Merge pull request #27 from nptscot/use_anime_join

add .devcontainer and replace rnet_merge with anime_join

.devcontainer/devcontainer.json

@@ -0,0 +1,37 @@
+// For format details, see https://aka.ms/devcontainer.json. For config options, see the
+// README at: https://github.com/devcontainers/templates/tree/main/src/debian
+{
+	"name": "Debian",
+	// Or use a Dockerfile or Docker Compose file. More info: https://containers.dev/guide/dockerfile
+	"image": "ghcr.io/geocompx/docker:rust",
+	"customizations": {
+		// Configure properties specific to VS Code.
+		"vscode": {
+		  // Add the IDs of extensions you want installed when the container is created.
+		  "extensions": ["reditorsupport.r",
+		  "GitHub.copilot-chat",
+		  "GitHub.copilot-labs",
+		  "GitHub.copilot",
+		  "yzhang.markdown-all-in-one",
+		  "quarto.quarto"
+		   ]
+		}
+	  },
+
+	// Features to add to the dev container. More info: https://containers.dev/features.
+	// "features": {},
+
+	// Use 'forwardPorts' to make a list of ports inside the container available locally.
+	// "forwardPorts": [],
+
+	// Configure tool-specific properties.
+	// "customizations": {},
+
+	// Uncomment to connect as root instead. More info: https://aka.ms/dev-containers-non-root.
+	// "remoteUser": "root"
+
+	// Run the script install-additional-dependencies.sh:
+	"postCreateCommand": "apt-get update && apt-get install -y dos2unix && dos2unix ./.devcontainer/postCreateCommand.sh && bash ./.devcontainer/postCreateCommand.sh"
+
+}
+

.devcontainer/postCreateCommand.sh

@@ -0,0 +1,47 @@
+# Check the Linux distro we're running:
+cat /etc/os-release
+
+# Add cargo to the path both temporarily and permanently:
+export PATH="$HOME/.cargo/bin:$PATH"
+echo 'export PATH="$HOME/.cargo/bin:$PATH"' >> ~/.profile
+
+# Ensure cargo command is available
+command -v cargo
+
+# Install odjitter using cargo
+cargo install --git https://github.com/dabreegster/odjitter --rev 32fb58bf7f0d68afd3b76b88cf6b1272c5c66828
+# Add local instance of odjitter to the /usr/local/bin directory:
+which odjitter
+sudo ln -s ~/.cargo/bin/odjitter /usr/local/bin/odjitter
+
+# Ensure R is installed and execute the R script
+Rscript code/install.R
+
+# Ensure apt repository is up-to-date and install Python packages
+apt-get update
+apt-get install -y software-properties-common python3 python3-pip
+
+# Install Python dependencies:
+pip install -r requirements.txt
+
+# Clone and install tippecanoe if not already installed
+cd /tmp
+if [ ! -d "tippecanoe" ]; then
+    git clone https://github.com/felt/tippecanoe.git
+fi
+cd tippecanoe
+make -j$(nproc)
+sudo make install
+tippecanoe --version
+
+# Install git and GitHub CLI (gh)
+apt-get install -y git
+apt-get install -y gh
+
+# Configure git settings
+git config --global core.autocrlf input
+git config --global core.fileMode false
+git update-index --refresh
+
+# Make sure there's a newline at the end of the script
+echo "Script execution completed successfully."

R/anime_join.R

@@ -0,0 +1,68 @@
+# Define a helper function that runs anime and aggregates an attribute.
+anime_join = function(source_data,
+                       target_data,
+                       attribute,      # character name of the attribute to aggregate
+                       new_name,       # character name for the output column
+                       agg_fun = sum,  # aggregation function (e.g., sum, mean, max)
+                       weights,        # character vector of weight columns (e.g., "target_weighted")
+                       angle_tolerance = 35,
+                       distance_tolerance = 15) {
+                        library(dplyr)
+  library(sf)
+  library(rlang)
+  library(purrr)
+  library(anime)  
+
+  get_fun_name = function(fn) {
+    deparse(substitute(fn))
+  }
+
+  cat("Aggregating attribute:", attribute, "using function:", get_fun_name(sum), "\n")
+
+  # Run anime matching between source and target
+  matches = anime::anime(
+    source = source_data,
+    target = target_data,
+    angle_tolerance = angle_tolerance,
+    distance_tolerance = distance_tolerance
+  )
+  matches_df = anime::get_matches(matches)
+
+  # Join the source data (with a source_id) to the matches.
+  net_source_matches = source_data %>%
+    mutate(source_id = row_number()) %>%
+    st_drop_geometry() %>%
+    left_join(matches_df, by = "source_id")
+
+  # Convert the attribute and weight names to symbols.
+  attr_sym = sym(attribute)
+  weight_syms = syms(weights)
+
+  # For max aggregatio
+  if (identical(agg_fun, max)) {
+    mult_expr = purrr::reduce(weight_syms, function(acc, w) expr((!!acc)), .init = attr_sym)
+  } else {
+    mult_expr = purrr::reduce(weight_syms, function(acc, w) expr((!!acc) * (!!w)), .init = attr_sym)
+  }
+
+  # Group by target id (assumed to be provided by anime as "target_id") and aggregate.
+  net_target_aggregated = net_source_matches %>%
+    group_by(row_number = target_id) %>%
+    summarise(!!sym(new_name) := agg_fun(!!mult_expr, na.rm = TRUE))
+
+  # For max, we also want to round the final result.
+  if (identical(agg_fun, max)) {
+    net_target_aggregated = net_target_aggregated %>%
+      mutate(!!sym(new_name) := round(!!sym(new_name)))
+  }
+
+  # Join the aggregated values back to the target data.
+  net_target_joined = target_data %>%
+    mutate(row_number = row_number()) %>%
+    st_drop_geometry() %>%
+    select(-any_of(new_name)) %>%   # Remove the original column (if it exists)
+    left_join(net_target_aggregated, by = "row_number") %>%
+    select(id, !!sym(new_name))
+
+  return(net_target_joined)
+}

R/corenet.R

@@ -33,8 +33,7 @@ utils::globalVariables(c("edge_paths", "influence_network", "all_fastest_bicycle
 #' NPT_demo_3km = sf::st_set_crs(NPT_demo_3km, 27700)
 #' base_network = sf::st_transform(os_edinburgh_demo_3km, crs = 27700)
 #' influence_network = sf::st_transform(NPT_demo_3km, crs = 27700)
-#' target_zone = zonebuilder::zb_zone("Edinburgh", n_circles = 2) |>
-#'                sf::st_transform(crs = "EPSG:27700")
+#' target_zone = zonebuilder::zb_zone("Edinburgh", n_circles = 3) |> sf::st_transform(crs = "EPSG:27700")        
 #'
 #' # Prepare the cohesive network
 #' OS_NPT_demo = cohesive_network_prep( base_network = base_network, 
@@ -46,7 +45,7 @@ utils::globalVariables(c("edge_paths", "influence_network", "all_fastest_bicycle
 #'
 
 
-cohesive_network_prep = function(base_network, influence_network, target_zone, crs = "EPSG:27700",  key_attribute = "road_function", attribute_values = c("A Road", "B Road", "Minor Road"), use_stplanr = TRUE) {
+cohesive_network_prep = function(base_network, influence_network, target_zone, crs = "EPSG:27700",  key_attribute = "road_function", attribute_values = c("A Road", "B Road", "Minor Road"), use_stplanr = FALSE) {
     base_network = sf::st_transform(base_network, crs)    
     influence_network = sf::st_transform(influence_network, crs)
     target_zone = sf::st_transform(target_zone, crs)
@@ -62,35 +61,69 @@ cohesive_network_prep = function(base_network, influence_network, target_zone, c
                         dplyr::filter(!!rlang::sym(key_attribute) %in% attribute_values ) |> 
                         sf::st_transform(crs) |> 
                         sf::st_zm()
-    if (use_stplanr) {
-      # Assign functions for data aggregation based on network attribute values 
-      name_list = names(NPT_zones)
+    # if (use_stplanr) {
+    #   # Assign functions for data aggregation based on network attribute values 
+    #   name_list = names(NPT_zones)
 
-      funs = list()
-      for (name in name_list) {
-        if (name == "geometry") {
-          next  # Correctly skip the current iteration if the name is "geometry"
-        } else if (name %in% c("gradient", "quietness")) {
-          funs[[name]] = mean  # Assign mean function for specified fields
-        } else {
-          funs[[name]] = mean  # Assign sum function for all other fields
-        }
-      }
+    #   funs = list()
+    #   for (name in name_list) {
+    #     if (name == "geometry") {
+    #       next  # Correctly skip the current iteration if the name is "geometry"
+    #     } else if (name %in% c("gradient", "quietness")) {
+    #       funs[[name]] = mean  # Assign mean function for specified fields
+    #     } else {
+    #       funs[[name]] = mean  # Assign sum function for all other fields
+    #     }
+    #   }
 
-      # Merge road networks with specified parameters
-      filtered_OS_NPT_zones = stplanr::rnet_merge(filtered_OS_zones, NPT_zones, dist = 10, funs = funs, segment_length = 20, max_angle_diff = 10)
-    } else {
-      print("Using sf::st_join")
-      filtered_OS_NPT_zones = sf::st_join(filtered_OS_zones, NPT_zones, join = sf::st_intersects)
-    }
+    #   # Merge road networks with specified parameters
+    #   filtered_OS_NPT_zones = stplanr::rnet_merge(filtered_OS_zones, NPT_zones, dist = 10, funs = funs, segment_length = 20, max_angle_diff = 10)
+
+    # } else {
+    #   print("Using sf::st_join")
+    #   filtered_OS_NPT_zones = sf::st_join(filtered_OS_zones, NPT_zones, join = sf::st_intersects)
+    # }
 
+    NPT_zones = sf::st_cast(NPT_zones, "LINESTRING")
+    filtered_OS_zones = sf::st_cast(filtered_OS_zones, "LINESTRING")
+    NPT_zones$id = 1:nrow(NPT_zones)
+    filtered_OS_zones$id = 1:nrow(filtered_OS_zones)
+
+    params = list(
+                  list(
+                    source = NPT_zones,
+                    target = filtered_OS_zones,
+                    attribute = "all_fastest_bicycle_go_dutch",
+                    new_name = "all_fastest_bicycle_go_dutch",
+                    agg_fun = sum,
+                    weights = c("target_weighted")
+                  )
+                )
+
+    results_list = purrr::map(params, function(p) {
+      anime_join(
+        source_data = p$source,
+        target_data = p$target,
+        attribute = p$attribute,
+        new_name = p$new_name,
+        agg_fun = p$agg_fun,
+        weights = p$weights,
+        angle_tolerance = 35,
+        distance_tolerance = 15
+      )
+    })
+
+
+    filtered_OS_NPT_zones = reduce(results_list, function(x, y) {
+      left_join(x, y, by = "id")
+    }, .init = filtered_OS_zones)
+
     filtered_OS_NPT_zones = filtered_OS_NPT_zones |>
                             dplyr::mutate(dplyr::across(dplyr::where(is.numeric), as.integer))
 
-
-  print("Finished preparing the network data")
-
-  return(filtered_OS_NPT_zones)
+    print("Finished preparing the network data")
+
+    return(filtered_OS_NPT_zones)
 }
 
 
@@ -160,9 +193,11 @@ corenet = function(influence_network, cohesive_base_network, target_zone, key_at
 
     # Perform DBSCAN clustering
     coordinates = sf::st_coordinates(centroids)
-    clusters = dbscan::dbscan(coordinates, eps = 18, minPts = 1)
-    centroids$cluster = clusters$cluster
-    unique_centroids = centroids[!duplicated(centroids$cluster), ]   
+    coordinates_clean = coordinates[complete.cases(coordinates), ]
+    clusters = dbscan::dbscan(coordinates_clean, eps = 18, minPts = 1)
+    centroids_clean = centroids[complete.cases(coordinates), ]
+    centroids_clean$cluster = clusters$cluster
+    unique_centroids = centroids_clean[!duplicated(centroids_clean$cluster), ]   
 
     # create a buffer of 10 meters around the cohesive_base_network
     cohesive_base_network_buffer = sf::st_buffer(cohesive_base_network, dist = 20)

R/pkgs.R

@@ -0,0 +1,27 @@
+get_pkgs = function() {
+   c(
+  "crew", # For targets with workers
+  "collapse", # Needed for bind_sf
+  "cyclestreets", # For routing
+  "geojsonsf", # For converting geojson to sf
+  "geos", # For geometric operations
+  "gert", # For interactive with git
+  "glue", # For string interpolation
+  "lubridate", # For working with dates and times
+  "lwgeom", # For working with spatial data
+  "nngeo", # Nearest neighbour functions
+  "osmextract", # For extracting OpenStreetMap data
+  "pct", # PCT interface
+  "remotes", # For installing packages from remote sources
+  "sf", # For working with spatial data
+  "simodels", # For spatial interaction models
+  "snakecase", # For converting strings to snake case
+  "stplanr", # For sustainable transport planning
+  "targets", # For managing targets in a workflow
+  "tidyverse", # Includes dplyr, ggplot2, tidyr, stringr etc.
+  "zonebuilder", # For creating zones for spatial analysis
+  "iterators", # For creating iterators
+  "doParallel", # For parallel processing
+  "httr"
+  )
+}

code/install.R

@@ -0,0 +1,28 @@
+# Run as part of _targets.R to install packages
+# Do you want to reinstall github packages, set to TRUE for first run
+update_github_packages = TRUE
+options(Ncpus = 4)
+source("R/pkgs.R")
+pkgs = get_pkgs()
+remotes::install_cran(pkgs)
+
+# rsgeo
+install.packages(
+  'rsgeo', 
+  repos = c('https://josiahparry.r-universe.dev', 'https://cloud.r-project.org')
+)
+
+# Repeated builds can it GitHub API limit, set to TRUE in _targets.R to check for package updates
+if (update_github_packages) {
+  remotes::install_dev("cyclestreets")
+  remotes::install_github("dabreegster/odjitter", subdir = "r")
+  remotes::install_github("robinlovelace/ukboundaries")
+  remotes::install_github("robinlovelace/simodels")
+  remotes::install_version("rgeos", version = "0.6-3")
+  remotes::install_dev("od")
+  remotes::install_dev("osmextract")
+  remotes::install_github("nptscot/corenet")
+  remotes::install_github("nptscot/osmactive")
+  install.packages("pak")
+  pak::pak("JosiahParry/anime/r")
+}

requirements.txt

@@ -0,0 +1,3 @@
+geopandas
+shapely
+pandas