[GSoC] Set up HDF5 file structure for datasets

[gabrielasd]

Modify the current Python code to store the atomic species data (e.g., electronic configurations, electronic density properties, etc.) in an HDF5 structure. Enable initialization of a Species instance from this HDF5 file.

For each dataset (e.g., gaussian, nist, etc.), the goal is to have a single HDF5 file per element. For example for Carbon atom, the HDF5 file would have the rough structure:

├── C
    ├── C_-01_004_000
    │   ├── Table(s)
    ├── C_-02_003_000
    │   ├── Table(s)
...

Module associated with this change: species.py

Some of the methods/functions that are expected to change are:
SpeciesData
Species
load
dump

Tasks

• Design HDF5 structure for species datasets.
• Add methods to load and read data from HDF5 files
• Modify species.py module to initialize a Species from HDF5 data and dump an instance of Species to an HDF5 file
• Add tests

[gabrielasd]

@enjyashraf18
For this issue lets take the slater dataset as the toy dataset given that it is currently set at the default one

AtomDB/atomdb/utils.py

Line 40 in 464f116

DEFAULT_DATASET = "slater"

As first step, define the schema/description for the dataset inside the run.py file. Include this description at the top, e.g. at the level of L33

This table from the documentation can be used as a guide for what properties to include. The relevant column there is the one labelled Slater HF.
From the properties listed exclude from atmass to dispersion_c6 as these will be pulled from the elements HDF5 data.

[msricher]

The run.py for each dataset should look like:

class DefinitionClass:
    #.... (or something that defines the data structure for the dataset)
 
def run(...):
    #....
    return <definitionclass object>

And then the compile_species function adds it to the DB. The structure you showed us today will work well for this.