ATOMRefine

[RP-Bioinfo]

Hi, I followed all installation instructions for ATOMRefine but when I try to launch the example prediction I get the following error:

$ sh refine.sh example/T1062.pdb T1062 35 output
Using backend: pytorch
Traceback (most recent call last):
File "/home/rpracella/ATOMRefine/predict.py", line 21, in
from amber import relax
File "/home/rpracella/ATOMRefine/amber/relax.py", line 18, in
from amber import amber_minimize
File "/home/rpracella/ATOMRefine/amber/amber_minimize.py", line 29, in
import jax
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/jax/init.py", line 63, in
from jax._src.api import (
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/jax/_src/api.py", line 48, in
from jax._src import callback as jcb
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/jax/_src/callback.py", line 26, in
from jax._src import dispatch
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/jax/_src/dispatch.py", line 58, in
from etils import epath
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/etils/epath/init.py", line 19, in
from etils.epath import testing
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/etils/epath/testing.py", line 27, in
from etils.epath import gpath
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/etils/epath/gpath.py", line 61, in
@functools.cache
AttributeError: module 'functools' has no attribute 'cache'

Best Regards,

Riccardo Pracella

[ZhiYeG]

Hi
This may caused by the conflict of Jax version, could you kindly tell me what Jax you use now?

[RP-Bioinfo]

I use jax 0.3.23

[ZhiYeG]

please try to update jax to jax==0.4.13 and jaxlib==0.4.13

[RP-Bioinfo]

Ok this solution solved the previous issue, but now the command gives me another error message:

$ sh refine.sh example/T1062.pdb T1062 35 output
Traceback (most recent call last):
File "/home/rpracella/ATOMRefine/predict.py", line 14, in
from network2 import SE3Refine
File "/home/rpracella/ATOMRefine/network2.py", line 8, in
import dgl
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/dgl/init.py", line 13, in
from .backend import load_backend, backend_name
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/dgl/backend/init.py", line 95, in
load_backend(get_preferred_backend())
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/dgl/backend/init.py", line 41, in load_backend
from .._ffi.base import load_tensor_adapter # imports DGL C library
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/dgl/_ffi/base.py", line 44, in
_LIB, _LIB_NAME, _DIR_NAME = _load_lib()
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/dgl/_ffi/base.py", line 34, in _load_lib
lib = ctypes.CDLL(lib_path[0])
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/ctypes/init.py", line 373, in init
self._handle = _dlopen(self._name, mode)
OSError: libcudart.so.10.2: cannot open shared object file: No such file or directory

[ZhiYeG]

Hi, these is cause by you environment setup. You should check your the CUDA driver version you used and install the relevant PyTorch version. Also you may need to make sure you have the GPU environment

[RP-Bioinfo]

Ok I fixed my CUDA environment but now It seems that my numpy is too much upgraded. I have this error:
$ sh refine.sh example/T1062.pdb T1062 35 output
Using backend: pytorch
Traceback (most recent call last):
File "/home/rpracella/ATOMRefine/predict.py", line 20, in
from amber import protein
File "/home/rpracella/ATOMRefine/amber/protein.py", line 19, in
from amber import residue_constants
File "/home/rpracella/ATOMRefine/amber/residue_constants.py", line 576, in
restype_atom37_to_rigid_group = np.zeros([21, 37], dtype=np.int)
File "/home/rpracella/anaconda3/envs/ATOMRefine/lib/python3.8/site-packages/numpy/init.py", line 305, in getattr
raise AttributeError(former_attrs[attr])
AttributeError: module 'numpy' has no attribute 'int'.
np.int was a deprecated alias for the builtin int. To avoid this error in existing code, use int by itself. Doing this will not modify any behavior and is safe. When replacing np.int, you may wish to use e.g. np.int64 or np.int32 to specify the precision. If you wish to review your current use, check the release note link for additional information.
The aliases was originally deprecated in NumPy 1.20; for more details and guidance see the original release note at:
https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations

The problem is that my conda don't allow me to downgrade numpy for conflicts with versions of other packages. How Can I fix it? Thanks in advance

Best Regards,

Riccardo

[bernym12]

For me personally, I just went in and modified all instances of np.int to int in /home/rpracella/ATOMRefine/amber/residue_constants.py. It's rather simple and not a breaking change as I've been running ATOMRefine with no issue in regards to this. There may be similar instances of this that crop up so just perform the same change to the mentioned files and then you should be good to go.