obijupyterhub/Dockerfile.builder

@@ -32,6 +32,7 @@ RUN apt-get update \
        libpng-dev \
        libtiff5-dev \
        libjpeg-dev \
+       libuv1-dev \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 

start-jupyterhub.sh

@@ -20,6 +20,9 @@ FORCE_REBUILD=false
 STOP_SERVER=false
 UPDATE_LECTURES=false
 BUILD_OBIDOC=false
+REBUILD_BUILDER=false
+REBUILD_STUDENT=false
+REBUILD_HUB=false
 
 usage() {
     cat <<EOF
@@ -27,7 +30,10 @@ Usage: ./start-jupyterhub.sh [options]
 
 Options:
   --no-build | --offline   Skip Docker image builds (use existing images)
-  --force-rebuild          Rebuild images without cache
+  --force-rebuild          Rebuild all images without cache
+  --rebuild-builder        Force rebuild the builder image only
+  --rebuild-student        Force rebuild the student image only
+  --rebuild-hub            Force rebuild the JupyterHub image only
   --stop-server            Stop the stack and remove student containers, then exit
   --update-lectures        Rebuild the course website only (no Docker stop/start)
   --build-obidoc           Force rebuild of obidoc documentation
@@ -43,6 +49,9 @@ while [[ $# -gt 0 ]]; do
     case "$1" in
         --no-build|--offline) NO_BUILD=true ;;
         --force-rebuild) FORCE_REBUILD=true ;;
+        --rebuild-builder) REBUILD_BUILDER=true ;;
+        --rebuild-student) REBUILD_STUDENT=true ;;
+        --rebuild-hub) REBUILD_HUB=true ;;
         --stop-server) STOP_SERVER=true ;;
         --update-lectures) UPDATE_LECTURES=true ;;
         --build-obidoc) BUILD_OBIDOC=true ;;
@@ -102,6 +111,7 @@ get_file_timestamp() {
 check_if_image_needs_rebuild() {
     local image_name="$1"
     local dockerfile="$2"
+    local force="${3:-false}"
 
     echo -e "${BLUE}Checking image ${image_name}...${NC}"
 
@@ -111,8 +121,8 @@ check_if_image_needs_rebuild() {
         return 0  # Need to build (image doesn't exist)
     fi
 
-    # If force rebuild, always rebuild
-    if $FORCE_REBUILD; then
+    # If force rebuild (global or per-image), always rebuild
+    if $FORCE_REBUILD || $force; then
         echo -e "${YELLOW}Docker image build is forced.${NC}"
         return 0  # Need to rebuild
     fi
@@ -137,9 +147,9 @@ check_if_image_needs_rebuild() {
 }
 
 build_builder_image() {
-    if check_if_image_needs_rebuild "$BUILDER_IMAGE" "Dockerfile.builder"; then
+    if check_if_image_needs_rebuild "$BUILDER_IMAGE" "Dockerfile.builder" "$REBUILD_BUILDER"; then
         local build_flag=()
-        if $FORCE_REBUILD; then
+        if $FORCE_REBUILD || $REBUILD_BUILDER; then
             build_flag+=(--no-cache)
         fi
 
@@ -178,25 +188,28 @@ build_images() {
         return
     fi
 
-    local build_flag=()
-    if $FORCE_REBUILD; then
-        build_flag+=(--no-cache)
-    fi
-
     # Check and build student image
-    if check_if_image_needs_rebuild "jupyterhub-student:latest" "Dockerfile"; then
+    if check_if_image_needs_rebuild "jupyterhub-student:latest" "Dockerfile" "$REBUILD_STUDENT"; then
+        local student_flag=()
+        if $FORCE_REBUILD || $REBUILD_STUDENT; then
+            student_flag+=(--no-cache)
+        fi
         echo ""
         echo -e "${BLUE}Building student image...${NC}"
-        docker build "${build_flag[@]}" -t jupyterhub-student:latest -f Dockerfile .
+        docker build "${student_flag[@]}" -t jupyterhub-student:latest -f Dockerfile .
     else
         echo -e "${BLUE}Student image is up to date, skipping build.${NC}"
     fi
 
     # Check and build JupyterHub image
-    if check_if_image_needs_rebuild "jupyterhub-hub:latest" "Dockerfile.hub"; then
+    if check_if_image_needs_rebuild "jupyterhub-hub:latest" "Dockerfile.hub" "$REBUILD_HUB"; then
+        local hub_flag=()
+        if $FORCE_REBUILD || $REBUILD_HUB; then
+            hub_flag+=(--no-cache)
+        fi
         echo ""
         echo -e "${BLUE}Building JupyterHub image...${NC}"
-        docker build "${build_flag[@]}" -t jupyterhub-hub:latest -f Dockerfile.hub .
+        docker build "${hub_flag[@]}" -t jupyterhub-hub:latest -f Dockerfile.hub .
     else
         echo -e "${BLUE}JupyterHub image is up to date, skipping build.${NC}"
     fi

tools/install_quarto_deps.R

@@ -1,17 +1,15 @@
 #!/usr/bin/env Rscript
-# Script to dynamically detect and install R dependencies from Quarto files
-# Uses the {attachment} package to scan .qmd files for library()/require() calls
+# Script to dynamically detect and install R dependencies from Quarto files.
+# Scans library()/require() calls and remotes::install_git/github() calls.
 
 args <- commandArgs(trailingOnly = TRUE)
 quarto_dir <- if (length(args) > 0) args[1] else "."
 
-# Target library for installing packages (the mounted volume)
 target_lib <- "/usr/local/lib/R/site-library"
 
 cat("Scanning Quarto files in:", quarto_dir, "\n")
 cat("Target library:", target_lib, "\n")
 
-# Find all .qmd files
 qmd_files <- list.files(
   path = quarto_dir,
   pattern = "\\.qmd$",
@@ -26,34 +24,119 @@ if (length(qmd_files) == 0) {
 
 cat("Found", length(qmd_files), "Quarto files\n")
 
-# Extract dependencies using attachment
-deps <- attachment::att_from_rmds(qmd_files, inline = TRUE)
-
-if (length(deps) == 0) {
-  cat("No R package dependencies detected.\n")
-  quit(status = 0)
+# Extract package names from library()/require() calls
+extract_cran_packages <- function(files) {
+  pattern <- "(?:library|require)\\s*\\(\\s*['\"]?([A-Za-z0-9._]+)['\"]?"
+  pkgs <- character(0)
+  for (f in files) {
+    lines <- tryCatch(readLines(f, warn = FALSE), error = function(e) character(0))
+    m <- regmatches(lines, gregexpr(pattern, lines, perl = TRUE))
+    hits <- unlist(m)
+    if (length(hits) > 0) {
+      extracted <- sub(
+        "(?:library|require)\\s*\\(\\s*['\"]?([A-Za-z0-9._]+)['\"]?.*",
+        "\\1", hits, perl = TRUE
+      )
+      pkgs <- c(pkgs, extracted)
+    }
+  }
+  unique(pkgs)
 }
 
-cat("\nDetected R packages:\n")
-cat(paste(" -", deps, collapse = "\n"), "\n\n")
+# Extract git/github URLs from remotes::install_git/github() calls
+extract_git_packages <- function(files) {
+  # Matches remotes::install_git('url') or remotes::install_github('user/repo')
+  pattern <- "remotes::install_(git|github)\\s*\\(\\s*['\"]([^'\"]+)['\"]"
+  result <- list()
+  for (f in files) {
+    lines <- tryCatch(readLines(f, warn = FALSE), error = function(e) character(0))
+    text <- paste(lines, collapse = "\n")
+    m <- gregexpr(pattern, text, perl = TRUE)
+    hits <- regmatches(text, m)[[1]]
+    for (hit in hits) {
+      type <- sub("remotes::install_(git|github).*", "\\1", hit, perl = TRUE)
+      url  <- sub("remotes::install_(?:git|github)\\s*\\(\\s*['\"]([^'\"]+)['\"].*",
+                  "\\1", hit, perl = TRUE)
+      result[[length(result) + 1]] <- list(type = type, url = url)
+    }
+  }
+  result
+}
+
+cran_deps <- extract_cran_packages(qmd_files)
+git_deps  <- extract_git_packages(qmd_files)
+
+# Quarto's implicit runtime dependencies — must be in target_lib (the persistent
+# volume), not just somewhere in libPaths, because Quarto spawns its own R session.
+quarto_required <- c("rmarkdown", "knitr")
+if (length(git_deps) > 0) quarto_required <- c(quarto_required, "remotes")
+
+cat("\nDetected CRAN packages:\n")
+cat(paste(" -", unique(c(quarto_required, cran_deps)), collapse = "\n"), "\n")
+
+if (length(git_deps) > 0) {
+  cat("\nDetected git/github packages:\n")
+  for (d in git_deps) cat(" -", d$type, ":", d$url, "\n")
+}
+cat("\n")
+
+# --- Install CRAN packages ---
 
-# Filter out base R packages that are always available
 base_pkgs <- rownames(installed.packages(priority = "base"))
-deps <- setdiff(deps, base_pkgs)
 
-# Check which packages are not installed
+# quarto_required: check only in target_lib so they are guaranteed to be there
+installed_in_target <- rownames(installed.packages(lib.loc = target_lib))
+quarto_missing <- setdiff(quarto_required, c(base_pkgs, installed_in_target))
+
+# other deps: check anywhere in libPaths (they just need to be loadable)
+cran_deps  <- setdiff(cran_deps, c(base_pkgs, quarto_required))
 installed  <- rownames(installed.packages())
-to_install <- setdiff(deps, installed)
+to_install <- unique(c(quarto_missing, setdiff(cran_deps, installed)))
 
 if (length(to_install) == 0) {
-  cat("All required packages are already installed.\n")
+  cat("All CRAN packages already installed.\n")
 } else {
-  cat("Installing missing packages:", paste(to_install, collapse = ", "), "\n\n")
-  install.packages(
-    to_install,
-    lib = target_lib,
-    repos = "https://cloud.r-project.org/",
-    dependencies = TRUE
+  cat("Installing CRAN packages:", paste(to_install, collapse = ", "), "\n\n")
+  failed <- character(0)
+  for (pkg in to_install) {
+    result <- tryCatch({
+      withCallingHandlers(
+        install.packages(pkg, lib = target_lib, repos = "https://cloud.r-project.org/",
+                         dependencies = TRUE, quiet = FALSE),
+        warning = function(w) {
+          if (grepl("not available", conditionMessage(w))) invokeRestart("muffleWarning")
+        }
       )
-  cat("\nPackage installation complete.\n")
+      if (!requireNamespace(pkg, quietly = TRUE)) "unavailable" else "ok"
+    }, error = function(e) "error")
+
+    if (result %in% c("unavailable", "error")) {
+      cat("  [SKIP]", pkg, "- not available on CRAN\n")
+      failed <- c(failed, pkg)
+    } else {
+      cat("  [OK]", pkg, "\n")
+    }
+  }
+  if (length(failed) > 0)
+    cat("\nNot installed (not on CRAN):", paste(failed, collapse = ", "), "\n")
 }
+
+# --- Install git/github packages ---
+
+if (length(git_deps) > 0) {
+  cat("\nInstalling git/github packages...\n")
+  for (d in git_deps) {
+    tryCatch({
+      if (d$type == "git") {
+        remotes::install_git(d$url, lib = target_lib, upgrade = "never")
+      } else {
+        remotes::install_github(d$url, lib = target_lib, upgrade = "never")
+      }
+      cat("  [OK]", d$url, "\n")
+    }, error = function(e) {
+      cat("  [FAIL]", d$url, "-", conditionMessage(e), "\n")
+    })
+  }
+}
+
+cat("\nDependency installation complete.\n")