⚡ Bolt: Optimize balanceTranslator membership check

.jules/bolt.md

@@ -0,0 +1,3 @@
+## 2024-05-19 - Optimize `balanceTranslator` membership testing
+**Learning:** Checking membership inside a loop with a freshly-constructed list comprehension gives an O(N * M) performance impact. Pre-computing a set of keys can reduce it to O(N + M).
+**Action:** When finding a list membership test `x not in [y for y in Y]` inside a loop, elevate the inner list creation out of the loop and convert it to a `set` to reduce time complexity to O(1).

bionetgen/atomizer/writer/bnglWriter.py

@@ -116,17 +116,13 @@ def balanceTranslator(reactant, product, translator):
     for rMolecule in rMolecules:
         for pMolecule in pMolecules:
             if rMolecule.name == pMolecule.name:
+                p_names = {y.name for y in pMolecule.components}
+                r_names = {y.name for y in rMolecule.components}
                 overFlowingComponents = [
-                    x
-                    for x in rMolecule.components
-                    if x.name not in [y.name for y in pMolecule.components]
+                    x for x in rMolecule.components if x.name not in p_names
                 ]
                 overFlowingComponents.extend(
-                    [
-                        x
-                        for x in pMolecule.components
-                        if x.name not in [y.name for y in rMolecule.components]
-                    ]
+                    [x for x in pMolecule.components if x.name not in r_names]
                 )
                 rMolecule.removeComponents(overFlowingComponents)
                 pMolecule.removeComponents(overFlowingComponents)