tud_lbm.io.analysis.accelerations.regime_classification
=======================================================

.. py:module:: tud_lbm.io.analysis.accelerations.regime_classification

.. autoapi-nested-parse::

   Classify a droplet run into pinning / viscous / inertial / unknown.

   Built on top of :mod:`tud_lbm.io.analysis.accelerations.acceleration_analysis`:
   pinning is decided directly from the centroid excursion, while
   viscous/inertial/unknown are decided from the sign of the Ca(t) slope within
   the acceleration-derived slope window.



Classes
-------

.. autoapisummary::

   tud_lbm.io.analysis.accelerations.regime_classification.Regime
   tud_lbm.io.analysis.accelerations.regime_classification.RegimeResult


Functions
---------

.. autoapisummary::

   tud_lbm.io.analysis.accelerations.regime_classification.is_pinned
   tud_lbm.io.analysis.accelerations.regime_classification.classify_regime


Module Contents
---------------

.. py:class:: Regime

   Bases: :py:obj:`enum.StrEnum`


   Centralized regime labels used by acceleration-based classification.


   .. py:attribute:: PINNING
      :value: 'Pinning'



   .. py:attribute:: DISSIPATIVE
      :value: 'Dissipative'



   .. py:attribute:: INERTIAL
      :value: 'Inertial'



   .. py:attribute:: UNKNOWN
      :value: 'unknown'



.. py:class:: RegimeResult

   Bases: :py:obj:`NamedTuple`


   Outcome of regime classification for one run.


   .. py:attribute:: regime
      :type:  Regime


   .. py:attribute:: slope
      :type:  float | None


   .. py:attribute:: is_pinned
      :type:  bool


   .. py:attribute:: window
      :type:  tuple[int, int] | None


.. py:function:: is_pinned(cm_x: numpy.ndarray, r_zero: float) -> bool

   True when the centroid excursion is below half the initial radius.


.. py:function:: classify_regime(cm_x: numpy.ndarray, r_zero: float, accel_result: tud_lbm.io.analysis.accelerations.acceleration_analysis.AccelerationResult) -> RegimeResult

   Classify a run as pinning, viscous, inertial, or unknown.

   Pinning takes priority and does not require acceleration data. Otherwise,
   the slope of ``Ca`` vs iteration is fit within the acceleration-derived
   slope window: negative slope is viscous, non-negative is inertial. No
   usable window (no clean peak pair, or fewer than two points) is unknown.