up

index.qmd

@@ -14,73 +14,116 @@ format:
   revealjs:
     theme: [simple, style.scss] # beige blood dark default league moon night serif simple sky solarized
     incremental: true
-    toc: false
-    toc-depth: 2
+    toc: true
+    toc-depth: 1
     slide-number: true
     overview: true
     code-line-numbers: false
+    highlight-style: ayu
 execute:
     echo: true
 ---
 
-## Read/Write Backends
+# What is a raster?
+
+. . .
+
+:::: {.columns}
+
+::: {.column width="50%"}
+![](https://raw.githubusercontent.com/NEONScience/NEON-Data-Skills/main/graphics/raster-general/raster_concept.png)
+:::
+
+::: {.column width="50%"}
+- like an image, but not RGB
+- values of some variable accross a gridded space
+- usually has X/Y spatial dimensions 
+  - (e.g. lattitude/longitude)
+- with a coordinate reference system
+:::
+
+::::
+
+---
+
+:::: {.columns}
+
+::: {.column width="50%"}
+![](https://esd.copernicus.org/articles/11/201/2020/esd-11-201-2020-f03-web.png)
+:::
+
+::: {.column width="50%"}
+- may be collected in a dataset with multiple variables
+- may have more dimensions, like time
+:::
+
+::::
+
+
+# DimensionalData.jl integration
+
+- Rasters extends DimensionalData.jl
+  - `Raster <: AbstractDimArray` and `RasterStack <: AbstractDimStack`
+- This gives is the foundation for spatial work
+- Rasters adds:
+  - coordinate reference systems
+  - missing value handling
+  - File IO 
+  - GIS tools
+
+## File Read/Write Backends
 
 File types                      | Package
 ------------------------------- | ----------
 Netcdf/hdf5                     | NCDatasets.jl
 Grib (read only)                | GRIBDatasets.jl
 Zarr (PR nearly done!)          | ZarrDatasets.jl
-R grd (simple Mmap data from R) | native
+grd (simple Mmap data from R)   | native
 GeoTIFF and everything else     | ArchGDAL.jl
 
 
 ## Backend detection
 
-From filename extension:
+Backend detected in constructors:
 
 ```julia
 # Single array
-rast = Raster("myrasterfile.ext")
+rast = Raster("myraster.tif")    # Will use ArchGDAL.jl
+rast = Raster("myraster.nc")     # Will use NCDatasets
 
 # Multi-array
-st = RasterStack("myrasterstack.ext")
-
-# Series of rasters or stacks may be detectable from folders
-st = RasterSeries("myrasterfolder", Ti(DateTime))
+st = RasterStack("mystack.nc")   # Will use NCDatasets.jl
+st = RasterStack("mystack.grib") # Will use GRIBDatasets.jl
 ```
 
-# Lazy loading
-
-For manipulating larger-than-memory data
+# DiskArrays.jl integration  
 
-## DiskArrays.jl integration  
+For larger-than-memory data
 
-Loads data lazily:
+## Lazy loading
 
 ```julia
 rast = Raster(filename; lazy=true)
 ```
+. . .
 
 Still lazy after broadcasts:
 ```julia
 rast10 = rast .* 10
 ```
+. . .
 
 Reads from disk/network only on `getindex`:
 ```julia
 rast10[X=100 .. 135, Y=20 .. 40]
 ```
 
-Or `read`:
-```julia
-read(rast10)
-```
+## Chunk patterns 
 
-## Change chunk patterns 
 For more efficient lazy reads:
 \
 ```julia
-write("rechunked.tif", raster; chunks=(256, 256))
+write("rechunked.tif", mem_rast; chunks=(X(256), Y(256)))
 ```
 
 ## RasterDataSources.jl integration
@@ -97,6 +140,9 @@ using Rasters, RasterDataSources, ArchGDAL
 bioclim_filename = RasterDataSources.getraster(WorldClim{BioClim}, 5)
 bioclim5 = Raster(bioclim_filename);
 ```
+
+. . .
+
 \
 Or use RasterDataSources.jl syntax directly:
 
@@ -140,27 +186,36 @@ fig
 
 
 # Common GIS methods
+- for working with raster data
+- for using vector/geometry data with raster data
 
 ## Native rasterization engine
 
 - accepts all GeoInterface.jl geometries
 - extremely fast + threaded
 - detailed correctness warnings
 - consistent behaviour and syntax for:
+
+. . .
+
+::: {.nonincremental}
   - `rasterize`
   - `coverage`
   - `mask`
   - `boolmask`/`missingmask`
   - `zonal`
+:::
 
 ## Other common methods
 
+::: {.nonincremental}
 - extract
 - crop/extend
 - trim
 - mosaic
 - aggregate
 - resample
+:::
 
 # Examples
 \
@@ -193,19 +248,15 @@ extract(clim, occ; name=(:bio1, :bio4, :bio7)) |> DataFrame
 
 ## `mask` + `trim`
 
-Mask climate data wth country border and trim:
-
----
-code-annotations: hover
----
+Mask climate rasters wth country border :
 
 ```{julia}
 #| output-location: fragment
 clim = RasterStack(WorldClim{Climate}, (:tmin, :tmax, :prec, :wind); month=July)
 countries = naturalearth("ne_10m_admin_0_countries") |> DataFrame
-norway = subset(countries, :NAME => ByRow(==("Norway"))).geometry
-norway_clim = mask(clim; with=norway)[Y=0 .. 90] |> trim
-Plots.plot(norway_clim; size=(800, 400))
+finland = subset(countries, :NAME => ByRow(==("Finland"))).geometry
+finland_clim = mask(clim; with=finland) |> trim
+Plots.plot(finland_clim; size=(800, 400))
 ```
 
 ## `zonal` statistics
@@ -228,10 +279,15 @@ filtered = subset(countries, :july_maxtemp => ByRow(!isnan))
 sort!(filtered, :july_maxtemp).NAME
 ```
 
-# Thanks!
+## Thanks
 
+Especially to Rasters.jl contributors!
+
+![](https://contrib.rocks/image?repo=rafaqz/Rasters.jl)
+
+\
 Any problems, make github issues at\
 https://github.com/rafaqz/Rasters.jl
 \
 \
-Please include all files in a MWE!
+(Please include all files in a MWE!)