AttributeError: module 'numpy' has no attribute 'int'

[lubilok]

Hi,
I was trying to find out the true interface of the liquid/vapor interface in my system using Pytim modules, but I am getting an error like below

import MDAnalysis as mda
import pytim
u = mda.Universe("md_0_1.gro")
interface = pytim.ITIM(u)
interface.writepdb('layers.pdb',centered='middle')
upper_layer = interface.layers[0,0]

AttributeError Traceback (most recent call last)
/tmp/ipykernel_80284/2357494757.py in
4 # initialize the interface, this has to be called only once
5 # before iterating over the trajectory!
----> 6 interface = pytim.ITIM(u)
7
8 # write a pdb with the information on the surface layers

~/anaconda3/lib/python3.9/site-packages/pytim/itim.py in init(self, universe, group, alpha, normal, molecular, max_layers, radii_dict, cluster_cut, cluster_threshold_density, extra_cluster_groups, info, centered, warnings, mesh, autoassign, **kargs)
217 cluster_threshold_density, extra_cluster_groups)
218 sanity.assign_normal(normal)
--> 219 sanity.assign_radii(radii_dict=radii_dict)
220
221 self.grid = None

~/anaconda3/lib/python3.9/site-packages/pytim/sanity_check.py in assign_radii(self, radii_dict)
146 self.patch_radii_dict()
147
--> 148 _missing_attributes(self.interface, self.interface.universe)
149
150 try:

~/anaconda3/lib/python3.9/site-packages/pytim/properties.py in _missing_attributes(interface, universe)
72 _check_missing_attribute(interface, 'elements', 'Elements', universe.atoms,
73 1)
---> 74 _extra_attributes(interface, universe)
75
76

~/anaconda3/lib/python3.9/site-packages/pytim/properties.py in _extra_attributes(interface, universe)
81 for key in attr.keys():
82 if key not in dir(universe.atoms):
---> 83 vals = np.zeros(len(universe.atoms), dtype=np.int) - 1
84 universe.add_TopologyAttr(attrkey)
85

~/anaconda3/lib/python3.9/site-packages/numpy/init.py in getattr(attr)
282 return Tester
283
--> 284 raise AttributeError("module {!r} has no attribute "
285 "{!r}".format(name, attr))
286

AttributeError: module 'numpy' has no attribute 'int'

Please tell me how I can fix it.
Thank you.

[Marcello-Sega]

Hi and thanks for reporting this! Could you let me know which numpy version you are using? I think I know what the problem is and I’ll fix it as soon as possible.

…

On Thu, 9 May 2024 at 19:22, lubilok ***@***.***> wrote: Hi, I was trying to find out the true interface of the liquid/vapor interface in my system using Pytim modules, but I am getting an error like below import MDAnalysis as mda import pytim u = mda.Universe("md_0_1.gro") interface = pytim.ITIM(u) interface.writepdb('layers.pdb',centered='middle') upper_layer = interface.layers[0,0] AttributeError Traceback (most recent call last) /tmp/ipykernel_80284/2357494757.py in 4 # initialize the interface, this has to be called only once 5 # before iterating over the trajectory! ----> 6 interface = pytim.ITIM(u) 7 8 # write a pdb with the information on the surface layers ~/anaconda3/lib/python3.9/site-packages/pytim/itim.py in init(self, universe, group, alpha, normal, molecular, max_layers, radii_dict, cluster_cut, cluster_threshold_density, extra_cluster_groups, info, centered, warnings, mesh, autoassign, **kargs) 217 cluster_threshold_density, extra_cluster_groups) 218 sanity.assign_normal(normal) --> 219 sanity.assign_radii(radii_dict=radii_dict) 220 221 self.grid = None ~/anaconda3/lib/python3.9/site-packages/pytim/sanity_check.py in assign_radii(self, radii_dict) 146 self.patch_radii_dict() 147 --> 148 _missing_attributes(self.interface, self.interface.universe) 149 150 try: ~/anaconda3/lib/python3.9/site-packages/pytim/properties.py in _missing_attributes(interface, universe) 72 _check_missing_attribute(interface, 'elements', 'Elements', universe.atoms, 73 1) ---> 74 _extra_attributes(interface, universe) 75 76 ~/anaconda3/lib/python3.9/site-packages/pytim/properties.py in _extra_attributes(interface, universe) 81 for key in attr.keys(): 82 if key not in dir(universe.atoms): ---> 83 vals = np.zeros(len(universe.atoms), dtype=np.int) - 1 84 universe.add_TopologyAttr(attrkey <https://github.com/MDAnalysis/mdanalysis/discussions/vals>) 85 ~/anaconda3/lib/python3.9/site-packages/numpy/init.py in getattr(attr) 282 return Tester 283 --> 284 raise AttributeError("module {!r} has no attribute " 285 "{!r}".format(name, attr)) 286 AttributeError: module 'numpy' has no attribute 'int' Please tell me how I can fix it. Thank you. — Reply to this email directly, view it on GitHub <#405>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AADRZDYOPF65BWPA3HWG65TZBO5ERAVCNFSM6AAAAABHPIYN36VHI2DSMVQWIX3LMV43ERDJONRXK43TNFXW4OZWGYZTMNZYGU> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[lubilok]

Thank you so much for your help and for giving me ideas about ITIM.

[lubilok]

Hi,
After running the code you mentioned above, I got pdb, but some molecules are breaking.

[Marcello-Sega]

That looks ok to me, you are working with periodic boundary conditions. If you prefer, for the sake of visualisation, you can center the liquid slabe before saving, have a look at the online manual:

https://marcello-sega.github.io/pytim/center.html

[lubilok]

Hi,
Could you tell me how you decide this value (alpha=2, molecular=True, max_layers=3, cluster_cut=3.5)? I have tried using one system with cluster_cut=3.5. I didn't get any 0.50 value for the vapor phase, but when I change the value of cluster_cut to 2.5, I'm getting all the values in 0.5, however, in the case of 3, I get a smaller number of 0.50 values, so I don't know how to decide this thing.

What is the logic behind deciding the sphere, and how is it implemented in the code?

[Marcello-Sega]

For water, the first mininum of the rdf is located at around 3.5 Angstrom over a wide range of temperatures. This is usually a good choice to distinguish vapour from liquid molecules, at least at low temperatures, because most contact neighbors in a liquid cluster will be found below that distance. Of course, if you are simulating another liquid, you'll have to compute the rdf and check where the first minimum is located.

I wouldn't choose any other value than the location of the rdf minimum, unless you need to go to high temperatures. Closer to the critical point you should use more refined, density-based clustering methods (see https://pubs.rsc.org/en/content/articlehtml/2017/cp/c7cp02918g) ; In particular, for its application to the liquid/vapour interface of water see https://pubs.rsc.org/en/content/articlehtml/2023/sm/d3sm00176h
In pytim you can switch this on by passing, e.g., cluster_cut=12. , cluster_threshold_density='auto' , where cluster_cut now needs to be large enough to sample the local density. For water, usually 12 Angstrom is a good choice; don't go below 10 and much above 15. See the convergence plots in the 2017 paper if you're interested in reproducing the same for your system.

Regading your surprise at not finidng molecules in the vapor phase: at low temperatures (300 K) is completely normal to find from zero to 4-5 molecules in the vapour phase for a simulation of ~ 1000 molecules. Just convert the vapor and liquid densities from the coexistence diagram into number of molecules in your simulation box and you'll see what I mean.