more efficient BitStrToIntList (#111)

* more efficient BitStrToIntList

* add SureChEMBL in the README

FPSim2/io/chem.py

@@ -1,4 +1,4 @@
-from typing import Any, Callable, Iterable as IterableType, Dict, List, Tuple, Union
+from typing import Any, Callable, Iterable as IterableType, Dict, Tuple, Union
 from FPSim2.FPSim2lib.utils import BitStrToIntList, PyPopcount
 from collections.abc import Iterable
 from rdkit.Chem import rdMolDescriptors

FPSim2/src/utils.cpp

@@ -23,8 +23,13 @@ namespace utils {
 
     py::list BitStrToIntList(const std::string &bit_string) {
         py::list efp;
-        for (size_t i = 0; i < bit_string.length(); i += 64) {
-            efp.append(std::stoull(bit_string.substr(i, 64), 0, 2));
+        size_t len = bit_string.length();
+        for (size_t i = 0; i < len; i += 64) {
+            uint64_t value = 0;
+            for (size_t j = 0; j < 64 && (i + j) < len; ++j) {
+                value = (value << 1) | (bit_string[i + j] - '0');
+            }
+            efp.append(value);
         }
         return efp;
     }

README.md

@@ -7,7 +7,7 @@
 
 # FPSim2: Simple package for fast molecular similarity searches
 
-FPSim2 is a small NumPy centric Python/C++ RDKit based package to run fast compound similarity searches. FPSim2 performs better with high search thresholds (>=0.7). Currently used in the [ChEMBL](http://www.ebi.ac.uk/chembl/) interface.
+FPSim2 is a small NumPy centric Python/C++ RDKit based package to run fast compound similarity searches. FPSim2 performs better with high search thresholds (>=0.7). Currently used in the [ChEMBL](http://www.ebi.ac.uk/chembl/) and [SureChEMBL](https://www.surechembl.org/) interfaces.
 
 Highlights:
 - Using CPU POPCNT instruction