odvm 0.1.0

Open, Dynamic, and Versatile Modeling AutoML system

O D V M

From Raw Data to Intelligent Insights

Where data clarity meets decision power through automated intelligence

Overview

ODVM is a modular AutoML pipeline that automates the entire data science workflow:

Exploratory Data Analysis (EDA) → Smart Preprocessing → Model Selection & Training → Reporting

Multi-backend support (Pandas, Dask) and extensible architecture accelerate your ML projects while maintaining transparency.

---

Documentation Summary

Module / Class	Method / Function	Description (Humanized)	Sample Code Example
core.runner.ODVM	__init__	Initialize the pipeline with data, target column, and config settings.	python\nodvm = ODVM(data="data.csv", target="price", config=config)\n
	_load_data	Load data from file path or DataFrame.	python\ndf = odvm._load_data("data.csv")\n
	_validate	Check that the target column exists in the dataset.	python\nodvm._validate()\n
	run	Run the full pipeline with options for each stage (EDA, preprocessing, modeling, reporting, deployment).	python\nodvm.run(eda=True, preprocess=True, model=True)\n
	run_eda	Perform exploratory data analysis with charts and stats.	python\nodvm.run_eda()\n
	run_preprocessing	Clean, encode, split, and scale the data preparing for modeling.	python\nodvm.run_preprocessing()\n
	run_modeling	Select, tune, train, and evaluate models automatically.	python\nodvm.run_modeling()\n
	generate_report	(Work in progress) Generate detailed performance reports.	python\nodvm.generate_report()\n
	deploy_model	(Work in progress) Deploy model via REST API or dashboard.	python\nodvm.deploy_model()\n
	save_best_model	Save the best trained model to a pickle file for later use.	python\nodvm.save_best_model(results, models_dict)\n
assistant.task_detector.TaskDetector	detect	Automatically detect task type: classification, regression, clustering, etc.	python\ntask_info = TaskDetector(df, target="price").detect()\n
eda.analyzer.EDAAnalyzer	get_shape	Get the shape of the dataset (rows and columns).	python\nrows, cols = analyzer.get_shape()\n
	get_columns_info	Get information about columns and their data types.	python\ninfo = analyzer.get_columns_info()\n
	missing_values	Analyze and report missing values in each column.	python\nmissings = analyzer.missing_values()\n
	save_summary	Save EDA summary statistics to a JSON file.	python\nanalyzer.save_summary("eda_summary.json")\n
eda.visualizer.EDAVisualizer	plot_distributions	Plot distributions for numerical and categorical features.	python\nvisualizer.plot_distributions()\n
	plot_boxplots	Plot boxplots to check outliers and spread.	python\nvisualizer.plot_boxplots()\n
	plot_correlation	Plot correlation heatmap for numerical features.	python\nvisualizer.plot_correlation()\n
	plot_target_distribution	Plot distribution of the target variable.	python\nvisualizer.plot_target_distribution()\n
	plot_pairplot	Plot pairwise scatterplots for features.	python\nvisualizer.plot_pairplot()\n
preprocess.cleaner.DataCleaner	clean	Clean data by handling missing values, duplicates, and outliers.	python\ncleaned_df = cleaner.clean()\n
preprocess.encoder.Encoder	encode	Encode categorical/text features to numerical form.	python\nencoded_df = encoder.encode()\n
preprocess.scaler.Scaler	fit_transform	Fit scaler on training data and transform it.	python\nX_train_scaled = scaler.fit_transform(X_train)\n
	transform	Transform validation/test data using fitted scaler.	python\nX_test_scaled = scaler.transform(X_test)\n
preprocess.splitter.DataSplitter	split	Split dataset into train, validation, and test sets.	python\nX_train, X_val, X_test, y_train, y_val, y_test = splitter.split()\n
modeling.model_selector.ModelSelector	get_models	Get a dictionary of models suitable for the detected task.	python\nmodels = selector.get_models()\n
modeling.tuner.ModelTuner	tune	Tune model hyperparameters using grid or random search.	python\nbest_model = tuner.tune()\n
modeling.trainer.ModelTrainer	train_all	Train all selected (and tuned) models.	python\ntrained_models, params = trainer.train_all()\n
modeling.evaluator.ModelEvaluator	evaluate	Evaluate models on test data and return performance metrics.	python\nresults = evaluator.evaluate()\n
reporting.report_builder.ReportBuilder	build	Build performance reports in HTML or Excel format.	python\nreport.build()\n

---

Quick Start Example

from core.runner import ODVM

config = {
    "eda": {"visualize": True, "save_summary": True},
    "preprocessing": {"missing_strategy": "mean", "encoding": "label", "scaling": "standard"},
    "split": {"test_size": 0.2, "val_size": 0.1, "random_state": 42},
    "modeling": {"allowed_models": ["LinearRegression"], "tuning": True, "cv": 3, "scoring": "r2"}
}

odvm = ODVM(data="data/housing.csv", target="median_house_value", config=config)
odvm.run(eda=True, preprocess=True, model=True)

Contributing

We welcome all contributions, big or small!

To get started:

1. Fork the repository
2. Create a new branch
3. Commit your changes
4. Open a Pull Request

---

Contact