---
title: "Input formats and coordinate reference systems"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Input formats and coordinate reference systems}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r plot-options, include=FALSE}
knitr::opts_chunk$set(fig.bg = "white", dev.args = list(bg = "white"))
```

```{r setup}
library(hydromeso)
```

Data frames and matrices use named or positional field selection. CSV files are
read with base R. Existing columns and row order are retained.

```{r}
classify_mesohabitat_table(hydromeso_example, 3, 4)
```

`SpatVector` objects and vector paths supported by `terra::vect()` retain their
geometry, CRS, feature order, and attributes. This applies to points, lines,
and polygons. GeoPackage is preferable for examples; Shapefile export warns and
uses short class-field names.

```{r}
v <- mesohabitat_example_vector()
classified_v <- classify_mesohabitat_vector(v, "depth", "velocity")
plot_mesohabitat(classified_v)
```

Rasters can be separate objects/paths or selected layers in one multilayer
raster. Geometry mismatches fail by default. With
`align = "velocity_to_depth"`, velocity is projected or resampled directly to
the depth template with bilinear interpolation. Missing CRS values are never
silently assigned. `assume_crs` assigns a declared CRS without transforming
coordinates and reports that distinction.

An AOI is projected to the raster CRS if necessary. `mask_aoi = TRUE` performs
both a crop and polygon mask; `FALSE` performs only a rectangular crop.