Bug: OSError: could not get source code #609
Description
I downloaded the pre_processed data and tried running the tutorial here https://github.com/DeepRank/deeprank2/blob/main/tutorials/training.ipynb
At the stage of
trainer.train(
nepoch=epochs,
batch_size=batch_size,
earlystop_patience=earlystop_patience,
earlystop_maxgap=earlystop_maxgap,
min_epoch=min_epoch,
validate=True,
filename=os.path.join(output_path, f"gnn_{task}", "model.pth.tar"),
)
The output is:
INFO:deeprank2.trainer:Training set loaded
INFO:deeprank2.trainer:Validation set loaded
INFO:deeprank2.trainer:Epoch 0:
INFO:deeprank2.trainer:training loss 0.8269949356714884 | time 0.3814964294433594
INFO:deeprank2.trainer:validation loss 0.8072875671916537 | time 0.0742034912109375
INFO:deeprank2.trainer:Epoch 1:
INFO:deeprank2.trainer:training loss 0.7365012235111661 | time 0.36728739738464355
INFO:deeprank2.trainer:validation loss 0.6814224786228604 | time 0.07352352142333984
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/apps/software/Anaconda3/2022.05/envs/deeprank2/lib/python3.10/site-packages/deeprank2/trainer.py", line 649, in train
checkpoint_model = self._save_model()
File "/apps/software/Anaconda3/2022.05/envs/deeprank2/lib/python3.10/site-packages/deeprank2/trainer.py", line 949, in _save_model
str_expr = inspect.getsource(key["transform"])
File "/apps/software/Anaconda3/2022.05/envs/deeprank2/lib/python3.10/inspect.py", line 1139, in getsource
lines, lnum = getsourcelines(object)
File "/apps/software/Anaconda3/2022.05/envs/deeprank2/lib/python3.10/inspect.py", line 1121, in getsourcelines
lines, lnum = findsource(object)
File "/apps/software/Anaconda3/2022.05/envs/deeprank2/lib/python3.10/inspect.py", line 958, in findsource
raise OSError('could not get source code')
OSError: could not get source code
I have deeprank2 installed in a cluster, with permissions rwxr.x... and I run it with my user that belongs to the group (ie running it with permissions r.x). I tried running it with full rwx as well, but got the same error.
Also, I noticed that the file in https://github.com/DeepRank/deeprank2/blob/main/deeprank2/trainer.py is not identical to the file that gets installed. At least, in the save_model() function they differ:
In the file:
if features_transform_to_save:
for key in features_transform_to_save.values():
if key["transform"] is None:
continue
str_expr = inspect.getsource(key["transform"])
match = re.search(r"[\"|\']transform[\"|\']:.*(lambda.*).*,.*[\"|\']standardize[\"|\'].*", str_expr).group(1)
key["transform"] = match
But in github:
if features_transform_to_save:
for key in features_transform_to_save.values():
if key["transform"] is None:
continue
# Serialize the function
serialized_func = dill.dumps(key["transform"])
# Deserialize the function
deserialized_func = dill.loads(serialized_func) # noqa: S301
str_expr = inspect.getsource(deserialized_func)
match = re.search(r"[\"|\']transform[\"|\']:.*(lambda.*).*,.*[\"|\']standardize[\"|\'].*", str_expr).group(1)
key["transform"] = match
Any hints as to the OS error?