Merge pull request #97 from firefly-cpp/84-cli

🧑‍💻added `niaaml` CLI

README.md

@@ -33,6 +33,7 @@
 <p align="center">
   <a href="https://github.com/firefly-cpp/NiaAML?tab=readme-ov-file#-installation">📦 Installation</a> •
   <a href="https://github.com/firefly-cpp/NiaAML?tab=readme-ov-file#-graphical-user-interface">💻 Graphical User Interface</a> •
+  <a href="https://github.com/firefly-cpp/NiaAML?tab=readme-ov-file#-command-line-interface">🧑‍💻 Command Line Interface</a> •
   <a href="https://github.com/firefly-cpp/NiaAML?tab=readme-ov-file#-api">📮 API</a> •
   <a href="https://github.com/firefly-cpp/NiaAML?tab=readme-ov-file#-implemented-components">✨ Implemented Components</a> •
   <a href="https://github.com/firefly-cpp/NiaAML?tab=readme-ov-file#-optimization-process-and-parameter-tuning">💪 Optimization Process And Parameter Tuning</a> •
@@ -117,6 +118,22 @@ $ nix-shell -p python311Packages.niaaml
 
 There is a simple Graphical User Interface for the NiaAML package available [here](https://github.com/lukapecnik/NiaAML-GUI).
 
+## 🧑‍💻 Command Line Interface
+
+We also provide a CLI for quick pipeline optimizations and inference from the terminal without the need to write custom scripts.
+
+When you install the package as instructed above, you will already have access to the `niaaml` command with sub-commands `optimize` and `infer`
+
+For usage information, add the `--help` flag:
+
+![niaaml help](.github/images/niaaml_cli_help.png)
+
+![niaaml infer help](.github/images/niaaml_cli_infer_help.png)
+
+An example Invocation of `optimize`:
+
+![niaaml optimize example](.github/images/niaaml_cli_optimize_example.png)
+
 ## 📮 API
 
 There is a simple API for remote work with NiaAML package available [here](https://github.com/alenrajsp/NiaAML-API).

niaaml/__init__.py

@@ -11,6 +11,7 @@
 from niaaml.pipeline import Pipeline
 from niaaml.pipeline_component import PipelineComponent
 from niaaml.logger import Logger
+from niaaml import cli
 
 __all__ = [
     "classifiers",
@@ -26,6 +27,7 @@
     "Pipeline",
     "PipelineComponent",
     "Logger",
+    "cli",
 ]
 
 __project__ = "niaaml"

niaaml/cli.py

@@ -0,0 +1,104 @@
+"""🧑‍💻 command line interface for NiaAML"""
+
+from pathlib import Path
+from typing import Optional
+
+from loguru import logger
+import pandas as pd
+import typer
+from typing_extensions import Annotated
+
+from niaaml import PipelineOptimizer, Pipeline
+from niaaml.data.csv_data_reader import CSVDataReader
+
+
+app = typer.Typer(
+    help="🌳 a command line interface for NiaAML.",
+    no_args_is_help=True
+)
+
+
+@app.command()
+def optimize(
+    data_csv_file: Path,
+    has_header: bool = True,
+    ignore_columns: list[int] = [],
+    classifiers: list[str] = ['AdaBoost', 'Bagging', 'MultiLayerPerceptron', 'RandomForest', 'ExtremelyRandomizedTrees', 'LinearSVC'],
+    feature_selection_algorithms: list[str] = ['SelectKBest', 'SelectPercentile', 'ParticleSwarmOptimization', 'VarianceThreshold'],
+    feature_transform_algorithms: list[str] = ['Normalizer', 'StandardScaler'],
+    categorical_features_encoder: Annotated[Optional[str], typer.Option()] = "OneHotEncoder",
+    imputer: Annotated[Optional[str], typer.Option()] = None,
+    fitness_name: str = 'Accuracy',
+    pipeline_population_size: int = 15,
+    inner_population_size: int = 15,
+    number_of_pipeline_evaluations: int = 100,
+    number_of_inner_evaluations: int = 100,
+    optimization_algorithm: str = 'ParticleSwarmAlgorithm',
+    inner_optimization_algorithm: Annotated[Optional[str], typer.Option()] = None,
+    result_file: Path = Path("pipeline.ppln"),
+) -> None:
+    """🦾 optimizes a NiaAML pipeline on a given dataset."""
+    # 📄 load and setup data
+    logger.info(f"📄 reading `{data_csv_file}`")
+    data_reader = CSVDataReader(
+        src=str(data_csv_file),
+        has_header=has_header,
+        contains_classes=True,
+        ignore_columns=ignore_columns
+    )
+
+    # 🦾 setup pipeline
+    logger.info("🦾 start the optimization process ...")
+    pipeline_optimizer = PipelineOptimizer(
+        data=data_reader,
+        classifiers=classifiers,
+        feature_selection_algorithms=feature_selection_algorithms,
+        feature_transform_algorithms=feature_transform_algorithms,
+        categorical_features_encoder=categorical_features_encoder,
+        imputer=imputer,
+    )
+
+    # 📈 optimize pipeline
+    pipeline = pipeline_optimizer.run(fitness_name, pipeline_population_size, inner_population_size, number_of_pipeline_evaluations, number_of_inner_evaluations, optimization_algorithm, inner_optimization_algorithm)
+
+    # 💾 save pipeline
+    logger.success(f"💾 saving optimized pipeline to `{result_file}`")
+    pipeline.export(result_file)
+
+@app.command()
+def infer(
+    data_csv_file: Path,
+    has_header: bool = True,
+    ignore_columns: list[int] = [],
+    pipeline_file: Path = Path("pipeline.ppln"),
+    predictions_csv_file: Path = Path("preds.csv"),
+) -> None:
+    """🔮 use an optimized NiaAML pipeline to make predictions."""
+    # 💾 load pipeline
+    pipeline = Pipeline.load(pipeline_file)
+
+    # 📄 load and setup data
+    logger.info(f"📄 reading `{data_csv_file}`")
+    reader = CSVDataReader(
+        src=str(data_csv_file),
+        has_header=has_header,
+        contains_classes=True,
+        ignore_columns=ignore_columns
+    )
+    reader._read_data()
+    x: pd.DataFrame = reader._x
+
+    # 🔮 make predictions
+    logger.info(f"🔮 using `{pipeline_file}` to make predictions on the data")
+    x['preds'] = pipeline.run(x)
+
+    # 💾 save predictions
+    logger.success(f"💾 saving predictions to `{predictions_csv_file}`")
+    x.to_csv(predictions_csv_file)
+
+def main():
+    """🚪 typer entry point for the CLI."""
+    app()
+
+if __name__ == "__main__":
+    main()

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "NiaAML"
-version = "2.0.0"
+version = "2.1.0"
 description = "Python automated machine learning framework"
 license = "MIT"
 authors = ["Luka Pečnik <[email protected]>", "Iztok Fister Jr. <[email protected]>", "Laurenz Farthofer <[email protected]>"]
@@ -16,12 +16,17 @@ include = [
     { path="COMPONENTS.md", format="sdist" }
 ]
 
+[tool.poetry.scripts]
+niaaml = "niaaml.cli:main"
+
 [tool.poetry.dependencies]
 python = "^3.9"
 numpy = "^1.19.1"
 scikit-learn = "^1.1.2"
 niapy = "^2.0.5"
 pandas = "^2.1.1"
+typer = "^0.12.3"
+loguru = "^0.7.2"
 
 [tool.poetry.dev-dependencies]
 sphinx = "^3.3.1"