Validate Config Command
=======================

The ``validate-config`` command validates a configuration file without running the pipeline. This is useful for catching configuration errors before starting a potentially long training run.

Usage
-----

.. code-block:: bash

    respredai validate-config <path_to_config.ini> [--check-data]

Options
-------

Required
~~~~~~~~

- ``config`` - Path to the configuration file (INI format)

  - Positional argument (no flag needed)

Optional
~~~~~~~~

- ``--check-data, -d`` - Also validate that the data file exists and can be loaded

  - Checks that ``data_path`` points to a valid CSV file
  - Verifies all specified targets and features exist in the data
  - Reports dataset dimensions and group information

What Gets Validated
-------------------

Without ``--check-data``
~~~~~~~~~~~~~~~~~~~~~~~~

1. **File existence** - Config file exists and is readable
2. **File format** - Config file has ``.ini`` extension
3. **Syntax** - Valid INI format with all required sections
4. **Required parameters** - All mandatory parameters are present
5. **Parameter values** - Values are valid (e.g., valid model names, threshold methods)

With ``--check-data``
~~~~~~~~~~~~~~~~~~~~~

All of the above, plus:

6. **Data file existence** - CSV file specified in ``data_path`` exists
7. **Data loading** - CSV can be parsed without errors
8. **Column validation** - Target columns exist in the data
9. **Missing values** - Data doesn't contain unexpected missing values
10. **Group validation** - Group column exists (if specified)

See Also
--------

- :doc:`run-command` - Execute the full pipeline with nested CV
- :doc:`train-command` - Train models on entire dataset for cross-dataset validation
- :doc:`create-config-command` - Generate template configs