Correct regex parsing + binary strings dec

valence-labs · Aug 8, 2024 · dab04ef · dab04ef
1 parent a1061a8
commit dab04ef
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 6 deletions.
diff --git a/openqdc/datasets/potential/ani.py b/openqdc/datasets/potential/ani.py
@@ -154,7 +154,7 @@ def convert_forces(self, x):
         return super().convert_forces(x) * 0.529177249  # correct the Dataset error
 
     def __smiles_converter__(self, x):
-        return x
+        return "-".join(x.decode("ascii").split("-")[:-1])
 
 
 class ANI1CCX(ANI1):
@@ -195,10 +195,7 @@ class ANI1CCX(ANI1):
     __links__ = {"ani1x.hdf5.gz": "https://zenodo.org/record/4081694/files/292.hdf5.gz"}
 
     def __smiles_converter__(self, x):
-        """util function to convert string to smiles: useful if the smiles is
-        encoded in a different format than its display format
-        """
-        return x
+        return x.decode("ascii")
 
 
 class ANI1CCX_V2(ANI1CCX):

diff --git a/openqdc/datasets/potential/maceoff.py b/openqdc/datasets/potential/maceoff.py
@@ -14,7 +14,7 @@
 def parse_mace_xyz(xyzpath):
     energy_re = re.compile(r"energy=(\S+)")
     smiles_re = re.compile(r"smiles=(\S+)")
-    subset_re = re.compile(r"config_type=(\S+)")
+    subset_re = re.compile(r"config_type=([^;]+)\ MACE_energy")
     with open(xyzpath, "r") as f:
         n_atoms = None
         counter = 0