Elastic and NMR OUTCAR needs to be refactored

[shyuep]

System

• Pymatgen version: master

  Summary

• Currently, the Outcar init is a big mess with elastic property and NMR reading all bundled in. Note that OUTCAR parsing is by nature very calculation dependent. We cannot cater to ALL kinds of run in the init method. Hence, other than the most basic parameters like energies and magmoms, all other special parsing should be done via the read_property methods, with the use of read_pattern as a general approach.
• @montoyjh Pls refactor the elastic and piezo parsing.
• @xhqu1981 Pls refactor the NMR parsing.

Thanks.

[xhqu1981]

Will do. Thanks.

[shyuep]

@montoyjh @xhqu1981 Can I have a status update pls.

[xhqu1981]

I will push the solution tonight or tomorrow.

On Sat, May 7, 2016 at 7:29 AM, Shyue Ping Ong notifications@github.com wrote:

@montoyjh https://github.com/montoyjh @xhqu1981 https://github.com/xhqu1981 Can I have a status update pls.

— You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub https://github.com/materialsproject/pymatgen/issues/350#issuecomment-217641196

[xhqu1981]

Done. @shyuep I have to add the read_table_pattern() method. This is because the NMR data are tables and context related. Very hard to defined by a single line pattern. E.g. the chemical shift table have two sections one is without core contribution, another is with core contributions. Only the second one is desired. The rows from both section are matching exactly the same line pattern. As such, I have to add one more function read_table_pattern(). Both chemical shift and EFG parsing are built on it. Concept of of named and unamed regex capturing groups is involved. It increases some complexity, however, it helps users to get the desired field only.

[montoyjh]

The read_table_pattern is a lot more flexible than my initial approach to the elastic/piezo, so I'm going to redo my changes using that method.

[xhqu1981]

@montoyjh Thanks.

[shyuep]

Thanks,