xarray backend for Selafin formats

Supports lazy loading by default.

Install

To have the backend working in xarray, you need to follow these steps:

pip install xarray-selafin

or, if you are using conda/mamba:

conda install -c conda-forge xarray_selafin

Read Selafin

import xarray as xr

with xr.open_dataset("tests/data/r3d_tidal_flats.slf", engine="selafin") as ds:
    print(ds)  # do something with `ds`...
    # `ds.close()` not necessary

ds = xr.open_dataset("tests/data/r3d_tidal_flats.slf", lang="fr", engine="selafin")  # if variables are in French
print(ds)  # do something with `ds`...
ds.close()  # avoid a ResourceWarning (unclosed file)

<xarray.Dataset> Size: 5MB
Dimensions:  (time: 17, plan: 21, node: 648)
Coordinates:
    x        (node) float32 3kB ...
    y        (node) float32 3kB ...
  * time     (time) datetime64[ns] 136B 1900-01-01 ... 1900-01-02T20:26:40
Dimensions without coordinates: plan, node
Data variables:
    Z        (time, plan, node) float32 925kB ...
    U        (time, plan, node) float32 925kB ...
    V        (time, plan, node) float32 925kB ...
    W        (time, plan, node) float32 925kB ...
    MUD      (time, plan, node) float32 925kB ...
Attributes:
    title:       Sloped flume Rouse profile test
    language:    en
    float_size:  4
    endian:      >
    params:      (1, 0, 0, 0, 0, 0, 21, 5544, 0, 1)
    ipobo:       [   1  264  263 ... 5411 5412 5413]
    ikle2:       [[155 153 156]\n [310 307 305]\n [308 310 305]\n ...\n [537 ...
    ikle3:       [[  155   153   156   803   801   804]\n [  310   307   305 ...
    variables:   {'Z': ('ELEVATION Z', 'M'), 'U': ('VELOCITY U', 'M/S'), 'V':...
    date_start:  (1900, 1, 1, 0, 0, 0)

Indexing

ds_last = ds.isel(time=-1)  # last frame

Manipulate variables

ds = ds.assign(UTIMES100=lambda x: x.U * 100)  # Add a new variable
# ds.attrs["variables"]["UTIMES100"] = ("UTIMES100", "My/Unit")  # To provide variable name and unit (optional)
ds.drop_vars(["W"])  # Remove variable `VELOCITY W`

Extracting a specific layer from a 3D DataSet

ds_bottom = ds.selafin.get_dataset_as_2d(plan=0)  # bottom layer

Write Selafin

ds.selafin.write("output_file.slf")

DataSet content

Dimensions

In 2D:

1. time
2. node

in 3D:

1. time
2. plan
3. node

Coordinates

Coordinate	Description
x	East mesh coordinates
y	North mesh coordinates
time	Datetime serie

Attributes

All attributes are optional except ikle2:

Attribute	Description	Default value
title	Serafin title	"Converted with array-serafin"
language	Language for variable detection	"en"
float_size	Float size	4 (single precision)
endian	File endianness	">"
params	Table of integer parameters	(can be rebuilt)
ikle2	Connectivity table in 2D (1-indexed)	-
ikle3	Connectivity table in 3D (1-indexed, only in 3D, optional)	(can be rebuilt from 2D)
variables	Dictionary with variable names and units (key is variable abbreviation)	-
date_start	Starting date with integers (year to seconds)	(from first time serie)