1 files changed, 45 insertions, 3 deletions

diff --git a/skate/zippy.go b/skate/zippy.go
index c949069..7aff7a6 100644
--- a/skate/zippy.go
+++ b/skate/zippy.go
@@ -319,7 +319,7 @@ func ZippyVerifyRefsOpenLibrary(olr, refs io.Reader, w io.Writer) error {
 // We can identify, which docs have been matched by checking the ref key and index.
 func ZippyBrefAugment(bref, raw io.Reader, w io.Writer) error {
 	var (
-		_       = json.NewEncoder(w)
+		enc     = json.NewEncoder(w)
 		keyer   = makeKeyFunc("\t", 1)
 		grouper = func(g *zipkey.Group) error {
 			// g.G0 contains a matched docs for a given work id, g.G1 all raw
@@ -327,13 +327,26 @@ func ZippyBrefAugment(bref, raw io.Reader, w io.Writer) error {
 
 			// First, iterate over all matches and sort out duplicates, e.g.
 			// docs that have the same source and target id.
-			unique, err := uniqueMatches(g.G0)
+			matched, err := uniqueMatches(g.G0)
 			if err != nil {
 				return err
 			}
+			var refs = make([]*Ref, len(g.G1))
+			for i := 0; i < len(refs); i++ {
+				var ref Ref
+				if err := json.Unmarshal([]byte(g.G1[i]), &ref); err != nil {
+					return err
+				}
+				refs[i] = &ref
+			}
+			matchedRefsExtend(matched, refs)
+			for _, bref := range matched {
+				if err := enc.Encode(bref); err != nil {
+					return err
+				}
+			}
 			// We want to find all items in g.G1, which are not in unique. This
 			// is a set like operation, but we want a custom comparator.
-			log.Println(unique)
 
 			return nil
 		}
@@ -342,6 +355,35 @@ func ZippyBrefAugment(bref, raw io.Reader, w io.Writer) error {
 	return zipper.Run()
 }
 
+// matchedRefsExtend takes a set of (unique) biblioref docs and will emit that
+// set of biblioref docs (unchanged) plus raw references as biblioref, which
+// did not result in a match (determined by ref key and index).
+func matchedRefsExtend(matched []*BiblioRef, refs []*Ref) {
+	s := set.New() // store key + index of matched items
+	for _, m := range matched {
+		s.Add(m.Key + fmt.Sprintf("%d", m.RefIndex))
+	}
+	for _, r := range refs {
+		if s.Contains(r.Key + fmt.Sprintf("%d", r.Index)) {
+			continue
+		}
+		var bref BiblioRef
+		bref.Key = fmt.Sprintf("%s_%d", r.ReleaseIdent, r.Index)
+		bref.RefIndex = r.Index
+		bref.RefKey = r.Key
+		bref.SourceReleaseIdent = r.ReleaseIdent
+		bref.SourceReleaseStage = r.ReleaseStage
+		bref.SourceWorkIdent = r.WorkIdent
+		bref.SourceYear = fmt.Sprintf("%d", r.ReleaseYear)
+		bref.TargetUnstructured = r.Biblio.Unstructured
+		// Reuse fields for debugging, for now.
+		bref.MatchStatus = StatusUnmatched.Short()
+		bref.MatchReason = ReasonUnknown.Short()
+		matched = append(matched, &bref)
+	}
+	return
+}
+
 // uniqueMatches takes a list of bref docs (unserialized) and will return a
 // list of serialized bref docs, containing unique matches.
 func uniqueMatches(docs []string) (result []*BiblioRef, err error) {