.. # split from OEMedChemFunctions.rst:135


.. oe:namespace:: OEMedChem

.. meta::
   :keywords: OECreateIDStrings

.. _OEMedChem_OECreateIDStrings:

.. oe:function:: OECreateIDStrings

   .. oe:language:: c++ java csharp

      .. oe:api::

         OESystem::OEIterBase<const std::string> * OECreateIDStrings(const OEChem::OEMolBase& mol,
                                                                     const std::string& atomTag = "")

   .. oe:language:: python

      .. oe:api:: python

         OECreateIDStrings(mol: Union[OEGraphMol,OEMol,OEQMol], arg2: str) -> Iterable[str]

   This pair of functions exposes the graph-edit similarity used in clustering BROOD hitlists.
   Because this is a naturally O(N^2), the API is designed for efficiency using an asymmetric
   approach.  For any 'accepted' molecule, the :oe:func:`OEMedChem::OECreateIDStrings` function is used to generate
   a set of all the canonical SMILES of fragments which would be considered similar to the input
   molecule (this can be stored for fast lookup).  If one wants to determine if a new molecule is
   similar to the original molecule, :oe:func:`OEMedChem::OECreateIDString` is called, and the single string
   generated is compared to all the strings previously created.  If the new string is an exact
   match to any of the multiple strings created from the prior molecule, then the two molecules
   are deemed similar.

   The optional ``atomTag`` in the :oe:func:`OEMedChem::OECreateIDStrings` function enables working
   with a subset of atoms in the molecule. This is useful in reducing the scope of the function when
   clustering BROOD hitlists for multi-query replacements.


