How to convert CSV data to a compliant openPMD 1.0.0 external field HDF5(.h5) file?

[dinhyi]

Hello,

I am a graduate student working with WarpX v25.04. I am trying to create an external magnetic field from my own CSV data file.

I have an example HDF5 file (example-femm-3d.h5) that works correctly with my simulation. I used a script to inspect its structure, and this is the "blueprint" I need to create:

/ (Root Group)
├── attributes:
│ ├── basePath: b'/data/%T/'
│ ├── iterationEncoding: b'groupBased'
│ ├── meshesPath: b'meshes/'
│ ├── openPMD: b'1.1.0'
│ └── ... and other root attributes
└── Group: data
└── Group: 1
├── attributes:
│ ├── dt: 1.0
│ ├── time: 0.0
│ └── timeUnitSI: 1.0
└── Group: meshes
└── Group: B
├── attributes:
│ ├── gridGlobalOffset: [-1.15, -1.15, -0.375]
│ ├── gridSpacing: [0.05, 0.05, 0.125]
│ └── ... and other mesh attributes
└── Dataset: x (Shape: (47, 47, 47))
└── Dataset: y (Shape: (47, 47, 47))
└── Dataset: z (Shape: (47, 47, 47))

My source data is a CSV file with scattered points (r, z, Hr, Hz) from a magnet design. My WarpX version seems to require the openPMD 1.0.0 standard.

My questions are:

1. What is the recommended way to write a Python script with h5py to convert my CSV data (after interpolation) into an HDF5 file with the exact structure shown above?

2. Is there an official tutorial or example script for this specific task (converting external CSV/array data to a compliant openPMD file) that I might have missed? If so, could you please provide a link?

Thank you very much for your time and help.

[RemiLehe]

Hi @dinhyi
Thanks for your interest in WarpX.

Here is an example reading fields from a text file (as a 2D array, as a function of r and z), and writing a corresponding openPMD file:
https://github.com/openPMD/openPMD-example-datasets/blob/draft/scripts/femm/openpmd_create_rz.py
Instead of using h5py, we recommend using openPMD-api (as in the above example).

Note that, in the above example, the format of the text file seems to be a bit different than the one that you have, but hopefully the example Python script linked above should be easily adaptable.

[dinhyi]

Thank you so much for your help.