id summary reporter owner description type status priority milestone component version resolution keywords cc blockedby blocking notify_on_close platform project 1063 Collections don't accept Numpy arrays with mismatched type Tristan Croll Tom Goddard "{{{ import numpy m = session.models.list()[0] m.atoms.coords = m.atoms.coords.astype(numpy.float32) --------------------------------------------------------------------------- ArgumentError Traceback (most recent call last) in () ----> 1 m.atoms.coords = m.atoms.coords.astype(numpy.float32) ~/apps/chimerax/lib/python3.6/site-packages/chimerax/core/atomic/molc.py in set_prop(self, values) 194 va[:] = values 195 v = pointer(va) --> 196 cset(self._c_pointers, n, v) 197 198 return property(get_prop, set_prop, doc = doc) ArgumentError: argument 3: : expected LP_c_double instance instead of LP_c_float }}} This came up for me because I have a Collection class with a 64-bit cvec_property that uses Atoms.elements.masses (which returns a float32 array) as an input to calculate spring constants. Of course I can cast it to the required type myself, but would it be better to change line 181 of molc.py to the following? {{{ if isinstance(values,ndarray) and values.ndim == vdim and values.dtype == value_type: }}}" defect closed moderate Core not a bug all ChimeraX