Argo Data accessed locally by the software are assumed to be organised following the GDAC folder structure

This requirement is crucial for reproducibility and to foster broad use-cases while contributing to the robustness of the service.

If the software takes as input any of the GDAC files and support local data, the software must assume that local data are organised following the GDAC server structure.

To the opposite, the software must not depends on any random localisation of a GDAC file.

Typically, the software must assumed that input files are in any of the following folders:

```markdown
./
│
├── ar_index_global_meta.txt
├── ...
├── readme_before_using_the_data.txt
│
├── dac/ # Data sorted by Data Assembly Centre (DAC)
│ ├── aoml/
│ ├── bodc/
│ │ ├── *.nc
│ │ └── profiles/
│ │ └── **.nc
│ ├── ...
│ └── nmdis/
│
├── geo/ # Data sorted by Ocean and Year/Month
│ ├── atlantic_ocean/
│ │ ├── 1997
│ │ ├── ...
│ │ └── 2026/
│ │ ├── 01/
│ │ │ ├── 20260101_prof.nc
│ │ │ ├── ...
│ │ │ └── 20260131_prof.nc
│ │ ├── ...
│ │ └── 12/
│ ├── indian_ocean/
│ │ └── ...
│ └── pacific_ocean/
│ └── ...
│
├── aux/ # Auxiliary data sorted by DACs, free form
│ ├── aoml
│ ├── ...
│ ├── coriolis
│ │ └── /
│ │ └── _tech_aux.nc
│ └── meds
│
├── etc/ # Other data
│ ├── argo_aux-profile_index.txt
│ ├── ...
│ └── argo-index/
│ ├── argo_profile_detailled_index.txt
│ └── ...
│
└── latest_data/ # Last 4 months data
└── _prof_<0/1/2/...>.nc
```
```