QM7 QM7b QM9

valence-labs · Jun 26, 2024 · ea59515 · ea59515
1 parent 1ab09ea
commit ea59515
Show file tree

Hide file tree

Showing 2 changed files with 36 additions and 5 deletions.
diff --git a/openqdc/datasets/potential/qmx.py b/openqdc/datasets/potential/qmx.py
@@ -88,22 +88,49 @@ def preprocess_path(self):
 
     def read_raw_entries(self):
         raw_path = p_join(self.root, f"{self.__name__}.h5.gz")
-        samples = read_qc_archive_h5(raw_path, self.__name__, self.energy_target_names, self.force_target_names)
+        samples = read_qc_archive_h5(raw_path, self.__name__, self.energy_target_names, None)
         return samples
 
 
 # ['smiles', 'E1-CC2', 'E2-CC2', 'f1-CC2', 'f2-CC2', 'E1-PBE0', 'E2-PBE0', 'f1-PBE0', 'f2-PBE0', 'E1-PBE0.1', 'E2-PBE0.1', 'f1-PBE0.1', 'f2-PBE0.1', 'E1-CAM', 'E2-CAM', 'f1-CAM', 'f2-CAM']
 class QM7(QMX):
     __links__ = {"qm7.hdf5.gz": "https://zenodo.org/record/3588337/files/150.hdf5.gz?download=1"}
     __name__ = "qm7"
+
+    energy_target_names = ['B2PLYP-D3(BJ):aug-cc-pvdz', 'B2PLYP-D3(BJ):aug-cc-pvtz', 'B2PLYP-D3(BJ):def2-svp',
+                           'B2PLYP-D3(BJ):def2-tzvp', 'B2PLYP-D3(BJ):sto-3g', 'B2PLYP-D3:aug-cc-pvdz',
+                           'B2PLYP-D3:aug-cc-pvtz', 'B2PLYP-D3:def2-svp', 'B2PLYP-D3:def2-tzvp',
+                           'B2PLYP-D3:sto-3g', 'B2PLYP-D3M(BJ):aug-cc-pvdz', 'B2PLYP-D3M(BJ):aug-cc-pvtz',
+                           'B2PLYP-D3M(BJ):def2-svp', 'B2PLYP-D3M(BJ):def2-tzvp', 'B2PLYP-D3M(BJ):sto-3g',
+                           'B2PLYP-D3M:aug-cc-pvdz', 'B2PLYP-D3M:aug-cc-pvtz', 'B2PLYP-D3M:def2-svp', 
+                           'B2PLYP-D3M:def2-tzvp', 'B2PLYP-D3M:sto-3g', 'B2PLYP:aug-cc-pvdz', 
+                           'B2PLYP:aug-cc-pvtz', 'B2PLYP:def2-svp', 'B2PLYP:def2-tzvp', 
+                           'B2PLYP:sto-3g', 'B3LYP-D3(BJ):aug-cc-pvdz', 'B3LYP-D3(BJ):aug-cc-pvtz',
+                           'B3LYP-D3(BJ):def2-svp', 'B3LYP-D3(BJ):def2-tzvp', 'B3LYP-D3(BJ):sto-3g', 
+                           'B3LYP-D3:aug-cc-pvdz', 'B3LYP-D3:aug-cc-pvtz', 'B3LYP-D3:def2-svp',
+                           'B3LYP-D3:def2-tzvp', 'B3LYP-D3:sto-3g', 'B3LYP-D3M(BJ):aug-cc-pvdz',
+                           'B3LYP-D3M(BJ):aug-cc-pvtz', 'B3LYP-D3M(BJ):def2-svp', 'B3LYP-D3M(BJ):def2-tzvp',
+                           'B3LYP-D3M(BJ):sto-3g', 'B3LYP-D3M:aug-cc-pvdz', 'B3LYP-D3M:aug-cc-pvtz', 
+                           'B3LYP-D3M:def2-svp', 'B3LYP-D3M:def2-tzvp', 'B3LYP-D3M:sto-3g', 
+                           'B3LYP:aug-cc-pvdz', 'B3LYP:aug-cc-pvtz', 'B3LYP:def2-svp', 'B3LYP:def2-tzvp',
+                           'B3LYP:sto-3g', 'HF:aug-cc-pvdz', 'HF:aug-cc-pvtz', 'HF:def2-svp',
+                           'HF:def2-tzvp', 'HF:sto-3g', 'MP2:aug-cc-pvdz', 'MP2:aug-cc-pvtz',
+                           'MP2:def2-svp', 'MP2:def2-tzvp', 'MP2:sto-3g', 'PBE0:aug-cc-pvdz', 
+                           'PBE0:aug-cc-pvtz', 'PBE0:def2-svp', 'PBE0:def2-tzvp', 'PBE0:sto-3g', 
+                           'PBE:aug-cc-pvdz', 'PBE:aug-cc-pvtz', 'PBE:def2-svp', 'PBE:def2-tzvp',
+                           'PBE:sto-3g', 'WB97M-V:aug-cc-pvdz', 'WB97M-V:aug-cc-pvtz', 'WB97M-V:def2-svp', 
+                           'WB97M-V:def2-tzvp', 'WB97M-V:sto-3g', 'WB97X-D:aug-cc-pvdz', 
+                           'WB97X-D:aug-cc-pvtz', 'WB97X-D:def2-svp', 'WB97X-D:def2-tzvp', 
+                           'WB97X-D:sto-3g']
+
 
-    def read_raw_entries(self):
-        "h5.gz"
 
 
 class QM7b(QMX):
     __links__ = {"qm7b.hdf5.gz": "https://zenodo.org/record/3588335/files/200.hdf5.gz?download=1"}
     __name__ = "qm7b"
+    energy_target_names = ['CCSD(T0):cc-pVDZ', 'HF:cc-pVDZ', 'HF:cc-pVTZ', 'MP2:cc-pVTZ', 
+                            'B2PLYP-D3:aug-cc-pvdz', 'B2PLYP-D3:aug-cc-pvtz', 'B2PLYP-D3:def2-svp', 'B2PLYP-D3:def2-tzvp', 'B2PLYP-D3:sto-3g', 'B2PLYP-D3M(BJ):aug-cc-pvdz', 'B2PLYP-D3M(BJ):aug-cc-pvtz', 'B2PLYP-D3M(BJ):def2-svp', 'B2PLYP-D3M(BJ):def2-tzvp', 'B2PLYP-D3M(BJ):sto-3g', 'B2PLYP-D3M:aug-cc-pvdz', 'B2PLYP-D3M:aug-cc-pvtz', 'B2PLYP-D3M:def2-svp', 'B2PLYP-D3M:def2-tzvp', 'B2PLYP-D3M:sto-3g', 'B2PLYP:aug-cc-pvdz', 'B2PLYP:aug-cc-pvtz', 'B2PLYP:def2-svp', 'B2PLYP:def2-tzvp', 'B2PLYP:sto-3g', 'B3LYP-D3(BJ):aug-cc-pvdz', 'B3LYP-D3(BJ):aug-cc-pvtz', 'B3LYP-D3(BJ):def2-svp', 'B3LYP-D3(BJ):def2-tzvp', 'B3LYP-D3(BJ):sto-3g', 'B3LYP-D3:aug-cc-pvdz', 'B3LYP-D3:aug-cc-pvtz', 'B3LYP-D3:def2-svp', 'B3LYP-D3:def2-tzvp', 'B3LYP-D3:sto-3g', 'B3LYP-D3M(BJ):aug-cc-pvdz', 'B3LYP-D3M(BJ):aug-cc-pvtz', 'B3LYP-D3M(BJ):def2-svp', 'B3LYP-D3M(BJ):def2-tzvp', 'B3LYP-D3M(BJ):sto-3g', 'B3LYP-D3M:aug-cc-pvdz', 'B3LYP-D3M:aug-cc-pvtz', 'B3LYP-D3M:def2-svp', 'B3LYP-D3M:def2-tzvp', 'B3LYP-D3M:sto-3g', 'B3LYP:aug-cc-pvdz', 'B3LYP:aug-cc-pvtz', 'B3LYP:def2-svp', 'B3LYP:def2-tzvp', 'B3LYP:sto-3g', 'HF:aug-cc-pvdz', 'HF:aug-cc-pvtz', 'HF:cc-pvtz', 'HF:def2-svp', 'HF:def2-tzvp', 'HF:sto-3g', 'PBE0:aug-cc-pvdz', 'PBE0:aug-cc-pvtz', 'PBE0:def2-svp', 'PBE0:def2-tzvp', 'PBE0:sto-3g', 'PBE:aug-cc-pvdz', 'PBE:aug-cc-pvtz', 'PBE:def2-svp', 'PBE:def2-tzvp', 'PBE:sto-3g', 'SVWN:sto-3g', 'WB97M-V:aug-cc-pvdz', 'WB97M-V:aug-cc-pvtz', 'WB97M-V:def2-svp', 'WB97M-V:def2-tzvp', 'WB97M-V:sto-3g', 'WB97X-D:aug-cc-pvdz', 'WB97X-D:aug-cc-pvtz', 'WB97X-D:def2-svp', 'WB97X-D:def2-tzvp', 'WB97X-D:sto-3g']
 
 
 class QM8(QMX):
@@ -161,3 +188,7 @@ def read_raw_entries(self):
 class QM9(QMX):
     __links__ = {"qm9.hdf5.gz": "https://zenodo.org/record/3588339/files/155.hdf5.gz?download=1"}
     __name__ = "qm9"
+    energy_target_names = [
+        'Internal energy at 0 K',
+        'B3LYP:def2-svp', 'HF:cc-pvtz', 'HF:sto-3g', 'PBE:sto-3g', 'SVWN:sto-3g', 'WB97X-D:aug-cc-pvtz', 'WB97X-D:def2-svp', 'WB97X-D:def2-tzvp']
+
diff --git a/openqdc/utils/io.py b/openqdc/utils/io.py
@@ -150,8 +150,8 @@ def load_hdf5_file(hdf5_file_path: str):
 
     # inorder to enable multiprocessing:
     # https://github.com/fsspec/gcsfs/issues/379#issuecomment-839929801
-    fsspec.asyn.iothread[0] = None
-    fsspec.asyn.loop[0] = None
+    # fsspec.asyn.iothread[0] = None
+    # fsspec.asyn.loop[0] = None
 
     return file