From 0c6d4d3d0bfa0955d0328fd2fc18aa0071a3369c Mon Sep 17 00:00:00 2001
From: Alain Vaucher <avaucher@users.noreply.github.com>
Date: Thu, 5 Jan 2023 11:19:23 +0100
Subject: [PATCH] Add function to get list of individual compounds from any
 SMILES string (#16)

---
 src/rxn/chemutils/miscellaneous.py | 33 +++++++++++++++++++++++++++++-
 tests/test_miscellaneous.py        | 26 +++++++++++++++++++++++
 2 files changed, 58 insertions(+), 1 deletion(-)

diff --git a/src/rxn/chemutils/miscellaneous.py b/src/rxn/chemutils/miscellaneous.py
index 2015a9a..b398f7f 100644
--- a/src/rxn/chemutils/miscellaneous.py
+++ b/src/rxn/chemutils/miscellaneous.py
@@ -19,7 +19,12 @@
     apply_to_compounds,
     sort_compounds,
 )
-from .reaction_smiles import determine_format, parse_reaction_smiles, to_reaction_smiles
+from .reaction_smiles import (
+    determine_format,
+    parse_any_reaction_smiles,
+    parse_reaction_smiles,
+    to_reaction_smiles,
+)
 
 CHIRAL_CENTER_PATTERN = re.compile(
     r"\[([^],@]+)@+([^]]*)]"
@@ -220,3 +225,29 @@ def sort_any(any_smiles: str) -> str:
     else:
         # we call the same function for single- and multi-component SMILES
         return sort_multicomponent_smiles(any_smiles)
+
+
+def get_individual_compounds(any_smiles: str) -> List[str]:
+    """
+    Get the individual compound SMILES strings starting from any SMILES string
+    (multicomponent SMILES, reaction SMILES).
+
+    Single-component SMILES with dots are interpreted as multicomponent SMILES strings.
+
+    Args:
+        any_smiles: any kind of SMILES string.
+
+    Raises:
+        Exception: different kinds of exception may be raised during parsing.
+
+    Returns:
+        List of individual compound SMILES.
+    """
+    if ">" in any_smiles:
+        # We have a reaction SMILES
+        reaction = parse_any_reaction_smiles(any_smiles)
+        return list(reaction.iter_all_smiles())
+    else:
+        # We interpret it as a multicomponent SMILES.
+        # We use "~" as a fragment bond even if it is not actually needed.
+        return multicomponent_smiles_to_list(any_smiles, fragment_bond="~")
diff --git a/tests/test_miscellaneous.py b/tests/test_miscellaneous.py
index 01706f7..2b3f60d 100644
--- a/tests/test_miscellaneous.py
+++ b/tests/test_miscellaneous.py
@@ -10,6 +10,7 @@
     atom_type_counter,
     canonicalize_any,
     equivalent_smiles,
+    get_individual_compounds,
     is_valid_smiles,
     remove_chiral_centers,
     remove_double_bond_stereochemistry,
@@ -192,3 +193,28 @@ def test_sort_any() -> None:
     # reaction SMILES
     assert sort_any("B.A.E~D.A>>C.B") == "A.A.B.E~D>>B.C"
     assert sort_any("B.A.E.D.A>>C.B |f:2.3|") == "A.A.B.E.D>>B.C |f:3.4|"
+
+
+def test_get_individual_compounds() -> None:
+    # Single-component SMILES and/or multi-component SMILES
+    assert get_individual_compounds("A.C.C.B") == ["A", "C", "C", "B"]
+    assert get_individual_compounds("A.D~C.B") == ["A", "D.C", "B"]
+    assert get_individual_compounds("CBA") == ["CBA"]
+
+    # reaction SMILES
+    assert get_individual_compounds("B.A.E~D.A>>C.B") == [
+        "B",
+        "A",
+        "E.D",
+        "A",
+        "C",
+        "B",
+    ]
+    assert get_individual_compounds("B.A.E.D.A>>C.B |f:2.3|") == [
+        "B",
+        "A",
+        "A",
+        "E.D",  # Note that it is located elsewhere, due to how extended SMILES are parsed
+        "C",
+        "B",
+    ]