vhchavez/moly

Molecular Visualization powered by Plotly

👨‍💻🤸‍♂️
Software + Physics + UX

moly's Issues

Don't add images to repo

Adding rendered images to the repo is irreversible and can bloat the size if changed a lot. Consider adding Jupyter notebook that are auto rendered in nbsphinx for the docs.

Surface

Have we tried go.Surface?

Layering molecules updates incorrectly the max range

Given n molecules of considerable different size.
Layering them by decreasing size keeps the range of the latest molecule added instead of the larger.

moly/layers/cube.py negelecting first two characters of a directory

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of moly.

• (optional) I have confirmed this bug exists on the master branch of moly.

---

Code Sample, a copy-pastable example

import os
import psi4
import moly

os.mkdir('test')
psi4.set_options({'CUBEPROP_FILEPATH': 'test', 'CUBEPROP_TASKS': ['density']})
psi4.cubeprop(wfn)

fig = moly.Figure()
fig.add_cubes(directory='test')
fig.show()

Error:

---------------------------------------------------------------------------
FileNotFoundError                         Traceback (most recent call last)
<ipython-input-2-5fc546037018> in <module>
      3 # fig.add_molecule("Li", molecule)
      4 # fig.add_density(wfn=wfn, densities=['Da'], iso=0.1)
----> 5 fig.add_cubes(directory='test')
      6 # fig.add_density(wfn=wfn, densities='Da', iso=0.1)
      7 fig.show()

~/.local/lib/python3.6/site-packages/moly/figure/figure.py in add_cubes(self, directory, iso, style, colorscale, opacity)
    131     def add_cubes(self, directory=".", iso=0.03, style="ball_and_stick", colorscale="Portland_r", opacity=0.3):
    132         cubes, details = get_cubes(directory)
--> 133         geometry, symbols, atomic_numbers, spacing, origin = cube_to_molecule(details[0]["name"]+".cube")
    134         bonds = qcel.molutil.guess_connectivity(symbols, geometry)
    135         add_bonds(geometry, symbols, bonds, self.fig, style, self.surface)

~/.local/lib/python3.6/site-packages/moly/layers/cube.py in cube_to_molecule(cube_file)
    139 def cube_to_molecule(cube_file):
    140 
--> 141     _ , meta = cube_to_array(cube_file)
    142     origin = meta["origin"]
    143     atoms = meta["geometry"]

~/.local/lib/python3.6/site-packages/moly/layers/cube.py in cube_to_array(file)
    171     """
    172     cube_details = {}
--> 173     with open(file, 'r') as cube:
    174         cube.readline()
    175         cube.readline()  # ignore comments

FileNotFoundError: [Errno 2] No such file or directory: 'st/Psi_a_1_1-Ag.cube'

Problem description

In moly/layers/cube.py, line 64 has details.update({"name":cube_file[2:-5]}) and thus it is chopping off the first two characters of the directory. It should be details.update({"name":cube_file[0:-5]}). I have tested this locally and it works as expected.

OO API

An API like the following can be awkward:

fig.add_molecule("QC", molecule=ammonia_dimer)
fig.add_molecule("xyz", file='bucky.xyz')

It is a bit overloaded in general and doesn't follow any common python pattern.

Consider always using the pattern add_x(name, object) where we would have:

fig.add_molecule("mol1", Molecule)
fig.add_molecule("mol2", Molecule.from_file('Bucky.xyz"))
fig.add_orbital("Orbital1", Wavefunction)

It bring in an interesting question if we have two contexts:

fig.add_measurement("measure1", "mol1", [0, 1, 2])

Molpy?

Or Moly?

None of the functions has a docstring

Molecule Adapter Pattern

The Molecule adapter pattern in this case is overkill and adds additional overhead. I would highly recommend using QCElemental as the base molecule and relying on its readers.

Tests

Currently no tests are implemented.

[] Importing molecules
[] Adding layers
[] Showing plot (One needs to be creative on how to test this :))

Slider Bug

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of moly.

• (optional) I have confirmed this bug exists on the master branch of moly.

---

Code Sample, a copy-pastable example

fig = moly.Figure()
fig.add_cube('Psi_a_9_5-A1.cube', iso=[0.03, 0.04, 0.05], plot_geometry=True)
fig.show()

Error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-2-29b34361b222> in <module>
      1 fig = moly.Figure()
----> 2 fig.add_cube('Psi_a_9_5-A1.cube', iso=[0.03, 0.04, 0.05], plot_geometry=True)
      3 fig.show()

~/Python_stuff/psi4labs/moly/moly/figure/figure.py in add_cube(self, file, iso, plot_geometry, colorscale, opacity, style)
    127                 self.fig.add_trace(trace)
    128 
--> 129             slider = get_slider(iso, geometry_traces)
    130             self.fig.update_layout(sliders=slider)
    131 

~/Python_stuff/psi4labs/moly/moly/figure/widgets.py in get_slider(iso, geometry_traces)
     44         one_step = {'method': "restyle", 
     45                     'label': str(iso_i),
---> 46                     'args': [{"visible": [True] * geometry_traces + [False] * len(iso)}]}
     47         one_step["args"][0]['visible'][geometry_traces + i] = True
     48         steps.append(one_step)

TypeError: can't multiply sequence by non-int of type 'tuple'

Problem description

In moly/figure/figure.py, line 120 should have the len removed.
In moly/figure/widgets.py, lines 46 and 47 do not include the len for the geometry_traces parameter.

I have tested this locally and it works as expected.

See PR #15

Launch Binder Not working

Hello,

Your Launch Binder feature is not working but it is working on google colab. Can we switch the demo to there?

Atom colors are incomplete