added bounded tree to track exploration boundary

2 files changed, 242 insertions, 0 deletions

diff --git a/weed/util/bounded_tree/bounded_tree.go b/weed/util/bounded_tree/bounded_tree.go
new file mode 100644
index 000000000..fa172a764
--- /dev/null
+++ b/weed/util/bounded_tree/bounded_tree.go
@@ -0,0 +1,129 @@
+package bounded_tree
+
+import (
+	"fmt"
+
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/chrislusf/seaweedfs/weed/util"
+)
+
+type Node struct {
+	Parent   *Node
+	Name     string
+	Children map[string]*Node
+}
+
+type BoundedTree struct {
+	root *Node
+}
+
+func NewBoundedTree() *BoundedTree {
+	return &BoundedTree{
+		root: &Node{
+			Name: "/",
+		},
+	}
+}
+
+type VisitNodeFunc func(path util.FullPath) (childDirectories []string, err error)
+
+// If the path is not visited, call the visitFn for each level of directory
+// No action if the directory has been visited before or does not exist.
+// A leaf node, which has no children, represents a directory not visited.
+// A non-leaf node or a non-existing node represents a directory already visited, or does not need to visit.
+func (t *BoundedTree) EnsureVisited(p util.FullPath, visitFn VisitNodeFunc) {
+	println()
+	println("EnsureVisited", p)
+
+	if t.root == nil {
+		return
+	}
+	components := p.Split()
+	fmt.Printf("components %v %d\n", components, len(components))
+	if canDelete := t.ensureVisited(t.root, util.FullPath("/"), components, 0, visitFn); canDelete {
+		t.root = nil
+	}
+}
+
+func (t *BoundedTree) ensureVisited(n *Node, currentPath util.FullPath, components []string, i int, visitFn VisitNodeFunc) (canDeleteNode bool) {
+
+	println("ensureVisited", currentPath, i)
+
+	if n == nil {
+		fmt.Printf("%s null\n", currentPath)
+		return
+	}
+
+	if n.isVisited() {
+		fmt.Printf("%s visited %v\n", currentPath, n.Name)
+	} else {
+		fmt.Printf("ensure %v\n", currentPath)
+
+		children, err := visitFn(currentPath)
+		if err != nil {
+			glog.V(0).Infof("failed to visit %s: %v", currentPath, err)
+			return
+		}
+
+		if len(children) == 0 {
+			fmt.Printf("  canDelete %v without children\n", currentPath)
+			return true
+		}
+
+		n.Children = make(map[string]*Node)
+		for _, child := range children {
+			fmt.Printf("  add child %v %v\n", currentPath, child)
+			n.Children[child] = &Node{
+				Name: child,
+			}
+		}
+	}
+
+	if i >= len(components) {
+		return
+	}
+
+	fmt.Printf("  check child %v %v\n", currentPath, components[i])
+
+	toVisitNode, found := n.Children[components[i]]
+	if !found {
+		fmt.Printf("  did not find child %v %v\n", currentPath, components[i])
+		return
+	}
+
+	fmt.Printf("  ensureVisited %v %v\n", currentPath, toVisitNode.Name)
+
+	if canDelete := t.ensureVisited(toVisitNode, currentPath.Child(components[i]), components, i+1, visitFn); canDelete {
+
+		fmt.Printf("  delete %v %v\n", currentPath, components[i])
+		delete(n.Children, components[i])
+
+		if len(n.Children) == 0 {
+			fmt.Printf("  canDelete %v\n", currentPath)
+			return true
+		}
+	}
+
+	return false
+
+}
+
+func (n *Node) isVisited() bool {
+	if n == nil {
+		return true
+	}
+	if len(n.Children) > 0 {
+		return true
+	}
+	return false
+}
+
+func (n *Node) getChild(childName string) *Node {
+	if n == nil {
+		return nil
+	}
+	if len(n.Children) > 0 {
+		return n.Children[childName]
+	}
+	return nil
+}
diff --git a/weed/util/bounded_tree/bounded_tree_test.go b/weed/util/bounded_tree/bounded_tree_test.go
new file mode 100644
index 000000000..b8dce664c
--- /dev/null
+++ b/weed/util/bounded_tree/bounded_tree_test.go
@@ -0,0 +1,113 @@
+package bounded_tree
+
+import (
+	"fmt"
+	"testing"
+
+	"github.com/chrislusf/seaweedfs/weed/util"
+)
+
+
+var (
+
+	visitFn = func(path util.FullPath) (childDirectories []string, err error) {
+		fmt.Printf("  visit %v ...\n", path)
+		switch path {
+		case "/":
+			return []string{"a", "g", "h"}, nil
+		case "/a":
+			return []string{"b", "f"}, nil
+		case "/a/b":
+			return []string{"c", "e"}, nil
+		case "/a/b/c":
+			return []string{"d"}, nil
+		case "/a/b/c/d":
+			return []string{"i", "j"}, nil
+		case "/a/b/c/d/i":
+			return []string{}, nil
+		case "/a/b/c/d/j":
+			return []string{}, nil
+		case "/a/b/e":
+			return []string{}, nil
+		case "/a/f":
+			return []string{}, nil
+		}
+		return nil, nil
+	}
+
+
+	printMap = func(m map[string]*Node) {
+		for k := range m {
+			println("  >", k)
+		}
+	}
+
+
+)
+
+func TestBoundedTree(t *testing.T) {
+
+	// a/b/c/d/i
+	// a/b/c/d/j
+	// a/b/c/d
+	// a/b/e
+	// a/f
+	// g
+	// h
+
+	tree := NewBoundedTree()
+
+	tree.EnsureVisited(util.FullPath("/a/b/c"), visitFn)
+
+	printMap(tree.root.Children)
+
+	a := tree.root.getChild("a")
+
+	b := a.getChild("b")
+	if !b.isVisited() {
+		t.Errorf("expect visited /a/b")
+	}
+	c := b.getChild("c")
+	if !c.isVisited() {
+		t.Errorf("expect visited /a/b/c")
+	}
+
+	d := c.getChild("d")
+	if d.isVisited() {
+		t.Errorf("expect unvisited /a/b/c/d")
+	}
+
+	tree.EnsureVisited(util.FullPath("/a/b/c/d"), visitFn)
+	tree.EnsureVisited(util.FullPath("/a/b/c/d/i"), visitFn)
+	tree.EnsureVisited(util.FullPath("/a/b/c/d/j"), visitFn)
+	tree.EnsureVisited(util.FullPath("/a/b/e"), visitFn)
+	tree.EnsureVisited(util.FullPath("/a/f"), visitFn)
+
+	printMap(tree.root.Children)
+
+}
+
+func TestEmptyBoundedTree(t *testing.T) {
+
+	// g
+	// h
+
+	tree := NewBoundedTree()
+
+	visitFn := func(path util.FullPath) (childDirectories []string, err error) {
+		fmt.Printf("  visit %v ...\n", path)
+		switch path {
+		case "/":
+			return []string{"g", "h"}, nil
+		}
+		t.Fatalf("expected visit %s", path)
+		return nil, nil
+	}
+
+	tree.EnsureVisited(util.FullPath("/a/b"), visitFn)
+
+	tree.EnsureVisited(util.FullPath("/a/b"), visitFn)
+
+	printMap(tree.root.Children)
+
+}