can't use jnk3 and gsk3b models to call oracle functions

[HEU2021201622]

Describe the bug
now i'm using these model to get the ocacle scores. but i found that the size of current version model is 675B, it seems impossible to store a RF model and in fact , i can't use the model to predict because i met the problem "you need to call fit to get a trained model".

Then i try to use the older version models, but i can't get the right version of scikit-learn. do you have some advices?

To Reproduce
Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Expected behavior
A clear and concise description of what you expected to happen.

Screenshots
If applicable, add screenshots to help explain your problem.

Environment:

• OS:Linux
• Python version:3.8.12
• TDC version:0.4.1
• Any other relevant information:

Additional context

ValueError Traceback (most recent call last)
Cell In[8], line 2
1 from tdc import Oracle
----> 2 oracle = Oracle(name = 'JNK3')
3 smi = "O=C([O-])C(CC1=c2cccc(O)c2=[NH+]C1)c1cccc(Br)c1"
4 oracle(smi)

File ~/miniconda3/envs/pmo/lib/python3.8/site-packages/tdc/oracles.py:67, in Oracle.init(self, name, target_smiles, num_max_call, **kwargs)
65 self.name = name
66 self.evaluator_func = None
---> 67 self.assign_evaluator()
68 self.num_called = 0
70 if num_max_call is not None:

File ~/miniconda3/envs/pmo/lib/python3.8/site-packages/tdc/oracles.py:109, in Oracle.assign_evaluator(self)
106 elif "jnk3" in self.name:
107 from .chem_utils import jnk3
--> 109 oracle_object = jnk3()
110 self.evaluator_func = oracle_object
111 elif self.name == "similarity_meta":

File ~/miniconda3/envs/pmo/lib/python3.8/site-packages/tdc/chem_utils/oracle/oracle.py:699, in jnk3.init(self)
697 if SKLEARN_VERSION >= version.parse("0.24.0"):
698 jnk3_model_path = "oracle/jnk3_current.pkl"
--> 699 self.jnk3_model = load_pickled_model(jnk3_model_path)

File ~/miniconda3/envs/pmo/lib/python3.8/site-packages/tdc/chem_utils/oracle/oracle.py:457, in load_pickled_model(name)
455 try:
456 with open(name, "rb") as f:
--> 457 model = pickle.load(f)
458 except EOFError:
459 import sys

File sklearn/tree/_tree.pyx:728, in sklearn.tree._tree.Tree.setstate()

File sklearn/tree/_tree.pyx:1434, in sklearn.tree._tree._check_node_ndarray()

ValueError: node array from the pickle has an incompatible dtype:

• expected: {'names': ['left_child', 'right_child', 'feature', 'threshold', 'impurity', 'n_node_samples', 'weighted_n_node_samples', 'missing_go_to_left'], 'formats': ['<i8', '<i8', '<i8', '<f8', '<f8', '<i8', '<f8', 'u1'], 'offsets': [0, 8, 16, 24, 32, 40, 48, 56], 'itemsize': 64}
• got : [('left_child', '<i8'), ('right_child', '<i8'), ('feature', '<i8'), ('threshold', '<f8'), ('impurity', '<f8'), ('n_node_samples', '<i8'), ('weighted_n_node_samples', '<f8')]

[star-bo]

I encountered the same issue. Have you found a solution? Thank you!

[yukang123]

same here. Have you solved this issue? Thanks.

[CiaoHe]

Downgrade PyTDC to 1.1.14 works for me