Getting Started
===============

CorePlexML exposes a comprehensive REST API for automating every stage of
the ML lifecycle: data ingestion, model training, deployment, monitoring,
privacy compliance, and synthetic data generation.

There are two ways to interact with the API:

1. **Python SDK** -- A typed, resource-oriented client library.
   Install with ``pip install coreplexml`` and get started in minutes.

2. **REST API** -- Direct HTTP calls using ``curl``, ``requests``, or
   any HTTP library.  Every endpoint is documented with request/response
   examples in the :doc:`/api-reference/index` section.

Core Concepts
-------------

Projects
^^^^^^^^

Projects are the top-level organizational unit. Every dataset, experiment,
model, and deployment belongs to exactly one project. Think of a project
as a workspace for a single ML use case (e.g., "Customer Churn Prediction").

Datasets & Versions
^^^^^^^^^^^^^^^^^^^^

Upload CSV files to a project. CorePlexML automatically profiles the data,
detects column types, and creates an immutable **dataset version**. Each
subsequent upload creates a new version, preserving full data lineage.

Experiments
^^^^^^^^^^^

An experiment runs H2O AutoML against a dataset version and target column.
It automatically trains, tunes, and ranks dozens of candidate models
(GBM, XGBoost, Deep Learning, Stacked Ensembles, etc.). Experiments are
asynchronous background jobs -- you launch them and poll for completion.

Models
^^^^^^

Each experiment produces multiple ranked models. Inspect model metrics
(AUC, RMSE, MAE, etc.), feature importance, hyperparameters, and make
ad-hoc predictions.

Deployments (MLOps)
^^^^^^^^^^^^^^^^^^^

Deploy any model to a REST prediction endpoint. Deployments support
staging/production stages, canary rollouts, alerting, and automatic
retraining triggers.

Reports
^^^^^^^

Generate PDF reports for experiments, models, and projects with charts,
metrics tables, and optional AI-powered recommendations.

Enterprise Modules
^^^^^^^^^^^^^^^^^^

- **Privacy Suite** -- PII detection across 72+ types with automated
  masking, hashing, and redaction. HIPAA, GDPR, PCI-DSS, and CCPA
  compliance profiles.
- **SynthGen** -- Train CTGAN, CopulaGAN, or TVAE models to produce
  privacy-safe synthetic datasets that preserve statistical properties.
- **What-If Studio** -- Scenario-based model exploration. Change feature
  values and compare prediction outcomes side-by-side.

Next Steps
----------

- :doc:`authentication` -- Set up API keys for programmatic access
- :doc:`quickstart` -- Run your first ML pipeline end-to-end
- :doc:`/sdk/index` -- Install and use the Python SDK
- :doc:`/api-reference/index` -- Complete REST API reference