Reads a hand-curated CSV file and converts it to the same list structure as
gq_registry_read(). Multiple rows per layer_key with a class_field
and class_value produce a classification layer. Single rows produce
simple fill/stroke/mark/label styles.
Arguments
- path
Path to a CSV file with columns: layer_key, type, source_layer, class_field, class_value, fill_color, fill_opacity, stroke_color, stroke_width, stroke_opacity, mark_color, mark_shape, mark_radius, mark_stroke_color, mark_stroke_width, label_color, label_size, label_font, label_halo_color, label_halo_width, label_offset_x, label_offset_y, note.
Value
A list with name, version, source, and layers elements,
compatible with gq_tmap_style(), gq_mapgl_style(), etc.
Raster layers
A type of "raster" carries a QGIS paletted renderer: one row per palette
entry, class_value holding the band value and class_label the palette
label. habitat_lateral is the shipped example.
class_field is a sentinel for a raster, not a column name. The
classification branch here requires a non-NA class_field, and a paletted
raster keys on pixel value rather than on any attribute — so the convention
is the literal string "value", meaning band 1. A consumer with a real band
name passes it through gq_style(field = ).
Give class_value at least one non-numeric value across the file, or accept
that read.csv() will type the column as integer; the reader coerces back to
character, but a numeric column reads as an ordinal in every other tool.
What a raster row cannot carry: QGIS's per-value rasterTransparency
(habitat_lateral sets 30% on top of the renderer's 0.4 opacity), multi-band
renderers, and resampling. Those live in the QML, which is lossless — reach
for gq_style_qml() whenever the consumer is QGIS itself.
The tmap and mapgl translators do not render rasters. gq_tmap_style(),
gq_mapgl_style() and gq_mapgl_classes() all refuse a raster rather than
returning something plausible; gq_tmap_classes() works, since a palette is
a classification like any other.
Examples
path <- system.file("registry", "reg_custom.csv", package = "gq")
reg <- gq_reg_custom(path)
names(reg$layers)
#> [1] "bec_zone" "rivers_poly"
#> [3] "dam" "town"
#> [5] "harvest_area" "planting_site"
#> [7] "old_growth_management_areas" "national_park"
#> [9] "habitat_lateral"
# Classified layer (multiple rows per layer_key)
reg$layers$bec_zone$classification$field
#> [1] "ZONE"
names(reg$layers$bec_zone$classification$classes)
#> [1] "SBS" "ESSF" "ICH" "BWBS" "CWH" "MS" "SBPS" "SWB" "AT" "MH"
#> [11] "BG"
# Simple layer (single row)
reg$layers$rivers_poly$fill
#> $color
#> [1] "#7ba7cc"
#>
#> $opacity
#> [1] 0.7
#>