Optimize the effect and speed of scanning

bin/purge.sh

@@ -47,21 +47,119 @@ start_purge() {
         printf '\033[2J\033[H'
     fi
     printf '\n'
-    echo -e "${PURPLE_BOLD}Purge Project Artifacts${NC}"
 
     # Initialize stats file in user cache directory
     local stats_dir="${XDG_CACHE_HOME:-$HOME/.cache}/mole"
     ensure_user_dir "$stats_dir"
     ensure_user_file "$stats_dir/purge_stats"
     ensure_user_file "$stats_dir/purge_count"
+    ensure_user_file "$stats_dir/purge_scanning"
     echo "0" > "$stats_dir/purge_stats"
     echo "0" > "$stats_dir/purge_count"
+    echo "" > "$stats_dir/purge_scanning"
 }
 
 # Perform the purge
 perform_purge() {
+    local stats_dir="${XDG_CACHE_HOME:-$HOME/.cache}/mole"
+    local monitor_pid=""
+    
+    # Cleanup function
+    cleanup_monitor() {
+        # Remove scanning file to stop monitor
+        rm -f "$stats_dir/purge_scanning" 2> /dev/null || true
+        
+        if [[ -n "$monitor_pid" ]]; then
+            kill "$monitor_pid" 2> /dev/null || true
+            wait "$monitor_pid" 2> /dev/null || true
+        fi
+        if [[ -t 1 ]]; then
+            printf '\r\033[K\n\033[K\033[A'
+        fi
+    }
+    
+    # Set up trap for cleanup
+    trap cleanup_monitor INT TERM
+    
+    # Show scanning with spinner on same line as title
+    if [[ -t 1 ]]; then
+        # Print title first
+        printf '%s' "${PURPLE_BOLD}Purge Project Artifacts${NC} "
+        
+        # Start background monitor with ASCII spinner
+        (
+            local spinner_chars="|/-\\"
+            local spinner_idx=0
+            local last_path=""
+            
+            # Set up trap to exit cleanly
+            trap 'exit 0' INT TERM
+            
+            # Function to truncate path in the middle
+            truncate_path() {
+                local path="$1"
+                local max_len=80
+                
+                if [[ ${#path} -le $max_len ]]; then
+                    echo "$path"
+                    return
+                fi
+                
+                # Calculate how much to show on each side
+                local side_len=$(( (max_len - 3) / 2 ))
+                local start="${path:0:$side_len}"
+                local end="${path: -$side_len}"
+                echo "${start}...${end}"
+            }
+            
+            while [[ -f "$stats_dir/purge_scanning" ]]; do
+                local current_path=$(cat "$stats_dir/purge_scanning" 2> /dev/null || echo "")
+                local display_path=""
+                
+                if [[ -n "$current_path" ]]; then
+                    display_path="${current_path/#$HOME/~}"
+                    display_path=$(truncate_path "$display_path")
+                    last_path="$display_path"
+                elif [[ -n "$last_path" ]]; then
+                    display_path="$last_path"
+                fi
+                
+                # Get current spinner character
+                local spin_char="${spinner_chars:$spinner_idx:1}"
+                spinner_idx=$(( (spinner_idx + 1) % ${#spinner_chars} ))
+                
+                # Show title on first line, spinner and scanning info on second line
+                if [[ -n "$display_path" ]]; then
+                    printf '\r%s\n%s %sScanning %s\033[K\033[A' \
+                        "${PURPLE_BOLD}Purge Project Artifacts${NC}" \
+                        "${BLUE}${spin_char}${NC}" \
+                        "${GRAY}" "$display_path"
+                else
+                    printf '\r%s\n%s %sScanning...\033[K\033[A' \
+                        "${PURPLE_BOLD}Purge Project Artifacts${NC}" \
+                        "${BLUE}${spin_char}${NC}" \
+                        "${GRAY}"
+                fi
+                
+                sleep 0.05
+            done
+            exit 0
+        ) &
+        monitor_pid=$!
+    else
+        echo -e "${PURPLE_BOLD}Purge Project Artifacts${NC}"
+    fi
+    
     clean_project_artifacts
     local exit_code=$?
+    
+    # Clean up
+    trap - INT TERM
+    cleanup_monitor
+    
+    if [[ -t 1 ]]; then
+        echo -e "${PURPLE_BOLD}Purge Project Artifacts${NC}"
+    fi
 
     # Exit codes:
     # 0 = success, show summary
@@ -79,15 +177,11 @@ perform_purge() {
     local total_size_cleaned=0
     local total_items_cleaned=0
 
-    # Read stats from user cache directory
-    local stats_dir="${XDG_CACHE_HOME:-$HOME/.cache}/mole"
-
     if [[ -f "$stats_dir/purge_stats" ]]; then
         total_size_cleaned=$(cat "$stats_dir/purge_stats" 2> /dev/null || echo "0")
         rm -f "$stats_dir/purge_stats"
     fi
 
-    # Read count
     if [[ -f "$stats_dir/purge_count" ]]; then
         total_items_cleaned=$(cat "$stats_dir/purge_count" 2> /dev/null || echo "0")
         rm -f "$stats_dir/purge_count"

lib/clean/project.sh

@@ -45,7 +45,7 @@ readonly PURGE_TARGETS=(
 readonly MIN_AGE_DAYS=7
 # Scan depth defaults (relative to search root).
 readonly PURGE_MIN_DEPTH_DEFAULT=2
-readonly PURGE_MAX_DEPTH_DEFAULT=8
+readonly PURGE_MAX_DEPTH_DEFAULT=4
 # Search paths (default, can be overridden via config file).
 readonly DEFAULT_PURGE_SEARCH_PATHS=(
     "$HOME/www"
@@ -339,6 +339,11 @@ scan_purge_targets() {
     if [[ ! -d "$search_path" ]]; then
         return
     fi
+
+    # Update current scanning path
+    local stats_dir="${XDG_CACHE_HOME:-$HOME/.cache}/mole"
+    echo "$search_path" > "$stats_dir/purge_scanning" 2> /dev/null || true
+
     if command -v fd > /dev/null 2>&1; then
         # Escape regex special characters in target names for fd patterns
         local escaped_targets=()
@@ -356,28 +361,39 @@ scan_purge_targets() {
             "--type" "d"
             "--min-depth" "$min_depth"
             "--max-depth" "$max_depth"
-            "--threads" "4"
+            "--threads" "8"
             "--exclude" ".git"
             "--exclude" "Library"
             "--exclude" ".Trash"
             "--exclude" "Applications"
         )
-        fd "${fd_args[@]}" "$pattern" "$search_path" 2> /dev/null | while IFS= read -r item; do
+        # Write to temp file first, then filter - more efficient than piping
-            if is_safe_project_artifact "$item" "$search_path"; then
+        fd "${fd_args[@]}" "$pattern" "$search_path" 2> /dev/null > "$output_file.raw" || true
-                echo "$item"
+
-            fi
+        # Single pass: safe + nested + protected
-        done | filter_nested_artifacts | filter_protected_artifacts > "$output_file"
+        if [[ -f "$output_file.raw" ]]; then
+            while IFS= read -r item; do
+                # Check if we should abort (scanning file removed by Ctrl+C)
+                if [[ ! -f "$stats_dir/purge_scanning" ]]; then
+                    rm -f "$output_file.raw"
+                    return
+                fi
+
+                if [[ -n "$item" ]] && is_safe_project_artifact "$item" "$search_path"; then
+                    echo "$item"
+                    # Update scanning path to show current project directory
+                    local project_dir=$(dirname "$item")
+                    echo "$project_dir" > "$stats_dir/purge_scanning" 2> /dev/null || true
+                fi
+            done < "$output_file.raw" | filter_nested_artifacts | filter_protected_artifacts > "$output_file"
+            rm -f "$output_file.raw"
+        else
+            touch "$output_file"
+        fi
     else
         # Pruned find avoids descending into heavy directories.
-        local prune_args=()
-        local prune_dirs=(".git" "Library" ".Trash" "Applications")
-        for dir in "${prune_dirs[@]}"; do
-            prune_args+=("-name" "$dir" "-prune" "-o")
-        done
-        for target in "${PURGE_TARGETS[@]}"; do
-            prune_args+=("-name" "$target" "-print" "-prune" "-o")
-        done
         local find_expr=()
+        local prune_dirs=(".git" "Library" ".Trash" "Applications")
         for dir in "${prune_dirs[@]}"; do
             find_expr+=("-name" "$dir" "-prune" "-o")
         done
@@ -390,28 +406,49 @@ scan_purge_targets() {
             ((i++))
         done
         command find "$search_path" -mindepth "$min_depth" -maxdepth "$max_depth" -type d \
-            \( "${find_expr[@]}" \) 2> /dev/null | while IFS= read -r item; do
+            \( "${find_expr[@]}" \) 2> /dev/null > "$output_file.raw" || true
-            if is_safe_project_artifact "$item" "$search_path"; then
+
-                echo "$item"
+        # Single pass: safe + nested + protected
-            fi
+        if [[ -f "$output_file.raw" ]]; then
-        done | filter_nested_artifacts | filter_protected_artifacts > "$output_file"
+            while IFS= read -r item; do
+                # Check if we should abort (scanning file removed by Ctrl+C)
+                if [[ ! -f "$stats_dir/purge_scanning" ]]; then
+                    rm -f "$output_file.raw"
+                    return
+                fi
+
+                if [[ -n "$item" ]] && is_safe_project_artifact "$item" "$search_path"; then
+                    echo "$item"
+                    # Update scanning path to show current project directory
+                    local project_dir=$(dirname "$item")
+                    echo "$project_dir" > "$stats_dir/purge_scanning" 2> /dev/null || true
+                fi
+            done < "$output_file.raw" | filter_nested_artifacts | filter_protected_artifacts > "$output_file"
+            rm -f "$output_file.raw"
+        else
+            touch "$output_file"
+        fi
     fi
 }
-# Filter out nested artifacts (e.g. node_modules inside node_modules).
+# Filter out nested artifacts (e.g. node_modules inside node_modules, .build inside build).
+# Optimized: Sort paths to put parents before children, then filter in single pass.
 filter_nested_artifacts() {
-    while IFS= read -r item; do
+    # 1. Append trailing slash to each path (to ensure /foo/bar starts with /foo/)
-        local parent_dir=$(dirname "$item")
+    # 2. Sort to group parents and children (LC_COLLATE=C ensures standard sorting)
-        local is_nested=false
+    # 3. Use awk to filter out paths that start with the previous kept path
-        for target in "${PURGE_TARGETS[@]}"; do
+    # 4. Remove trailing slash
-            if [[ "$parent_dir" == *"/$target/"* || "$parent_dir" == *"/$target" ]]; then
+    sed 's|[^/]$|&/|' | LC_COLLATE=C sort | awk '
-                is_nested=true
+        BEGIN { last_kept = "" }
-                break
+        {
-            fi
+            current = $0
-        done
+            # If current path starts with last_kept, it is nested
-        if [[ "$is_nested" == "false" ]]; then
+            # Only check if last_kept is not empty
-            echo "$item"
+            if (last_kept == "" || index(current, last_kept) != 1) {
-        fi
+                print current
-    done
+                last_kept = current
+            }
+        }
+    ' | sed 's|/$||'
 }
 
 filter_protected_artifacts() {
@@ -703,17 +740,14 @@ clean_project_artifacts() {
         for temp in "${scan_temps[@]+"${scan_temps[@]}"}"; do
             rm -f "$temp" 2> /dev/null || true
         done
-        if [[ -t 1 ]]; then
+        # Clean up purge scanning file
-            stop_inline_spinner
+        local stats_dir="${XDG_CACHE_HOME:-$HOME/.cache}/mole"
-        fi
+        rm -f "$stats_dir/purge_scanning" 2> /dev/null || true
         echo ""
         exit 130
     }
     trap cleanup_scan INT TERM
-    # Start parallel scanning of all paths at once
+    # Scanning is started from purge.sh with start_inline_spinner
-    if [[ -t 1 ]]; then
-        start_inline_spinner "Scanning projects..."
-    fi
     # Launch all scans in parallel
     for path in "${PURGE_SEARCH_PATHS[@]}"; do
         if [[ -d "$path" ]]; then
@@ -730,9 +764,6 @@ clean_project_artifacts() {
     for pid in "${scan_pids[@]+"${scan_pids[@]}"}"; do
         wait "$pid" 2> /dev/null || true
     done
-    if [[ -t 1 ]]; then
-        stop_inline_spinner
-    fi
     # Collect all results
     for scan_output in "${scan_temps[@]+"${scan_temps[@]}"}"; do
         if [[ -f "$scan_output" ]]; then

tests/purge.bats

@@ -101,6 +101,27 @@ setup() {
     [[ "$result" == "2" ]]
 }
 
+@test "filter_nested_artifacts: removes Xcode build subdirectories (Mac projects)" {
+    # Simulate Mac Xcode project with nested .build directories:
+    # ~/www/testapp/build
+    # ~/www/testapp/build/Framework.build
+    # ~/www/testapp/build/Package.build
+    mkdir -p "$HOME/www/testapp/build/Framework.build"
+    mkdir -p "$HOME/www/testapp/build/Package.build"
+
+    result=$(bash -c "
+        source '$PROJECT_ROOT/lib/clean/project.sh'
+        printf '%s\n' \
+            '$HOME/www/testapp/build' \
+            '$HOME/www/testapp/build/Framework.build' \
+            '$HOME/www/testapp/build/Package.build' | \
+        filter_nested_artifacts | wc -l | tr -d ' '
+    ")
+    
+    # Should only keep the top-level 'build' directory, filtering out nested .build dirs
+    [[ "$result" == "1" ]]
+}
+
 # Vendor protection unit tests
 @test "is_rails_project_root: detects valid Rails project" {
     mkdir -p "$HOME/www/test-rails/config"