From 23910bdad2983a4b79aa2b5a1d36d3d0822cb06d Mon Sep 17 00:00:00 2001
From: Martin Czygan <martin.czygan@gmail.com>
Date: Mon, 14 Jun 2021 22:42:47 +0200
Subject: add cut batch helper

---
 skate/zippy.go      | 25 ++++++++++++++++++-------
 skate/zippy_test.go | 31 +++++++++++++++++++++++++++++++
 2 files changed, 49 insertions(+), 7 deletions(-)

diff --git a/skate/zippy.go b/skate/zippy.go
index 70d42fc..b4f2804 100644
--- a/skate/zippy.go
+++ b/skate/zippy.go
@@ -328,19 +328,22 @@ 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.
-			matched, err := uniqueMatches(g.G0)
+			matched, err := uniqueMatches(CutBatch(g.G0, 2))
 			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 {
+				var (
+					data []byte = []byte(Cut(g.G1[i], 2))
+					ref  Ref
+				)
+				if err := json.Unmarshal(data, &ref); err != nil {
 					return err
 				}
 				refs[i] = &ref
 			}
-			matchedRefsExtend(matched, refs) // XXX: Still duplicates (?)
+			matchedRefsExtend(matched, refs)
 			for _, bref := range matched {
 				if err := enc.Encode(bref); err != nil {
 					return err
@@ -419,13 +422,21 @@ func uniqueMatches(docs []string) (result []*BiblioRef, err error) {
 	return result, nil
 }
 
-// Cut returns a specific column (1-indexed, like CutSep) from a tabular
-// file, returns empty string if column is invalid.
+// CutBatch runs Cut over a list of lines.
+func CutBatch(lines []string, column int) (result []string) {
+	for _, line := range lines {
+		result = append(result, Cut(line, column))
+	}
+	return result
+}
+
+// Cut returns a specific column (1-indexed) from a line, returns empty string
+// if column is invalid.
 func Cut(line string, column int) string {
 	return CutSep(line, "\t", column)
 }
 
-// CutSep allows to specify a separator.
+// CutSep allows to specify a separator, column is 1-indexed.
 func CutSep(line, sep string, column int) string {
 	parts := strings.Split(strings.TrimSpace(line), sep)
 	if len(parts) < column {
diff --git a/skate/zippy_test.go b/skate/zippy_test.go
index ed982b7..d0529c2 100644
--- a/skate/zippy_test.go
+++ b/skate/zippy_test.go
@@ -29,6 +29,37 @@ func TestLineColumn(t *testing.T) {
 	}
 }
 
+func TestCutBatch(t *testing.T) {
+	var cases = []struct {
+		lines  []string
+		column int
+		result []string
+	}{
+		{
+			[]string{},
+			1,
+			nil,
+		},
+		{
+			[]string{},
+			9,
+			nil,
+		},
+		{
+			[]string{"1\t2\n", "3\t4\n"},
+			2,
+			[]string{"2", "4"},
+		},
+	}
+	for _, c := range cases {
+		result := CutBatch(c.lines, c.column)
+		if !reflect.DeepEqual(result, c.result) {
+			t.Fatalf("got %v (%d), want %v (%d)",
+				result, len(result), c.result, len(c.result))
+		}
+	}
+}
+
 func TestUniqueMatches(t *testing.T) {
 	var cases = []struct {
 		about  string
-- 
cgit v1.2.3