[Bug] AttributeError: can't set attribute

[bhadreshpsavani]

In the retrosynthesis tutorial

this code of synthon completion

synthon_optimizer = torch.optim.Adam(synthon_task.parameters(), lr=1e-3)
synthon_solver = core.Engine(synthon_task, synthon_train, synthon_valid,
                             synthon_test, synthon_optimizer,
                             gpus=[0], batch_size=128)
synthon_solver.train(num_epoch=1)
synthon_solver.evaluate("valid")
synthon_solver.save("g2gs_synthon_model.pth")

gives below error

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Input In [13], in <cell line: 11>()
      7 synthon_optimizer = torch.optim.Adam(synthon_task.parameters(), lr=1e-3)
      8 synthon_solver = core.Engine(synthon_task, synthon_train, synthon_valid,
      9                              synthon_test, synthon_optimizer,
     10                              gpus=[0], batch_size=128)
---> 11 synthon_solver.train(num_epoch=1)
     12 synthon_solver.evaluate("valid")
     13 synthon_solver.save("g2gs_synthon_model.pth")

File /usr/local/lib/python3.9/dist-packages/torchdrug/core/engine.py:155, in Engine.train(self, num_epoch, batch_per_epoch)
    152 if self.device.type == "cuda":
    153     batch = utils.cuda(batch, device=self.device)
--> 155 loss, metric = model(batch)
    156 if not loss.requires_grad:
    157     raise RuntimeError("Loss doesn't require grad. Did you define any loss in the task?")

File /usr/local/lib/python3.9/dist-packages/torch/nn/modules/module.py:1130, in Module._call_impl(self, *input, **kwargs)
   1126 # If we don't have any hooks, we want to skip the rest of the logic in
   1127 # this function, and just call forward.
   1128 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
   1129         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1130     return forward_call(*input, **kwargs)
   1131 # Do not call functions when jit is used
   1132 full_backward_hooks, non_full_backward_hooks = [], []

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:592, in SynthonCompletion.forward(self, batch)
    589 all_loss = torch.tensor(0, dtype=torch.float32, device=self.device)
    590 metric = {}
--> 592 pred, target = self.predict_and_target(batch, all_loss, metric)
    593 node_in_pred, node_out_pred, bond_pred, stop_pred = pred
    594 node_in_target, node_out_target, bond_target, stop_target, size = target

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:984, in SynthonCompletion.predict_and_target(self, batch, all_loss, metric)
    981 with reactant.graph():
    982     reactant.reaction = batch["reaction"]
--> 984 graph1, node_in_target1, node_out_target1, bond_target1, stop_target1 = self.all_edge(reactant, synthon)
    985 graph2, node_in_target2, node_out_target2, bond_target2, stop_target2 = self.all_stop(reactant, synthon)
    987 graph = self._cat([graph1, graph2])

File /usr/local/lib/python3.9/dist-packages/torch/autograd/grad_mode.py:27, in _DecoratorContextManager.__call__.<locals>.decorate_context(*args, **kwargs)
     24 @functools.wraps(func)
     25 def decorate_context(*args, **kwargs):
     26     with self.clone():
---> 27         return func(*args, **kwargs)

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:557, in SynthonCompletion.all_edge(self, reactant, synthon)
    555 assert (graph.num_edges % 2 == 0).all()
    556 # node / edge features may change because we mask some nodes / edges
--> 557 graph, feature_valid = self._update_molecule_feature(graph)
    559 return graph[feature_valid], node_in_target[feature_valid], node_out_target[feature_valid], \
    560        bond_target[feature_valid], stop_target[feature_valid]

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:398, in SynthonCompletion._update_molecule_feature(self, graphs)
    395 bond_type[edge_mask] = new_graphs.bond_type.to(device=graphs.device)
    397 with graphs.node():
--> 398     graphs.node_feature = node_feature
    399 with graphs.edge():
    400     graphs.edge_feature = edge_feature

File /usr/local/lib/python3.9/dist-packages/torchdrug/data/graph.py:160, in Graph.__setattr__(self, key, value)
    158 if hasattr(self, "meta_dict"):
    159     self._check_attribute(key, value)
--> 160 super(Graph, self).__setattr__(key, value)

File /usr/local/lib/python3.9/dist-packages/torchdrug/core/core.py:84, in _MetaContainer.__setattr__(self, key, value)
     82     if types:
     83         self.meta_dict[key] = types.copy()
---> 84 self._setattr(key, value)

File /usr/local/lib/python3.9/dist-packages/torchdrug/core/core.py:93, in _MetaContainer._setattr(self, key, value)
     92 def _setattr(self, key, value):
---> 93     return super(_MetaContainer, self).__setattr__(key, value)

AttributeError: can't set attribute

[FanmengWang]

hi, I have the same problem. May I ask if you have solved this problem?

[bhadreshpsavani]

Hi @FanmengWang, I couldn't figureout yet!

[DimGorr]

Seems like I finally solved the issue. However, now I got another issue The problem is with setting an attribute in retrosynthesis.py: graphs.node_feature=node_feature because in the Molecule.py property was used:

@property
def node_feature(self):
    return self.atom_feature

So we need to add this:

@node_feature.setter
def node_feature(self, value):
    self._node_feature = value

Then attributes are okay. However, now I got another issue: rdkit.Chem.rdchem.KekulizeException: Can't kekulize mol. Unkekulized atoms: 20. And it doesn't seem like the reason is in adding these strings of code. You might have a different number of these molecules as I've shortened the dataset so it's a bit different. Let me know if you have the same issue please

[bhadreshpsavani]

Hi @DimGorr, I tried this code changes and ended up with same exception

08:53:36   >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
08:53:36   Epoch 0 begin
08:53:37   >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
08:53:37   bond accuracy: 0.690909
08:53:37   bond ce loss: 0.947608
08:53:37   node in accuracy: 0.030303
08:53:37   node in ce loss: 2.71172
08:53:37   node out accuracy: 0
08:53:37   node out ce loss: 3.50957
08:53:37   stop accuracy: 0.546075
08:53:37   stop bce loss: 0.718129
08:53:37   total accuracy: 0.0170648
08:53:37   total loss: 7.88703
[08:53:37] Can't kekulize mol.  Unkekulized atoms: 4
---------------------------------------------------------------------------
KekulizeException                         Traceback (most recent call last)
Input In [16], in <cell line: 14>()
     10 synthon_optimizer = torch.optim.Adam(synthon_task.parameters(), lr=1e-3)
     11 synthon_solver = core.Engine(synthon_task, synthon_train, synthon_valid,
     12                              synthon_test, synthon_optimizer,
     13                              gpus=[0], batch_size=128)
---> 14 synthon_solver.train(num_epoch=1)
     15 synthon_solver.evaluate("valid")
     16 synthon_solver.save("g2gs_synthon_model.pth")

File /usr/local/lib/python3.9/dist-packages/torchdrug/core/engine.py:155, in Engine.train(self, num_epoch, batch_per_epoch)
    152 if self.device.type == "cuda":
    153     batch = utils.cuda(batch, device=self.device)
--> 155 loss, metric = model(batch)
    156 if not loss.requires_grad:
    157     raise RuntimeError("Loss doesn't require grad. Did you define any loss in the task?")

File /usr/local/lib/python3.9/dist-packages/torch/nn/modules/module.py:1130, in Module._call_impl(self, *input, **kwargs)
   1126 # If we don't have any hooks, we want to skip the rest of the logic in
   1127 # this function, and just call forward.
   1128 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
   1129         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1130     return forward_call(*input, **kwargs)
   1131 # Do not call functions when jit is used
   1132 full_backward_hooks, non_full_backward_hooks = [], []

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:592, in SynthonCompletion.forward(self, batch)
    589 all_loss = torch.tensor(0, dtype=torch.float32, device=self.device)
    590 metric = {}
--> 592 pred, target = self.predict_and_target(batch, all_loss, metric)
    593 node_in_pred, node_out_pred, bond_pred, stop_pred = pred
    594 node_in_target, node_out_target, bond_target, stop_target, size = target

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:984, in SynthonCompletion.predict_and_target(self, batch, all_loss, metric)
    981 with reactant.graph():
    982     reactant.reaction = batch["reaction"]
--> 984 graph1, node_in_target1, node_out_target1, bond_target1, stop_target1 = self.all_edge(reactant, synthon)
    985 graph2, node_in_target2, node_out_target2, bond_target2, stop_target2 = self.all_stop(reactant, synthon)
    987 graph = self._cat([graph1, graph2])

File /usr/local/lib/python3.9/dist-packages/torch/autograd/grad_mode.py:27, in _DecoratorContextManager.__call__.<locals>.decorate_context(*args, **kwargs)
     24 @functools.wraps(func)
     25 def decorate_context(*args, **kwargs):
     26     with self.clone():
---> 27         return func(*args, **kwargs)

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:557, in SynthonCompletion.all_edge(self, reactant, synthon)
    555 assert (graph.num_edges % 2 == 0).all()
    556 # node / edge features may change because we mask some nodes / edges
--> 557 graph, feature_valid = self._update_molecule_feature(graph)
    559 return graph[feature_valid], node_in_target[feature_valid], node_out_target[feature_valid], \
    560        bond_target[feature_valid], stop_target[feature_valid]

File /usr/local/lib/python3.9/dist-packages/torchdrug/tasks/retrosynthesis.py:384, in SynthonCompletion._update_molecule_feature(self, graphs)
    382 valid = [mol is not None for mol in mols]
    383 valid = torch.tensor(valid, device=graphs.device)
--> 384 new_graphs = type(graphs).from_molecule(mols, **self.feature_kwargs)
    386 node_feature = torch.zeros(graphs.num_node, *new_graphs.node_feature.shape[1:],
    387                            dtype=new_graphs.node_feature.dtype, device=graphs.device)
    388 edge_feature = torch.zeros(graphs.num_edge, *new_graphs.edge_feature.shape[1:],
    389                            dtype=new_graphs.edge_feature.dtype, device=graphs.device)

File /usr/local/lib/python3.9/dist-packages/torchdrug/utils/decorator.py:115, in deprecated_alias.<locals>.decorate.<locals>.wrapper(*args, **kwargs)
    112         warnings.warn("%s(): argument `%s` is deprecated in favor of `%s`" % (func.__name__, key, value))
    113         kwargs[value] = kwargs.pop(key)
--> 115 return func(*args, **kwargs)

File /usr/local/lib/python3.9/dist-packages/torchdrug/data/molecule.py:716, in PackedMolecule.from_molecule(cls, mols, atom_feature, bond_feature, mol_feature, with_hydrogen, kekulize)
    714     mol = Chem.AddHs(mol)
    715 if kekulize:
--> 716     Chem.Kekulize(mol)
    718 for atom in mol.GetAtoms():
    719     atom_type.append(atom.GetAtomicNum())

KekulizeException: Can't kekulize mol.  Unkekulized atoms: 4

[FanmengWang]

hi @bhadreshpsavani @DimGorr I figured out this problem (AttributeError: can't set attribute) by the similar way and met the new problem as you. Then I use try...except to solve this problem.

if kekulize:
    try:
         Chem.Kekulize(mol)
    except:
         kekulize = False
for atom in mol.GetAtoms():
    atom_type.append(atom.GetAtomicNum())

This approach may seem odd, but useful. We can get similar result as the authors for the synthon completion part.

[DimGorr]

@FanmengWang @bhadreshpsavani After applying except: pass I got an error different from yours:

Traceback (most recent call last):
  File "tests.py", line 133, in <module>
    predictions = synthon_task.predict_reactant(batch, num_beam=10, max_prediction=5)
  File "/root/anaconda3/lib/python3.7/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context
    return func(*args, **kwargs)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/tasks/retrosynthesis.py", line 913, in predict_reactant
    new_graph = self._apply_action(graph, action, logp)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/tasks/retrosynthesis.py", line 865, in _apply_action
    meta_dict=meta_dict, **data_dict)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/molecule.py", line 617, in __init__
    offsets=offsets, atom_type=atom_type, bond_type=bond_type, **kwargs)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 1104, in __init__
    num_relation=num_relation, **kwargs)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/molecule.py", line 58, in __init__
    super(Molecule, self).__init__(edge_list=edge_list, **kwargs)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 86, in __init__
    self.edge_feature = torch.as_tensor(edge_feature, device=self.device)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 161, in __setattr__
    self._check_attribute(key, value)
  File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 1179, in _check_attribute
    (key, self.num_edge, value.shape))
ValueError: Expect edge attribute `edge_feature` to have shape (1228, *), but found torch.Size([1160, 18])

Love this tutorial:) please let me know if you find out how to resolve it

[FanmengWang]

@DimGorr Hi, I review the tutorial and find that this part is not necessary for retrosynthesis task in fact. Therefore, I do not try this part now. I think maybe you can also ignore this part firstly.

batch = []
reaction_set = set()
for sample in synthon_valid:
    if sample["reaction"] not in reaction_set:
        reaction_set.add(sample["reaction"])
        batch.append(sample)
        if len(batch) == 4:
            break
batch = data.graph_collate(batch)
batch = utils.cuda(batch)
reactants, synthons = batch["graph"]
reactants = reactants.ion_to_molecule()
predictions = synthon_task.predict_reactant(batch, num_beam=10, max_prediction=5)

synthon_id = -1
i = 0
titles = []
graphs = []
for prediction in predictions:
    if synthon_id != prediction.synthon_id:
        synthon_id = prediction.synthon_id.item()
        i = 0
        graphs.append(reactants[synthon_id])
        titles.append("Truth %d" % synthon_id)
    i += 1
    graphs.append(prediction)
    if reactants[synthon_id] == prediction:
        titles.append("Prediction %d-%d, Correct!" % (synthon_id, i))
    else:
        titles.append("Prediction %d-%d" % (synthon_id, i))

# reset attributes so that pack can work properly
mols = [graph.to_molecule() for graph in graphs]
graphs = data.PackedMolecule.from_molecule(mols)
graphs.visualize(titles, save_file="uspto50k_synthon_valid.png", num_col=6)

[DimGorr]

@FanmengWang Are you sure we can skip this part? because it is beam search and, for instance, num_synthon_beam is used even in init part of Retrosynthesis in retrosynthesis.py. And in this part, you somehow modify mols and graphs. seems like they might be used in Retrosynthesis.py and cause this error. Please let me know if misunderstood smth

[FanmengWang]

@DimGorr Hi, I think that we just need g2gs_reaction_model.pth and g2gs_synthon_model.pth for retrosynthesis task in fact. Please let me know if I misunderstood it.

solver.load("g2gs_reaction_model.pth", load_optimizer=False)
solver.load("g2gs_synthon_model.pth", load_optimizer=False)
solver.evaluate("valid")

[FanmengWang]

@bhadreshpsavani @DimGorr Hi, after CenterIdentification part and SynthonCompletion part, I want to know whether you have the same problem for the retrosynthesis part as following? Thanks !

Traceback (most recent call last):
  File "C:/Users/Mn/Desktop/learn_torchdrug/retrosynthesis.py", line 102, in <module>
    solver.evaluate("valid")
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\autograd\grad_mode.py", line 28, in decorate_context
    return func(*args, **kwargs)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\core\engine.py", line 206, in evaluate
    pred, target = model.predict_and_target(batch)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\tasks\task.py", line 27, in predict_and_target
    return self.predict(batch, all_loss, metric), self.target(batch)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\tasks\retrosynthesis.py", line 1101, in predict
    reactant = self.synthon_completion.predict_reactant(synthon_batch, self.num_synthon_beam, self.max_prediction)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torch\autograd\grad_mode.py", line 28, in decorate_context
    return func(*args, **kwargs)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\tasks\retrosynthesis.py", line 885, in predict_reactant
    synthon = synthon[feature_valid]
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\data\graph.py", line 1474, in __getitem__
    return self.subbatch(index)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\data\graph.py", line 1638, in subbatch
    return self.graph_mask(index, compact=True)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\data\graph.py", line 1616, in graph_mask
    data_dict, meta_dict = self.data_mask(node_index, edge_index, graph_index=index)
  File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\data\graph.py", line 1433, in data_mask
    v = v[graph_index]
RuntimeError: CUDA error: device-side assert triggered

[DimGorr]

@FanmengWang Hi! Concerning your first note, yeah I agree, I mixed something and beam search is used in a different part. So it should be fine. And yeah, I have the same error.

[FanmengWang]

@DimGorr Ok, I see. If you have any idea about this problem, please tell me. Thanks!

[KiddoZhu]

File /usr/local/lib/python3.9/dist-packages/torchdrug/core/core.py:93, in _MetaContainer._setattr(self, key, value)
     92 def _setattr(self, key, value):
---> 93     return super(_MetaContainer, self).__setattr__(key, value)

AttributeError: can't set attribute

This is due to a mistake in the commit. @DimGorr Your solution is right. I just fixed it in d282313.

I just found another bug for the dataset. You may get incorrect node feature shape for the synthon solver. This is fixed in a1df0a2. If you are using any old commit, you can use the following lines to construct the dataset instead

synthon_dataset = datasets.USPTO50k("~/molecule-datasets/", as_synthon=True,
                                    atom_feature="synthon_completion",
                                    kekulize=True)

which replaces the argument node_feature with atom_feature.

[KiddoZhu]

File /usr/local/lib/python3.9/dist-packages/torchdrug/data/molecule.py:716, in PackedMolecule.from_molecule(cls, mols, atom_feature, bond_feature, mol_feature, with_hydrogen, kekulize)
    714     mol = Chem.AddHs(mol)
    715 if kekulize:
--> 716     Chem.Kekulize(mol)
    718 for atom in mol.GetAtoms():
    719     atom_type.append(atom.GetAtomicNum())

KekulizeException: Can't kekulize mol.  Unkekulized atoms: 4

I didn't reproduce this problem on my side, at least for molecules in USPTO50k. I guess it might be some issues in the old versions of RDKit. I am using version 2020.09.1.0 installed from conda, and the training part is okay.

 File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 161, in __setattr__
   self._check_attribute(key, value)
 File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 1179, in _check_attribute
   (key, self.num_edge, value.shape))
ValueError: Expect edge attribute `edge_feature` to have shape (1228, *), but found torch.Size([1160, 18])

This also happens to me. I am looking into it.

[KiddoZhu]

 File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 161, in __setattr__
   self._check_attribute(key, value)
 File "/root/anaconda3/lib/python3.7/site-packages/torchdrug/data/graph.py", line 1179, in _check_attribute
   (key, self.num_edge, value.shape))
ValueError: Expect edge attribute `edge_feature` to have shape (1228, *), but found torch.Size([1160, 18])

This is a compatible issue caused by update in f8d2de6. I just fixed that in 342c87e.

 File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\data\graph.py", line 1616, in graph_mask
   data_dict, meta_dict = self.data_mask(node_index, edge_index, graph_index=index)
 File "D:\Anaconda3\envs\pytorch\lib\site-packages\torchdrug\data\graph.py", line 1433, in data_mask
   v = v[graph_index]
RuntimeError: CUDA error: device-side assert triggered

This is actually the same issue as #115 if run on CPUs. Already fixed in #115.

[FanmengWang]

@KiddoZhu Ok, thanks! I will try it again.

[bhadreshpsavani]

This issue is resolved with RDKit version 2020.09.1.0 installed from conda.

[bhadreshpsavani]

May be in the installation step we should add this so others don't face such issue

[bhadreshpsavani]

I am closing the issue