Merge pull request ipfs/go-merkledag#50 from MichaelMure/dagutils

Migrate dagutils from go-ipfs

This commit was moved from ipfs/go-merkledag@cfc27e8

ipld/merkledag/dagutils/diff.go

@@ -0,0 +1,214 @@
+package dagutils
+
+import (
+	"context"
+	"fmt"
+	"path"
+
+	"github.com/ipfs/go-cid"
+	ipld "github.com/ipfs/go-ipld-format"
+
+	dag "github.com/ipfs/go-merkledag"
+)
+
+// ChangeType denotes type of change in Change
+type ChangeType int
+
+// These constants define the changes that can be applied to a DAG.
+const (
+	Add ChangeType = iota
+	Remove
+	Mod
+)
+
+// Change represents a change to a DAG and contains a reference to the old and
+// new CIDs.
+type Change struct {
+	Type   ChangeType
+	Path   string
+	Before cid.Cid
+	After  cid.Cid
+}
+
+// String prints a human-friendly line about a change.
+func (c *Change) String() string {
+	switch c.Type {
+	case Add:
+		return fmt.Sprintf("Added %s at %s", c.After.String(), c.Path)
+	case Remove:
+		return fmt.Sprintf("Removed %s from %s", c.Before.String(), c.Path)
+	case Mod:
+		return fmt.Sprintf("Changed %s to %s at %s", c.Before.String(), c.After.String(), c.Path)
+	default:
+		panic("nope")
+	}
+}
+
+// ApplyChange applies the requested changes to the given node in the given dag.
+func ApplyChange(ctx context.Context, ds ipld.DAGService, nd *dag.ProtoNode, cs []*Change) (*dag.ProtoNode, error) {
+	e := NewDagEditor(nd, ds)
+	for _, c := range cs {
+		switch c.Type {
+		case Add:
+			child, err := ds.Get(ctx, c.After)
+			if err != nil {
+				return nil, err
+			}
+
+			childpb, ok := child.(*dag.ProtoNode)
+			if !ok {
+				return nil, dag.ErrNotProtobuf
+			}
+
+			err = e.InsertNodeAtPath(ctx, c.Path, childpb, nil)
+			if err != nil {
+				return nil, err
+			}
+
+		case Remove:
+			err := e.RmLink(ctx, c.Path)
+			if err != nil {
+				return nil, err
+			}
+
+		case Mod:
+			err := e.RmLink(ctx, c.Path)
+			if err != nil {
+				return nil, err
+			}
+			child, err := ds.Get(ctx, c.After)
+			if err != nil {
+				return nil, err
+			}
+
+			childpb, ok := child.(*dag.ProtoNode)
+			if !ok {
+				return nil, dag.ErrNotProtobuf
+			}
+
+			err = e.InsertNodeAtPath(ctx, c.Path, childpb, nil)
+			if err != nil {
+				return nil, err
+			}
+		}
+	}
+
+	return e.Finalize(ctx, ds)
+}
+
+// Diff returns a set of changes that transform node 'a' into node 'b'.
+// It only traverses links in the following cases:
+// 1. two node's links number are greater than 0.
+// 2. both of two nodes are ProtoNode.
+// Otherwise, it compares the cid and emits a Mod change object.
+func Diff(ctx context.Context, ds ipld.DAGService, a, b ipld.Node) ([]*Change, error) {
+	// Base case where both nodes are leaves, just compare
+	// their CIDs.
+	if len(a.Links()) == 0 && len(b.Links()) == 0 {
+		return getChange(a, b)
+	}
+
+	var out []*Change
+	cleanA, okA := a.Copy().(*dag.ProtoNode)
+	cleanB, okB := b.Copy().(*dag.ProtoNode)
+	if !okA || !okB {
+		return getChange(a, b)
+	}
+
+	// strip out unchanged stuff
+	for _, lnk := range a.Links() {
+		l, _, err := b.ResolveLink([]string{lnk.Name})
+		if err == nil {
+			if l.Cid.Equals(lnk.Cid) {
+				// no change... ignore it
+			} else {
+				anode, err := lnk.GetNode(ctx, ds)
+				if err != nil {
+					return nil, err
+				}
+
+				bnode, err := l.GetNode(ctx, ds)
+				if err != nil {
+					return nil, err
+				}
+
+				sub, err := Diff(ctx, ds, anode, bnode)
+				if err != nil {
+					return nil, err
+				}
+
+				for _, subc := range sub {
+					subc.Path = path.Join(lnk.Name, subc.Path)
+					out = append(out, subc)
+				}
+			}
+			_ = cleanA.RemoveNodeLink(l.Name)
+			_ = cleanB.RemoveNodeLink(l.Name)
+		}
+	}
+
+	for _, lnk := range cleanA.Links() {
+		out = append(out, &Change{
+			Type:   Remove,
+			Path:   lnk.Name,
+			Before: lnk.Cid,
+		})
+	}
+	for _, lnk := range cleanB.Links() {
+		out = append(out, &Change{
+			Type:  Add,
+			Path:  lnk.Name,
+			After: lnk.Cid,
+		})
+	}
+
+	return out, nil
+}
+
+// Conflict represents two incompatible changes and is returned by MergeDiffs().
+type Conflict struct {
+	A *Change
+	B *Change
+}
+
+// MergeDiffs takes two slice of changes and adds them to a single slice.
+// When a Change from b happens to the same path of an existing change in a,
+// a conflict is created and b is not added to the merged slice.
+// A slice of Conflicts is returned and contains pointers to the
+// Changes involved (which share the same path).
+func MergeDiffs(a, b []*Change) ([]*Change, []Conflict) {
+	var out []*Change
+	var conflicts []Conflict
+	paths := make(map[string]*Change)
+	for _, c := range a {
+		paths[c.Path] = c
+	}
+
+	for _, c := range b {
+		if ca, ok := paths[c.Path]; ok {
+			conflicts = append(conflicts, Conflict{
+				A: ca,
+				B: c,
+			})
+		} else {
+			out = append(out, c)
+		}
+	}
+	for _, c := range paths {
+		out = append(out, c)
+	}
+	return out, conflicts
+}
+
+func getChange(a, b ipld.Node) ([]*Change, error) {
+	if a.Cid().Equals(b.Cid()) {
+		return []*Change{}, nil
+	}
+	return []*Change{
+		{
+			Type:   Mod,
+			Before: a.Cid(),
+			After:  b.Cid(),
+		},
+	}, nil
+}

ipld/merkledag/dagutils/diffenum.go

@@ -0,0 +1,99 @@
+package dagutils
+
+import (
+	"context"
+	"fmt"
+
+	cid "github.com/ipfs/go-cid"
+	ipld "github.com/ipfs/go-ipld-format"
+
+	mdag "github.com/ipfs/go-merkledag"
+)
+
+// DiffEnumerate fetches every object in the graph pointed to by 'to' that is
+// not in 'from'. This can be used to more efficiently fetch a graph if you can
+// guarantee you already have the entirety of 'from'
+func DiffEnumerate(ctx context.Context, dserv ipld.NodeGetter, from, to cid.Cid) error {
+	fnd, err := dserv.Get(ctx, from)
+	if err != nil {
+		return fmt.Errorf("get %s: %s", from, err)
+	}
+
+	tnd, err := dserv.Get(ctx, to)
+	if err != nil {
+		return fmt.Errorf("get %s: %s", to, err)
+	}
+
+	diff := getLinkDiff(fnd, tnd)
+
+	sset := cid.NewSet()
+	for _, c := range diff {
+		// Since we're already assuming we have everything in the 'from' graph,
+		// add all those cids to our 'already seen' set to avoid potentially
+		// enumerating them later
+		if c.bef.Defined() {
+			sset.Add(c.bef)
+		}
+	}
+	for _, c := range diff {
+		if !c.bef.Defined() {
+			if sset.Has(c.aft) {
+				continue
+			}
+			err := mdag.Walk(ctx, mdag.GetLinksDirect(dserv), c.aft, sset.Visit, mdag.Concurrent())
+			if err != nil {
+				return err
+			}
+		} else {
+			err := DiffEnumerate(ctx, dserv, c.bef, c.aft)
+			if err != nil {
+				return err
+			}
+		}
+	}
+
+	return nil
+}
+
+// if both bef and aft are not nil, then that signifies bef was replaces with aft.
+// if bef is nil and aft is not, that means aft was newly added
+// if aft is nil and bef is not, that means bef was deleted
+type diffpair struct {
+	bef, aft cid.Cid
+}
+
+// getLinkDiff returns a changeset between nodes 'a' and 'b'. Currently does
+// not log deletions as our usecase doesnt call for this.
+func getLinkDiff(a, b ipld.Node) []diffpair {
+	ina := make(map[string]*ipld.Link)
+	inb := make(map[string]*ipld.Link)
+	var aonly []cid.Cid
+	for _, l := range b.Links() {
+		inb[l.Cid.KeyString()] = l
+	}
+	for _, l := range a.Links() {
+		var key = l.Cid.KeyString()
+		ina[key] = l
+		if inb[key] == nil {
+			aonly = append(aonly, l.Cid)
+		}
+	}
+
+	var out []diffpair
+	var aindex int
+
+	for _, l := range b.Links() {
+		if ina[l.Cid.KeyString()] != nil {
+			continue
+		}
+
+		if aindex < len(aonly) {
+			out = append(out, diffpair{bef: aonly[aindex], aft: l.Cid})
+			aindex++
+		} else {
+			out = append(out, diffpair{aft: l.Cid})
+			continue
+		}
+	}
+	return out
+}