Questions about Output file

[Ricardo_Jauregui_UPC]

Dear All,

My name is Ricardo Jauregui, I work at UPC and I am in charge of FDTD
Post-processing and validation process. We need to clarify the order
of the output data as soon as possible.

Analysing the examples of AmeletHDF files() an having studied the
“AmeletHDF-release 1.4.0 (AmeletHD-1.4.0.pdf), we still do not fully
understand how is the output data organized.

If we have understood correctly, all the output value of field,
voltage, current, etc. it should be save in the “FloatingType” folder.
However, we can see that many things remain to be defined.

It would be more efficient and manageable if we organize the results
(FloatingType folder) in two folders, Time an Frequency domain. For
each output data, we can organize it into sub-folders with each axis
information or results. For example:

Output.h5/
`-- floatingType
|-- Time_Domain
| |-- E_field1
| | |-- X_value
| | |-- Y_value
| | `--Z_value
| |-- H_field1
| | |-- X_value
| | |-- Y_value
| | `--Z_value
| |-- Voltage_load1
| | `-- Data
| |-- Voltage_source1
| | `-- Data
| |-- Current_source1
| | `-- Data
| |
| `--Time_array
|
|-- Frequency_Domain
| |-- E_field
| | |-- X_value
| | | |-- Real_value
| | | |-- Imag_value
| | | `--ABS_value
| | |-- Y_value
| | | |-- Real_value
| | | |-- Imag_value
| | | `--ABS_value
| | |-- Z_value
| | | |-- Real_value
| | | |-- Imag_value
| | | `--ABS_value
| | `-- Total_field
| | |-- Real_value
| | |-- Imag_value
| | `--ABS_value
| |
| |-- Voltage_load1
| | |-- Real_value
| | |-- Imag_value
| | `--ABS_value
...

We think that this file structure, for the final partner, could enable
the required data to process, to be found quickly. In the same way, we
think it is important to define a standard name of all output data.

Does everyone agree? Or does anyone any improvement on this output
organization system?

Best regards,
Ricardo Jauregui

[cyril giraudon]

Hello Ricardo,

>
> Analysing the examples of AmeletHDF files() an having studied the
> “AmeletHDF-release 1.4.0 (AmeletHD-1.4.0.pdf), we still do not fully
> understand how is the output data organized.
>
> If we have understood correctly, all the output value of field,
> voltage, current, etc. it should be save in the “FloatingType” folder.
> However, we can see that many things remain to be defined.
>
> It would be more efficient and manageable if we organize the results
> (FloatingType folder) in two folders, Time an Frequency domain. For
> each output data, we can organize it into sub-folders with each axis
> information or results. For example:

From a human being point of you, it would certainly be more
comfortable to read data in organized folders.
But in the context of HIRF-SE, data are read by softwares and
softwares don't need classification.

Even if data were classified, a software must checked Output.h5:/
floatingType/Time_Domain/E_field1 is
an electric field.

In fact, this data shoudl be written in Amelet-HDF as follows :

Output.h5
`-- floatingType
`-- E_field1[@type=arraySet]
|-- data[@physicalNature=electric_field]
`-- ds
|-- dim1[@physicalNature=component]
| | x |
| | y |
| | z |
`-- dim2[@physicalNature=time
@unit=second]
|1e-9|
|... |
|... |
|10-9|

FloatingType's name is free but metadata show perfectly that E_field
contains electric_field stored in the time domain.

So you proposal brings redundancy and redundancy is dangerous.

Best Regards,

Cyril.