diff --git a/structtext.go b/structtext.go index 5722d5b..0cc88e1 100644 --- a/structtext.go +++ b/structtext.go @@ -23,6 +23,7 @@ package docx import ( "encoding/xml" "io" + "reflect" ) // Tab is the literal tab @@ -70,3 +71,129 @@ func (r *Text) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error { return nil } + +// RunMergeRule compares two runs and decides whether they can be merged +type RunMergeRule func(r1, r2 *Run) bool + +// MergeAllRuns ... +func MergeAllRuns(r1, r2 *Run) bool { + return true +} + +// MergeSamePropRuns merges runs with the same properties +func MergeSamePropRuns(r1, r2 *Run) bool { + if r1 == nil || r2 == nil { + return false + } + if r1.RunProperties == r2.RunProperties { + return true + } + if r1.RunProperties == nil && r2.RunProperties != nil { + return false + } + if r1.RunProperties != nil && r2.RunProperties == nil { + return false + } + rr1 := reflect.ValueOf(r1.RunProperties).Elem() + rr2 := reflect.ValueOf(r2.RunProperties).Elem() + for i := 1; i < rr1.NumField(); i++ { + x1 := rr1.Field(i) + x2 := rr2.Field(i) + if x1.IsZero() && x2.IsZero() { + continue + } + if x1.IsZero() && !x2.IsZero() { + return false + } + if !x1.IsZero() && x2.IsZero() { + return false + } + xx1 := x1.Elem() + if xx1.NumField() <= 1 { + continue + } + xx2 := x2.Elem() + for j := 1; j < xx1.NumField(); j++ { + if !xx1.Field(j).Equal(xx2.Field(j)) { + return false + } + } + } + return true +} + +// MergeText will merge contiguous run texts in a paragraph into one run +// +// note: np is not a deep-copy +func (p *Paragraph) MergeText(canmerge RunMergeRule) (np Paragraph) { + var prevrun *Run + np = *p + np.Children = make([]interface{}, 0, 64) + for _, c := range p.Children { + switch o := c.(type) { + case *Run: + r := *o + r.Children = make([]interface{}, 0, 16) + t := &Text{} + for _, c := range o.Children { + switch x := c.(type) { + case *Text: + if x.Text != "" { + t.Text += x.Text + } + default: + if t.Text != "" { + r.Children = append(r.Children, t) + t = &Text{} + } + r.Children = append(r.Children, x) + } + } + if t.Text != "" { + r.Children = append(r.Children, t) + } + if prevrun != nil && canmerge(prevrun, &r) { + var prevtext *Text + noappend := false + if len(prevrun.Children) == 0 { + prevtext = &Text{} + } else { + i := len(prevrun.Children) - 1 + if t, ok := prevrun.Children[i].(*Text); ok { + prevtext = t + noappend = true + } else { + prevtext = &Text{} + } + } + for _, c := range r.Children { + switch x := c.(type) { + case *Text: + if x.Text != "" { + prevtext.Text += x.Text + } + default: + if prevtext.Text != "" { + if noappend { + noappend = false + } else { + prevrun.Children = append(prevrun.Children, t) + } + prevtext = &Text{} + } + prevrun.Children = append(prevrun.Children, x) + } + } + if prevtext.Text != "" && !noappend { + prevrun.Children = append(prevrun.Children, t) + } + } else { + prevrun = &r + np.Children = append(np.Children, &r) + } + default: + np.Children = append(np.Children, o) + } + } + return +} diff --git a/structtext_test.go b/structtext_test.go new file mode 100644 index 0000000..a7cb7a9 --- /dev/null +++ b/structtext_test.go @@ -0,0 +1,227 @@ +package docx + +import ( + "encoding/xml" + "strings" + "testing" +) + +const xml2merge = ` + + + + + + + + + + + + + + + + 某某某大学 + + + + + + + + 20 + + + + + + + + 1 + + + + + + + 6 + + + + + + + + -20 + + + + + + + + 1 + + + + + + + 7 + + + + + + + + + 学年第 + + + + + + + + + + + + + + + + + + + 1 + + + + + + + + + + + + + + + + + 学期期 + + + + + + + + + + + + + + + + + + 末 + + + + + + + + + + + + + + + + + 考试 + + + + + + + + + + + + + + + + + A + + + + + + + + + + + + + + + + + 卷 + +` + +const ( + allmergedtext = `某某某大学2016-2017学年第1学期期末考试A卷` + propmergedtext = `某某某大学201|6|-201|7|学年第|1|学期期|末|考试||A||卷|` +) + +func TestMergeText(t *testing.T) { + p := Paragraph{} + err := xml.Unmarshal(StringToBytes(xml2merge), &p) + if err != nil { + t.Fatal(err) + } + np := p.MergeText(MergeAllRuns) + if len(np.Children) != 1 { + t.Fatal("expected only one run but has", len(np.Children)) + } + if len(np.Children[0].(*Run).Children) != 1 { + t.Fatal("expected only one run.child but has", len(np.Children[0].(*Run).Children)) + } + if np.Children[0].(*Run).Children[0].(*Text).Text != allmergedtext { + t.Fatal("expected merged text [", allmergedtext, "] but has [", np.Children[0].(*Run).Children[0].(*Text).Text, "]") + } + np = p.MergeText(MergeSamePropRuns) + if len(np.Children) != 13 { + t.Fatal("expected 13 runs but has", len(np.Children)) + } + sb := strings.Builder{} + for _, r := range np.Children { + if len(r.(*Run).Children) > 1 { + t.Fatal("expected 0/1 run.child but has", len(r.(*Run).Children)) + } + if len(r.(*Run).Children) == 1 { + sb.WriteString(r.(*Run).Children[0].(*Text).Text) + } + sb.WriteString("|") + } + if sb.String() != propmergedtext { + t.Fatal("expected merged text [", propmergedtext, "] but has [", sb.String(), "]") + } +}