I want to clarify that for structures (pulled using MPRester) with the label theoretical = False, the initial structure is that which was reported by the experimental study with no relaxation whatsoever.
As a followup, I’d also like to make sure that the default structure pulled using MPRester’s get_data() method is the “initial structure”, e.g.,
The short answer is there is no guarantee that initial_structure is an ICSD structure since it just corresponds to the initial structure from the geometry optimization that yielded the lowest energy structure. That initial structure may be from the ICSD, but it also may have been a hypothetical structure created by substituting another structure.
The ICSD tags come from structure matching after the material document is created.
Would it be reasonable to assume that the initial_structure is close to the experimental structure? Or if I’m wanting experimental structures, do I really need to go get things directly from ICSD?
Actually @shyamd I think it would be more instructive if you could just tell me where I can read more about the “structure matching” that you mentioned. That might give me an idea of the tolerances involved in comparing the ICSD experimental structure and the initial_structure