feat(性能优化): 优化文本搜索与流式导出

go.mod

@@ -9,7 +9,7 @@ require (
 	github.com/tdewolff/canvas v0.0.0-20260923214215-09804640d00c
 	github.com/tdewolff/font v0.0.0-20260913163313-54f98bb59ee6
 	github.com/xiaoqidun/jbig2 v1.1.1
-	github.com/xiaoqidun/pdfgo v0.0.0-20260927123406-320d8b5546b5
+	github.com/xiaoqidun/pdfgo v0.0.0-20260927125911-04f5476a9051
 	golang.org/x/crypto v0.57.0
 	golang.org/x/image v0.46.0
 	golang.org/x/text v0.42.0

go.sum

@@ -62,8 +62,8 @@ github.com/tdewolff/test v1.0.12 h1:7F21DqIajswxuche0geHdrUZRCWE4oko4b7bcmkkrxk=
 github.com/tdewolff/test v1.0.12/go.mod h1:XPuWBzvdUzhCuxWO1ojpXsyzsA5bFoS3tO/Q3kFuTG8=
 github.com/xiaoqidun/jbig2 v1.1.1 h1:SjliS9ZO0e9HX5oYUfBuXrfFsXhWricrSz+XY4zORDI=
 github.com/xiaoqidun/jbig2 v1.1.1/go.mod h1:8QonVWtENmh2kc6l+leqeGl7psb79DWlHDkCpdChjfQ=
-github.com/xiaoqidun/pdfgo v0.0.0-20260927123406-320d8b5546b5 h1:FyW2HSyxcrf33xo3F1eKVZmlizgpxML2H+qn63yXcDI=
-github.com/xiaoqidun/pdfgo v0.0.0-20260927123406-320d8b5546b5/go.mod h1:4Pf0JO0BEf1uzRY+uZjlMjN/OPn3PYa0YMi6cqPPTLQ=
+github.com/xiaoqidun/pdfgo v0.0.0-20260927125911-04f5476a9051 h1:VhoxfD1yFJMBcgThFZCo480VVBEU5dB8qAZXWCUSeqI=
+github.com/xiaoqidun/pdfgo v0.0.0-20260927125911-04f5476a9051/go.mod h1:4Pf0JO0BEf1uzRY+uZjlMjN/OPn3PYa0YMi6cqPPTLQ=
 github.com/xyproto/randomstring v1.0.5 h1:YtlWPoRdgMu3NZtP45drfy1GKoojuR7hmRcnhZqKjWU=
 github.com/xyproto/randomstring v1.0.5/go.mod h1:rgmS5DeNXLivK7YprL0pY+lTuhNQW3iGxZ18UQApw/E=
 github.com/yuin/goldmark v1.8.5 h1:r6N5afV5qj/5S4UTch8agZHJ8UxNCMwX7WjkkJam2NA=

ofdgo_render_output.go

@@ -166,7 +166,7 @@ func (r *Renderer) RenderToText(page *PageContent, writer io.Writer) error {
 	if err != nil {
 		return err
 	}
-	_, err = io.WriteString(writer, text.String())
+	_, err = text.WriteTo(writer)
 	return err
 }
 
@@ -191,17 +191,11 @@ func (r *Renderer) RenderToMultiPageText(writer io.Writer, indices ...int) error
 		if err != nil {
 			return fmt.Errorf("failed to extract page %d text: %w", index+1, err)
 		}
-		if value := text.String(); value != "" {
-			if written {
-				if _, err := io.WriteString(writer, "\n"); err != nil {
-					return err
-				}
-			}
-			if _, err := io.WriteString(writer, value); err != nil {
-				return err
-			}
-			written = true
+		n, err := text.writeTo(writer, written)
+		if err != nil {
+			return err
 		}
+		written = written || n > 0
 		if err := r.exportProgress(i+1, len(indices)); err != nil {
 			return err
 		}

ofdgo_search.go

@@ -17,6 +17,7 @@ package ofdgo
 import (
 	"encoding/xml"
 	"fmt"
+	"io"
 	"math"
 	"strings"
 	"unicode"
@@ -62,7 +63,7 @@ type PageText struct {
 }
 
 // TextRun 文本对象原文及逐字符区域，坐标以页面左上角为原点，单位为毫米
-// Boxes与Text的Unicode字符一一对应，缺少字体时使用文本对象区域
+// Boxes与Text的Unicode字符一一对应，缺少字体时使用文本对象区域，缺省区域不影响文字搜索
 type TextRun struct {
 	ID    string     `json:"id"`
 	Text  string     `json:"text"`
@@ -122,16 +123,47 @@ func (r *Renderer) PageTextByIndex(index int) (*PageText, error) {
 // 返回: string 页面原文
 func (p *PageText) String() string {
 	var text strings.Builder
+	_, _ = p.WriteTo(&text)
+	return text.String()
+}
+
+// WriteTo 按文本对象顺序输出UTF-8原文，以换行分隔非空对象，不拼接整页字符串
+// 入参: writer 输出流
+// 返回: int64 已写字节数, error 写入错误，出错时可能已有部分输出
+func (p *PageText) WriteTo(writer io.Writer) (int64, error) {
+	return p.writeTo(writer, false)
+}
+
+// writeTo 输出非空文本对象，可在首个对象前分隔上一页
+// 入参: writer 输出流, separator 是否需要前导换行
+// 返回: int64 已写字节数, error 写入错误
+func (p *PageText) writeTo(writer io.Writer, separator bool) (int64, error) {
+	var written int64
 	for _, run := range p.Runs {
 		if run.Text == "" {
 			continue
 		}
-		if text.Len() > 0 {
-			text.WriteByte('\n')
+		if separator {
+			n, err := io.WriteString(writer, "\n")
+			written += int64(n)
+			if err != nil {
+				return written, err
+			}
+			if n != 1 {
+				return written, io.ErrShortWrite
+			}
 		}
-		text.WriteString(run.Text)
+		n, err := io.WriteString(writer, run.Text)
+		written += int64(n)
+		if err != nil {
+			return written, err
+		}
+		if n != len(run.Text) {
+			return written, io.ErrShortWrite
+		}
+		separator = true
 	}
-	return text.String()
+	return written, nil
 }
 
 // Search 按原文进行忽略大小写的字面匹配，不跨文本对象拼接或识别图像
@@ -166,7 +198,7 @@ func (p *PageText) Search(query string) []TextMatch {
 				Text:   string(runes[start:end]),
 				After:  string(runes[end:min(len(runes), end+16)]),
 			}
-			for _, box := range run.Boxes[start:end] {
+			for _, box := range run.Boxes[min(start, len(run.Boxes)):min(end, len(run.Boxes))] {
 				if box.W > 0 && box.H > 0 && (len(match.Boxes) == 0 || match.Boxes[len(match.Boxes)-1] != box) {
 					match.Boxes = append(match.Boxes, box)
 				}
@@ -184,6 +216,12 @@ func (p *PageText) Search(query string) []TextMatch {
 // 返回: string 匹配用文字
 func foldText(text string) string {
 	return strings.Map(func(char rune) rune {
+		if char < utf8.RuneSelf {
+			if char >= 'a' && char <= 'z' {
+				return char - ('a' - 'A')
+			}
+			return char
+		}
 		folded := char
 		for next := unicode.SimpleFold(char); next != char; next = unicode.SimpleFold(next) {
 			folded = min(folded, next)