oasislmf.pytools.fm.portfolio_complexity
========================================

.. py:module:: oasislmf.pytools.fm.portfolio_complexity


Functions
---------

.. autoapisummary::

   oasislmf.pytools.fm.portfolio_complexity.compute_portfolio_complexity
   oasislmf.pytools.fm.portfolio_complexity.format_complexity_report


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

.. py:function:: compute_portfolio_complexity(static_path)

   Compute portfolio complexity metrics from Oasis static files.

   Works for GUL-only, IL, and IL+RI portfolios.

   **GUL-only** (no account file):  only ``portfolio_dimensions`` and the
   ``gul_work_units`` complexity score are populated.

   **IL** (account file present):  additionally populates ``il`` with the FM
   structure, profile complexity, and output count for the insured-loss layer.

   **RI** (reinsurance files present):  additionally populates ``ri`` with
   per-layer FM metrics and a total RI FM work-unit count.

   The two key per-event cost proxies in ``complexity_scores`` are:

   * ``gul_work_units = num_items``
   * ``fm_work_units = il_fm_work_units + ri_fm_work_units``
     where ``il_fm_work_units = total_il_nodes × il_layers`` and
     ``ri_fm_work_units = Σ(total_ri_nodes_layer_k × ri_layers_k)``

   Multiplying by the number of model events gives a per-dimension
   full-run cost proxy.

   Parameters
   ----------
   static_path : str
       Directory containing the Oasis static files.  For IL this is the
       directory with ``fm_programme.bin`` etc.  RI layer subdirectories
       are discovered automatically via ``ri_layers.json``.

   Returns
   -------
   dict


.. py:function:: format_complexity_report(metrics)

   Return a human-readable text summary of portfolio complexity metrics.

   Parameters
   ----------
   metrics : dict
       As returned by :func:`compute_portfolio_complexity`.

   Returns
   -------
   str