merge: pull request #6 from brsynth/GetCids

chore(rp2erxn): handle different cmp IDs

rp2paths/rp2erxn.py

@@ -10,6 +10,7 @@
 import sys
 import csv
 import argparse
+from re import match as re_match
 from rdkit import Chem
 from rp2paths.IDsHandler import IDsHandler
 
@@ -86,7 +87,10 @@ def SortCids(cids):
                 break
         # Sort IDs
         if mnx_case:
-            return sorted(cids, key=lambda x: int(x[4:]))
+            try:
+                return sorted(cids, key=lambda x: int(re_match(r'MNXM(\d+)\w*', x).group(1)))
+            except ValueError:
+                return sorted(cids)
         else:
             return sorted(cids)
 

tests/_main.py

@@ -6,10 +6,11 @@
 
 from unittest import TestCase
 
-from pathlib import Path
 from os import stat as os_stat
 from os import path as os_path
 
+from rp2paths.rp2erxn import Compound
+
 
 class Main(TestCase):
     __test__ = False

tests/test_rp2erxn.py

@@ -0,0 +1,36 @@
+"""
+Created on Oct 16 2023
+
+@author: Joan Hérisson
+"""
+
+from unittest import TestCase
+
+from os import stat as os_stat
+from os import path as os_path
+
+from rp2paths.rp2erxn import Compound
+
+
+class Test_rp2erxn(TestCase):
+
+    def test_SortCidsMNXM_1(self):
+        mnx_cids = ['MNXM03', 'MNXM02', 'MNXM5', 'MNXM1']
+        self.assertListEqual(
+            Compound.SortCids(mnx_cids),
+            ['MNXM1', 'MNXM02', 'MNXM03', 'MNXM5']
+        )
+
+    def test_SortCidsMNXM_2(self):
+        mnx_cids = ['MNXM03', 'MNXM02', 'MNXM5__64__MNXC3', 'MNXM1']
+        self.assertListEqual(
+            Compound.SortCids(mnx_cids),
+            ['MNXM1', 'MNXM02', 'MNXM03', 'MNXM5__64__MNXC3']
+        )
+
+    def test_SortCidsNotMNXM(self):
+        mnx_cids = ['M_2agpg181_c', 'M_2agpg161_c', 'M_2ahbut_c', 'M_2agpg180_c', 'MNXM1']
+        self.assertListEqual(
+            Compound.SortCids(mnx_cids),
+            ['MNXM1', 'M_2agpg161_c', 'M_2agpg180_c', 'M_2agpg181_c', 'M_2ahbut_c']
+        )