessing pipeline bundle (for categorical_features)
Read the input CSV
Extract PassengerId column
Drop columns listed in preprocessing.dropped_columns
Prepare data for CatBoost if needed
Generate integer predictions (0 or 1)
Save output CSV with PassengerId and Survived

Output CSV Format:

Column	Type	Description
`PassengerId`	int	Passenger identifier from input
`Survived`	int	Predicted survival (0 or 1)

Raises:

Exception	Condition
`FileNotFoundError`	If the model pickle or input CSV does not exist
`ValueError`	If the input CSV does not contain a `PassengerId` column

`main`¶

def main() -> None

CLI entry point. Parses --input and --output arguments using argparse, then calls run_inference().

src.inference¶