## ----include = FALSE----------------------------------------------------------
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.width = 7,
  fig.height = 5,
  fig.align = "center",
  message = FALSE,
  warning = FALSE
)

## ----setup--------------------------------------------------------------------
library(mapycusmaximus)
library(sf)
library(ggplot2)

# Use a minimal theme for clean visualizations
theme_set(theme_minimal())

## ----first-fisheye------------------------------------------------------------
# Examine the data
data(vic)
vic

## ----simple-fisheye, fig.cap = "A basic fisheye transformation of Victoria's LGAs"----
# Apply fisheye transformation
vic_warped <- sf_fisheye(
  vic,
  r_in = 0.3,        # Focus radius
  r_out = 0.6,       # Glue boundary
  zoom_factor = 2,   # Magnification strength
  squeeze_factor = 0.35
)

# Visualize with ggplot2
ggplot(vic_warped) +
  geom_sf(fill = "grey90", color = "white", linewidth = 0.3) +
  labs(title = "Victoria LGAs with Default Fisheye")

## ----center-geometry, fig.cap = "Fisheye centered on Melbourne CBD"-----------
# Extract Melbourne CBD as the focus
melbourne <- vic[vic$LGA_NAME == "MELBOURNE", ]

vic_melbourne <- sf_fisheye(
  vic,
  center = melbourne,      # Centroid becomes the warp center
  r_in = 0.34,
  r_out = 0.60,
  zoom_factor = 15,
  squeeze_factor = 0.35
)

ggplot() +
  geom_sf(data = vic_melbourne, fill = "grey92", color = "white", linewidth = 0.2) +
  geom_sf(data = melbourne, fill = NA, color = "tomato", linewidth = 0.8) +
  labs(title = "Melbourne CBD Magnified",
       subtitle = "Focus defined by Melbourne LGA geometry")

## ----center-lonlat, fig.cap = "Fisheye using lon/lat coordinates"-------------
# Melbourne CBD coordinates (WGS84)
melb_coords <- c(144.9631, -37.8136)

vic_coords <- sf_fisheye(
  vic,
  center = melb_coords,
  center_crs = "EPSG:4326",   # Explicitly state CRS
  r_in = 0.30,
  r_out = 0.55,
  zoom_factor = 12,
  squeeze_factor = 0.30
)

ggplot(vic_coords) +
  geom_sf(fill = "grey92", color = "white", linewidth = 0.2) +
  labs(title = "Center Specified as Lon/Lat",
       subtitle = "Coordinates: 144.96°E, 37.81°S")

## ----center-projected, eval = FALSE-------------------------------------------
# # Example: coordinates in the working CRS (meters)
# vic_projected <- sf_fisheye(
#   vic,
#   cx = 321000,      # Easting (meters)
#   cy = 5813000,     # Northing (meters)
#   r_in = 0.3,
#   r_out = 0.6,
#   zoom_factor = 10,
#   squeeze_factor = 0.35
# )

## ----parameter-demo-radii, fig.cap = "Effect of different radius settings", fig.height = 6----
# Small focus, narrow glue
vic_tight <- sf_fisheye(vic, center = melbourne, r_in = 0.2, r_out = 0.3, 
                        zoom_factor = 8, squeeze_factor = 0.35)

# Large focus, wide glue
vic_wide <- sf_fisheye(vic, center = melbourne, r_in = 0.4, r_out = 0.7, 
                       zoom_factor = 8, squeeze_factor = 0.35)

p1 <- ggplot(vic_tight) + 
  geom_sf(fill = "grey90", color = "white", linewidth = 0.2) +
  labs(title = "Tight Focus (r_in=0.2, r_out=0.3)")

p2 <- ggplot(vic_wide) + 
  geom_sf(fill = "grey90", color = "white", linewidth = 0.2) +
  labs(title = "Wide Focus (r_in=0.4, r_out=0.7)")

# Display side by side (requires patchwork or cowplot)
# p1 + p2
p1
p2

## ----parameter-demo-zoom, fig.cap = "Effect of zoom factor", fig.height = 6----
# Gentle zoom
vic_gentle <- sf_fisheye(vic, center = melbourne, r_in = 0.3, r_out = 0.5, 
                         zoom_factor = 3, squeeze_factor = 0.35)

# Aggressive zoom
vic_aggressive <- sf_fisheye(vic, center = melbourne, r_in = 0.3, r_out = 0.5, 
                             zoom_factor = 20, squeeze_factor = 0.35)

ggplot(vic_gentle) + 
  geom_sf(fill = "grey90", color = "white", linewidth = 0.2) +
  labs(title = "Gentle Magnification (zoom = 3)")

ggplot(vic_aggressive) + 
  geom_sf(fill = "grey90", color = "white", linewidth = 0.2) +
  labs(title = "Strong Magnification (zoom = 20)")

## ----parameter-demo-squeeze, fig.height = 6-----------------------------------
# Minimal squeeze (wider glue transition)
vic_loose <- sf_fisheye(vic, center = melbourne, r_in = 0.3, r_out = 0.5, 
                        zoom_factor = 8, squeeze_factor = 0.1)

# Strong squeeze (narrow glue transition)
vic_tight_squeeze <- sf_fisheye(vic, center = melbourne, r_in = 0.3, r_out = 0.5, 
                                zoom_factor = 8, squeeze_factor = 0.8)

ggplot(vic_loose) + 
  geom_sf(fill = "grey90", color = "white", linewidth = 0.2) +
  labs(title = "Loose Squeeze (0.1)")

ggplot(vic_tight_squeeze) + 
  geom_sf(fill = "grey90", color = "white", linewidth = 0.2) +
  labs(title = "Tight Squeeze (0.8)")

## ----multi-layer, fig.cap = "Multiple aligned layers with fisheye transformation"----
# Create centroids as a point layer
centroids <- st_centroid(vic)

# Add a layer identifier to each
vic_layer <- vic |> 
  dplyr::mutate(layer = "polygon")

centroids_layer <- centroids |> 
  dplyr::mutate(layer = "centroid")

# Combine layers before transformation
both_layers <- rbind(
  vic_layer[, c("LGA_NAME", "geometry", "layer")],
  centroids_layer[, c("LGA_NAME", "geometry", "layer")]
)

# Apply fisheye once to combined data
both_warped <- sf_fisheye(
  both_layers,
  center = melbourne,
  r_in = 0.34,
  r_out = 0.60,
  zoom_factor = 12,
  squeeze_factor = 0.35
)

# Separate for plotting
polygons_warped <- both_warped[both_warped$layer == "polygon", ]
points_warped <- both_warped[both_warped$layer == "centroid", ]

# Plot together
ggplot() +
  geom_sf(data = polygons_warped, fill = "grey92", color = "white", 
          linewidth = 0.2) +
  geom_sf(data = points_warped, color = "#2b6cb0", size = 1.2, alpha = 0.7) +
  labs(title = "Aligned Layers: Polygons and Centroids",
       subtitle = "Transformed together to ensure perfect alignment")

## ----crs-handling-------------------------------------------------------------
# Create data in WGS84
vic_lonlat <- st_transform(vic, "EPSG:4326")
st_crs(vic_lonlat)$proj4string

# Apply fisheye - auto-projects to GDA2020/MGA55 for Victoria
vic_auto <- sf_fisheye(
  vic_lonlat,
  center = melbourne,
  r_in = 0.3,
  r_out = 0.5,
  zoom_factor = 10,
  squeeze_factor = 0.35
)

# Original CRS is restored
st_crs(vic_auto)$proj4string

## ----custom-crs, eval = FALSE-------------------------------------------------
# vic_custom <- sf_fisheye(
#   vic_lonlat,
#   center = melbourne,
#   target_crs = "EPSG:3111",  # VicGrid
#   r_in = 0.3,
#   r_out = 0.5,
#   zoom_factor = 10,
#   squeeze_factor = 0.35
# )

## ----diagnostics, fig.cap = "Understanding the FGC transformation with a test grid", fig.width = 8, fig.height = 4----
# Create a regular grid
grid <- create_test_grid(range = c(-1, 1), spacing = 0.1)

# Apply transformation
warped <- fisheye_fgc(
  grid,
  r_in = 0.34,
  r_out = 0.5,
  zoom_factor = 1.3,
  squeeze_factor = 0.5
)

# Visualize the transformation
plot_fisheye_fgc(grid, warped, r_in = 0.34, r_out = 0.5)

## ----real-world-example, fig.cap = "Complete workflow: Metropolitan Melbourne in Victorian context", fig.width = 8----
# 1. Define the metropolitan region
metro_lgas <- c("MELBOURNE", "PORT PHILLIP", "STONNINGTON", "YARRA", 
                "MARIBYRNONG", "MOONEE VALLEY", "BOROONDARA", 
                "GLEN EIRA", "BAYSIDE")

metro_region <- vic[vic$LGA_NAME %in% metro_lgas, ]
metro_center <- st_union(metro_region) |> st_centroid()

# 2. Add a population indicator (example)
vic_pop <- vic |> 
  dplyr::mutate(is_metro = LGA_NAME %in% metro_lgas)

# 3. Apply fisheye
vic_focused <- sf_fisheye(
  vic_pop,
  center = metro_center,
  r_in = 0.25,
  r_out = 0.40,
  zoom_factor = 12,
  squeeze_factor = 0.35
)

# 4. Create publication-ready plot
ggplot(vic_focused) +
  geom_sf(aes(fill = is_metro), color = "white", linewidth = 0.2) +
  scale_fill_manual(
    values = c("TRUE" = "#d95f02", "FALSE" = "grey85"),
    labels = c("Metropolitan", "Regional"),
    name = NULL
  ) +
  theme_minimal() +
  theme(
    legend.position = c(0.85, 0.15),
    legend.background = element_rect(fill = "white", color = "grey70"),
    panel.grid = element_blank()
  ) +
  labs(
    title = "Metropolitan Melbourne with Regional Context",
    subtitle = "Fisheye magnification preserves spatial relationships",
    caption = "Transformation: r_in = 0.25, r_out = 0.40, zoom = 12"
  )

## ----best-practices-params, eval = FALSE--------------------------------------
# # Start here
# sf_fisheye(data, r_in = 0.3, r_out = 0.5, zoom_factor = 5, squeeze_factor = 0.35)
# 
# # Too distorted? Reduce zoom or widen glue
# sf_fisheye(data, r_in = 0.3, r_out = 0.6, zoom_factor = 3, squeeze_factor = 0.35)
# 
# # Need more magnification? Increase zoom gradually
# sf_fisheye(data, r_in = 0.3, r_out = 0.5, zoom_factor = 10, squeeze_factor = 0.35)

## ----best-practices-layers, eval = FALSE--------------------------------------
# # Good: Single transformation
# combined <- rbind(layer1, layer2, layer3)
# warped <- sf_fisheye(combined, ...)
# 
# # Avoid: Separate transformations
# layer1_warped <- sf_fisheye(layer1, ...)  # Different normalization
# layer2_warped <- sf_fisheye(layer2, ...)  # May not align perfectly

## ----best-practices-reproducibility, eval = FALSE-----------------------------
# # Explicit and reproducible
# result <- sf_fisheye(
#   data,
#   center = st_point(c(144.9631, -37.8136)),
#   center_crs = "EPSG:4326",
#   target_crs = "EPSG:7855",
#   r_in = 0.34,
#   r_out = 0.60,
#   zoom_factor = 12,
#   squeeze_factor = 0.35,
#   preserve_aspect = TRUE,
#   revolution = 0
# )

## ----performance, eval = FALSE------------------------------------------------
# # Pre-process large data
# data_clean <- data |>
#   filter(!st_is_empty(geometry)) |>
#   st_simplify(dTolerance = 100)  # Adjust tolerance as needed
# 
# # Transform once
# data_warped <- sf_fisheye(data_clean, ...)