feat(压缩编辑): 优化字体与压缩，规范文档解析

ofdgo_annot.go

@@ -18,6 +18,8 @@ import (
 	"context"
 	"encoding/xml"
 	"fmt"
+	"strconv"
+	"strings"
 )
 
 // Annotations 注释集合
@@ -93,6 +95,106 @@ type Appearance struct {
 	CompositeGraphicUnit []CompositeGraphicUnit `xml:"CompositeGraphicUnit"`
 }
 
+// UnmarshalXML 解析注释入口，按命名空间读取直接页面引用
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解析错误
+func (a *Annotations) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	if start.Name.Local != "Annotations" {
+		return fmt.Errorf("expected Annotations root, got %s", start.Name.Local)
+	}
+	*a = Annotations{XMLName: start.Name}
+	return decodeOFDCollection(d, start, "Page", &a.Page)
+}
+
+// UnmarshalXML 读取注释所属页面及文件路径，忽略扩展中的同名字段
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解析错误
+func (page *AnnPage) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	*page = AnnPage{PageID: attrValue(start, "PageID")}
+	return decodeOFDChildren(d, start, func(d *xml.Decoder, node xml.StartElement) error {
+		if node.Name.Local == "FileLoc" {
+			return d.DecodeElement(&page.FileLoc, &node)
+		}
+		return d.Skip()
+	})
+}
+
+// UnmarshalXML 解析分页注释文件，不将外部同名节点作为注释
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解析错误
+func (page *PageAnnot) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	if start.Name.Local != "PageAnnot" {
+		return fmt.Errorf("expected PageAnnot root, got %s", start.Name.Local)
+	}
+	*page = PageAnnot{XMLName: start.Name}
+	return decodeOFDCollection(d, start, "Annot", &page.Annot)
+}
+
+// UnmarshalXML 解析注释属性、说明、参数和外观，保留未指定可见性的默认语义
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解析错误
+func (a *Annotation) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	*a = Annotation{
+		ID: attrValue(start, "ID"), Type: attrValue(start, "Type"), Subtype: attrValue(start, "Subtype"),
+		Creator: attrValue(start, "Creator"), LastModDate: attrValue(start, "LastModDate"),
+	}
+	for _, attr := range start.Attr {
+		if attr.Name.Space != "" {
+			continue
+		}
+		var target *bool
+		switch attr.Name.Local {
+		case "Visible":
+			a.Visible = new(bool)
+			target = a.Visible
+		case "NoZoom":
+			target = &a.NoZoom
+		case "NoRotate":
+			target = &a.NoRotate
+		default:
+			continue
+		}
+		value, err := strconv.ParseBool(strings.TrimSpace(attr.Value))
+		if err != nil {
+			return err
+		}
+		*target = value
+	}
+	return decodeOFDChildren(d, start, func(d *xml.Decoder, node xml.StartElement) error {
+		switch node.Name.Local {
+		case "Remark":
+			return d.DecodeElement(&a.Remark, &node)
+		case "Parameters":
+			return d.DecodeElement(&a.Parameters, &node)
+		case "Appearance":
+			return d.DecodeElement(&a.Appearance, &node)
+		}
+		return d.Skip()
+	})
+}
+
+// UnmarshalXML 读取标准注释参数，保留参数顺序和重复名称
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解析错误
+func (p *AnnotationParameters) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	*p = AnnotationParameters{}
+	return decodeOFDCollection(d, start, "Parameter", &p.Parameter)
+}
+
+// UnmarshalXML 读取参数名称与文本，不采用带命名空间的名称属性
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解析错误
+func (p *AnnotationParameter) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	type plain AnnotationParameter
+	var value plain
+	start = ofdXMLStart(start)
+	if err := d.DecodeElement(&value, &start); err != nil {
+		return err
+	}
+	*p = AnnotationParameter(value)
+	return nil
+}
+
 // AnnotationInfos 按页面顺序获取注释信息，包含不可见注释
 // 返回: []AnnotationInfo 注释信息, error 错误信息
 func (r *Reader) AnnotationInfos() ([]AnnotationInfo, error) {

ofdgo_compression_images.go

@@ -167,7 +167,7 @@ func (r *Reader) compressionImageSizes(ctx context.Context, images []ImageInfo,
 	return result, ctx.Err()
 }
 
-// DrawObject 汇总图片、替换图及文字和路径底纹的尺寸，不使用裁剪后的局部面积
+// DrawObject 汇总图片、替换图及文字、路径和图像边框底纹的尺寸，不使用裁剪后的局部面积
 // 入参: object 图形对象, state 继承状态
 // 返回: error 取消或几何错误
 func (v *compressionImages) DrawObject(object *GraphicObject, state RenderState) error {
@@ -193,6 +193,12 @@ func (v *compressionImages) DrawObject(object *GraphicObject, state RenderState)
 	if err != nil {
 		return err
 	}
+	if border := img.Border; border != nil && border.BorderColor != nil && border.BorderColor.Pattern != nil {
+		matrix, _ := renderObjectMatrix(img.Boundary, IdentityMatrix, state)
+		if err := v.patternImages((*FillColor)(border.BorderColor), matrix); err != nil {
+			return err
+		}
+	}
 	local := NewMatrix(img.CTM)
 	if img.CTM == "" {
 		local = Matrix{a: box.W, d: box.H}

ofdgo_editor_font_binding.go

@@ -23,6 +23,7 @@ import (
 	"errors"
 	"fmt"
 	"maps"
+	"path"
 	"slices"
 	"strconv"
 	"strings"
@@ -57,6 +58,22 @@ type FontProcessOptions struct {
 	OnProgress func(stage string, completed, total int) error
 }
 
+// editorFontChange 汇总同一文档的字体修改，XML偏移均基于修改前的文件
+type editorFontChange struct {
+	patches map[string][]editorXMLPatch
+	files   map[string][]byte
+}
+
+// editorBindingFontCache 仅保留当前文档最近读取的字体文件及最近解析结果，操作结束后释放
+type editorBindingFontCache struct {
+	source   string
+	data     []byte
+	checksum [32]byte
+	name     string
+	index    int
+	font     *FontResource
+}
+
 // EmbedFont 嵌入当前文档指定字体，保留资源标识及全部使用位置，成功后可撤销
 // file为空时从编辑器字体来源精确匹配；显式文件表示调用方确认替换，仍校验字符和字形
 // 入参: ctx 取消上下文, id 字体标识, file 字体文件, index 集合索引
@@ -149,6 +166,7 @@ func (e *Editor) ProcessFontsWithOptions(ctx context.Context, options FontProces
 		if err != nil {
 			return nil, err
 		}
+		pending := &editorFontChange{patches: make(map[string][]editorXMLPatch), files: make(map[string][]byte)}
 		ids := slices.Sorted(maps.Keys(scan.fonts))
 		ids = slices.DeleteFunc(ids, func(id string) bool {
 			return (mode == FontEmbed) == (scan.fonts[id].location != "")
@@ -166,13 +184,20 @@ func (e *Editor) ProcessFontsWithOptions(ctx context.Context, options FontProces
 				return nil, err
 			}
 			item := FontChange{Document: document, ID: id, Name: declaration.font.FontName}
-			item.Changed, err = e.changeFontSnapshot(ctx, scan, id, mode, FontFile{}, 0)
+			change, err := e.prepareFontBinding(ctx, scan, id, mode, FontFile{}, 0, pending.files)
 			if err != nil {
 				if ctx.Err() != nil {
 					return nil, ctx.Err()
 				}
 				item.Reason = err.Error()
 			}
+			item.Changed = change != nil
+			if change != nil {
+				for name, patches := range change.patches {
+					pending.patches[name] = append(pending.patches[name], patches...)
+				}
+				maps.Copy(pending.files, change.files)
+			}
 			report = append(report, item)
 			if err := progress("fonts", len(report), total); err != nil {
 				return nil, err
@@ -182,13 +207,12 @@ func (e *Editor) ProcessFontsWithOptions(ctx context.Context, options FontProces
 				if mode == FontExternal {
 					externalFiles = append(externalFiles, declaration.location)
 				}
-				scan, err = scanEditorFonts(ctx, reader, document)
-				if err != nil {
+			}
+		}
+		if err := pending.apply(ctx, scan); err != nil {
 			return nil, err
 		}
 	}
-		}
-	}
 	if err := progress("fonts", len(report), len(report)); err != nil {
 		return nil, err
 	}
@@ -208,26 +232,40 @@ func (e *Editor) ProcessFontsWithOptions(ctx context.Context, options FontProces
 // 入参: ctx 取消上下文, scan 文档引用, id 字体标识, mode 保存方式, file 字体文件, index 集合索引
 // 返回: bool 是否修改, error 字体或结构错误
 func (e *Editor) changeFontSnapshot(ctx context.Context, scan *editorFontDocument, id string, mode FontMode, file FontFile, index int) (bool, error) {
+	change, err := e.prepareFontBinding(ctx, scan, id, mode, file, index, nil)
+	if err != nil || change == nil {
+		return false, err
+	}
+	if err := change.apply(ctx, scan); err != nil {
+		return false, err
+	}
+	return true, nil
+}
+
+// prepareFontBinding 校验单个字体并生成独立修改集，不改变文档及已准备的资源
+// 入参: ctx 取消上下文, scan 文档引用, id 字体标识, mode 保存方式, file 字体文件, index 集合索引, pending 已准备的资源文件
+// 返回: *editorFontChange 字体修改，无需处理时为nil, error 字体或结构错误
+func (e *Editor) prepareFontBinding(ctx context.Context, scan *editorFontDocument, id string, mode FontMode, file FontFile, index int, pending map[string][]byte) (*editorFontChange, error) {
 	declaration := scan.fonts[editorResourceID(id)]
 	if declaration == nil {
-		return false, fmt.Errorf("font not found: %s", id)
+		return nil, fmt.Errorf("font not found: %s", id)
 	}
 	if (mode == FontEmbed) == (declaration.location != "") {
-		return false, nil
+		return nil, nil
 	}
 	if scan.unsafe {
-		return false, fmt.Errorf("font references include unsupported content")
+		return nil, fmt.Errorf("font references include unsupported content")
 	}
 	if e.backends.FontResources == nil {
-		return false, ErrBackendUnavailable
+		return nil, ErrBackendUnavailable
 	}
 	var data []byte
 	var err error
 	explicit := mode == FontEmbed && len(file.Data) != 0
 	if mode == FontExternal {
-		data, err = scan.reader.readFile(declaration.location)
+		data, err = scan.fontCache.read(ctx, scan.reader, declaration.location)
 		if err != nil {
-			return false, err
+			return nil, err
 		}
 		file = FontFile{Data: data}
 		if bytes.HasPrefix(data, []byte("ttcf")) {
@@ -236,136 +274,207 @@ func (e *Editor) changeFontSnapshot(ctx context.Context, scan *editorFontDocumen
 	} else if len(file.Data) == 0 {
 		file, err = e.resolveBindingFont(scan, declaration)
 		if err != nil {
-			return false, err
+			return nil, err
 		}
 		index = 0
 	}
-	parsed, err := e.backends.FontResources.OpenFontResource(file, index)
+	parsed, err := scan.fontCache.open(ctx, e.backends.FontResources, file, index)
 	if err != nil {
-		return false, err
+		return nil, err
 	}
 	if mode == FontEmbed {
 		if err := validateFontEmbedding(parsed.Data); err != nil {
-			return false, err
+			return nil, err
 		}
 	}
-	changes := make(map[string][]byte)
+	change := &editorFontChange{patches: make(map[string][]editorXMLPatch), files: make(map[string][]byte)}
 	verifiedGlyphs := false
-	for _, name := range slices.Sorted(maps.Keys(scan.files)) {
+	for _, text := range scan.texts[editorResourceID(id)] {
 		if err := ctx.Err(); err != nil {
-			return false, err
-		}
-		file := scan.files[name]
-		var patches []editorXMLPatch
-		for _, node := range file.texts {
-			if editorResourceID(node.attr("Font")) != editorResourceID(id) {
-				continue
+			return nil, err
 		}
+		file, node := text.file, text.node
 		var object TextObject
 		if err := xml.Unmarshal(file.data[node.start:node.end], &object); err != nil {
-				return false, err
+			return nil, err
 		}
 		if err := checkFontBinding(object, parsed, mode); err != nil {
-				return false, err
+			return nil, err
 		}
 		if explicit && len(object.CGTransform) != 0 && !verifiedGlyphs {
 			source, err := e.resolveBindingFont(scan, declaration)
 			if err != nil || !sameBindingFont(source.Data, parsed.Data) {
-					return false, fmt.Errorf("font glyph mapping requires the original font")
+				return nil, fmt.Errorf("font glyph mapping requires the original font")
 			}
 			verifiedGlyphs = true
 		}
 		if mode == FontExternal {
 			for _, child := range node.children {
 				if child.matchesOFD("CGTransform") {
-						patches = append(patches, editorXMLPatch{start: child.start, end: child.end})
+					change.patches[file.name] = append(change.patches[file.name], editorXMLPatch{start: child.start, end: child.end})
 				}
 			}
 		}
 	}
-		if len(patches) != 0 {
-			changes[name] = editorPatchXML(file.data, patches)
-		}
-	}
 	data = declaration.file.data
 	node := declaration.node
-	if modified, ok := changes[declaration.file.name]; ok {
-		data = modified
-		node, err = fontBindingNode(data, id)
-		if err != nil {
-			return false, err
-		}
-	}
 	fontFile := node.child("FontFile")
+	var fragment []byte
 	if mode == FontEmbed {
 		digest := sha256.Sum256(parsed.Data)
-		name := fmt.Sprintf("Fonts/%x%s", digest, parsed.Extension)
-		if existing, err := scan.reader.readFile(name); err != nil || !bytes.Equal(existing, parsed.Data) {
-			name = packageAvailableName(scan.reader, changes, name)
-			changes[name] = bytes.Clone(parsed.Data)
+		name, reused, err := fontBindingFile(ctx, scan.reader, pending, fmt.Sprintf("Fonts/%x%s", digest, parsed.Extension), parsed.Data)
+		if err != nil {
+			return nil, err
+		}
+		if !reused {
+			change.files[name] = bytes.Clone(parsed.Data)
 		}
 		content := node.textXML("FontFile", "/"+name)
 		if fontFile != nil {
-			data = editorPatchXML(data, []editorXMLPatch{{fontFile.start, fontFile.end, content}})
+			fragment = editorPatchXML(data[node.start:node.end], []editorXMLPatch{{fontFile.start - node.start, fontFile.end - node.start, content}})
 		} else if node.open == node.end {
-			data = editorPatchXML(data, []editorXMLPatch{editorXMLContent(data, node, content)})
+			patch := editorXMLContent(data, node, content)
+			patch.start -= node.start
+			patch.end -= node.start
+			fragment = editorPatchXML(data[node.start:node.end], []editorXMLPatch{patch})
 		} else {
-			data = editorPatchXML(data, []editorXMLPatch{{node.close, node.close, content}})
+			fragment = editorPatchXML(data[node.start:node.end], []editorXMLPatch{{node.close - node.start, node.close - node.start, content}})
 		}
 	} else {
 		if fontFile == nil {
-			return false, fmt.Errorf("font file declaration is missing")
+			return nil, fmt.Errorf("font file declaration is missing")
 		}
-		fragment := editorPatchXML(data[node.start:node.end], []editorXMLPatch{{fontFile.start - node.start, fontFile.end - node.start, nil}})
+		fragment = editorPatchXML(data[node.start:node.end], []editorXMLPatch{{fontFile.start - node.start, fontFile.end - node.start, nil}})
 		for _, field := range []string{"FontName", "FamilyName"} {
 			root, err := parseEditorXML(fragment)
 			if err != nil {
-				return false, err
+				return nil, err
 			}
 			value := root.attr(field)
 			if stripped := externalFontName(value); stripped != value {
 				fragment, err = editorXMLAttribute(fragment, root, field, stripped)
 				if err != nil {
-					return false, err
+					return nil, err
 				}
 			}
 		}
-		data = editorPatchXML(data, []editorXMLPatch{{node.start, node.end, fragment}})
 	}
 	if explicit {
 		for _, field := range []struct{ name, value string }{{"FontName", parsed.Font.FontName}, {"FamilyName", parsed.Font.FamilyName}, {"Charset", "unicode"}} {
-			node, err := fontBindingNode(data, id)
-			if err != nil {
-				return false, err
-			}
-			fragment, err := editorXMLAttribute(data, node, field.name, field.value)
-			if err != nil {
-				return false, err
-			}
-			data = editorPatchXML(data, []editorXMLPatch{{node.start, node.end, fragment}})
-		}
-	}
-	changes[declaration.file.name] = data
-	if err := scan.apply(ctx, changes); err != nil {
-		return false, err
-	}
-	return true, nil
-}
-
-// fontBindingNode 在修改后的资源XML中重新定位字体声明
-// 入参: data 资源XML, id 字体标识
-// 返回: *editorXML 字体节点, error 解析或查找错误
-func fontBindingNode(data []byte, id string) (*editorXML, error) {
-	root, err := parseEditorXML(data)
+			root, err := parseEditorXML(fragment)
 			if err != nil {
 				return nil, err
 			}
-	for node := range editorResourceNodes(root, "Fonts", "Font") {
-		if editorResourceID(node.attr("ID")) == editorResourceID(id) {
-			return node, nil
+			fragment, err = editorXMLAttribute(fragment, root, field.name, field.value)
+			if err != nil {
+				return nil, err
+			}
+		}
+	}
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+	change.patches[declaration.file.name] = append(change.patches[declaration.file.name], editorXMLPatch{node.start, node.end, fragment})
+	return change, nil
+}
+
+// apply 合并同一文档的XML修改，再统一隔离其他文档及版本使用的文件
+// 入参: ctx 取消上下文, scan 修改前的文档引用
+// 返回: error 取消、引用或结构错误
+func (c *editorFontChange) apply(ctx context.Context, scan *editorFontDocument) error {
+	if err := ctx.Err(); err != nil {
+		return err
+	}
+	if len(c.patches) == 0 && len(c.files) == 0 {
+		return nil
+	}
+	changes := maps.Clone(c.files)
+	for name, patches := range c.patches {
+		if err := ctx.Err(); err != nil {
+			return err
+		}
+		changes[name] = editorPatchXML(scan.files[name].data, patches)
+	}
+	return scan.apply(ctx, changes)
+}
+
+// read 复用当前文档最近读取的字体，避免共享引用重复解压，不缓存错误或取消结果
+// 入参: ctx 取消上下文, reader 包快照, name 字体路径
+// 返回: []byte 只读字体数据, error 读取或取消错误
+func (c *editorBindingFontCache) read(ctx context.Context, reader *Reader, name string) ([]byte, error) {
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+	name = cleanPackagePath(name)
+	if c.source == name && c.data != nil {
+		return c.data, nil
+	}
+	c.source, c.data = "", nil
+	data, err := reader.readFileViewContext(ctx, name)
+	if err != nil {
+		return nil, err
+	}
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+	c.source, c.data = name, data
+	return data, nil
+}
+
+// open 按字体内容、名称及集合索引复用最近的解析结果，不缓存错误或取消结果
+// 入参: ctx 取消上下文, backend 字体资源后端, file 字体文件, index 集合索引
+// 返回: *FontResource 只读字体资源, error 解析或取消错误
+func (c *editorBindingFontCache) open(ctx context.Context, backend FontResourceBackend, file FontFile, index int) (*FontResource, error) {
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+	checksum := sha256.Sum256(file.Data)
+	if c.font != nil && c.checksum == checksum && c.name == file.Name && c.index == index {
+		return c.font, ctx.Err()
+	}
+	c.font = nil
+	font, err := backend.OpenFontResource(file, index)
+	if err != nil {
+		return nil, err
+	}
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+	c.checksum, c.name, c.index, c.font = checksum, file.Name, index, font
+	return font, nil
+}
+
+// fontBindingFile 查找未占用的字体路径，同名内容一致时复用已有文件或本次待写资源
+// 入参: ctx 取消上下文, reader 包快照, pending 待写资源, name 期望路径, data 字体内容
+// 返回: string 包内路径, bool 是否复用, error 取消错误
+func fontBindingFile(ctx context.Context, reader *Reader, pending map[string][]byte, name string, data []byte) (string, bool, error) {
+	base, extension := strings.TrimSuffix(name, path.Ext(name)), path.Ext(name)
+	for i := 0; ; i++ {
+		if err := ctx.Err(); err != nil {
+			return "", false, err
+		}
+		candidate := name
+		if i != 0 {
+			candidate = base + "_" + strconv.Itoa(i) + extension
+		}
+		existing, occupied := pending[candidate]
+		if !occupied {
+			existing, occupied = reader.files[candidate]
+			if !occupied {
+				_, occupied = reader.packageFile(candidate)
+				if occupied {
+					var err error
+					existing, err = reader.readFileViewContext(ctx, candidate)
+					if err != nil {
+						continue
+					}
+				}
+			}
+		}
+		if !occupied || bytes.Equal(existing, data) {
+			return candidate, occupied, ctx.Err()
 		}
 	}
-	return nil, fmt.Errorf("font declaration is missing")
 }
 
 // sameBindingFont 比较独立字体的全部数据表，忽略重新封装产生的校验和差异

ofdgo_editor_font_scan.go

@@ -30,6 +30,8 @@ type editorFontDocument struct {
 	index     int
 	files     map[string]*editorFontXML
 	fonts     map[string]*editorFontDeclaration
+	texts     map[string][]editorFontText
+	fontCache editorBindingFontCache
 	unsafe    bool
 }
 
@@ -37,8 +39,12 @@ type editorFontDocument struct {
 type editorFontXML struct {
 	name string
 	data []byte
-	root  *editorXML
-	texts []*editorXML
+}
+
+// editorFontText 关联文字节点及其原始文件，按字体标识归集
+type editorFontText struct {
+	file *editorFontXML
+	node *editorXML
 }
 
 // editorFontDeclaration 关联字体声明、所属资源文件及嵌入数据路径
@@ -53,7 +59,7 @@ type editorFontDeclaration struct {
 // 入参: ctx 取消上下文, reader 包快照, index 文档索引
 // 返回: *editorFontDocument 字体引用, error 读取或结构错误
 func scanEditorFonts(ctx context.Context, reader *Reader, index int) (*editorFontDocument, error) {
-	scan := &editorFontDocument{reader: reader, index: index, files: make(map[string]*editorFontXML), fonts: make(map[string]*editorFontDeclaration)}
+	scan := &editorFontDocument{reader: reader, index: index, files: make(map[string]*editorFontXML), fonts: make(map[string]*editorFontDeclaration), texts: make(map[string][]editorFontText)}
 	entry, err := reader.documentEntryContext(ctx, index)
 	if err != nil {
 		return nil, err
@@ -76,7 +82,7 @@ func scanEditorFonts(ctx context.Context, reader *Reader, index int) (*editorFon
 		if err != nil {
 			return nil, err
 		}
-		file := &editorFontXML{name: name, data: data, root: root}
+		file := &editorFontXML{name: name, data: data}
 		scan.files[name] = file
 		base := ""
 		if root.matchesOFD("Res") {
@@ -122,8 +128,9 @@ func scanEditorFonts(ctx context.Context, reader *Reader, index int) (*editorFon
 				scan.fonts[id] = &editorFontDeclaration{font: font, file: file, node: node, location: location}
 			}
 			if node.matchesOFD("TextObject") || node.matchesOFD("Text") && node.parent != nil && node.parent.matchesOFD("Area") {
-				file.texts = append(file.texts, node)
-				if node.attr("Font") == "" {
+				id := editorResourceID(node.attr("Font"))
+				scan.texts[id] = append(scan.texts[id], editorFontText{file: file, node: node})
+				if id == "" {
 					scan.unsafe = true
 				}
 			}

ofdgo_object.go

@@ -17,6 +17,7 @@ package ofdgo
 import (
 	"encoding/xml"
 	"fmt"
+	"io"
 	"iter"
 	"strconv"
 	"strings"
@@ -31,6 +32,32 @@ type graphicObjectTarget struct {
 	composite *[]CompositeGraphicUnit
 }
 
+// graphicAttributeTokens 提供图形节点的起止令牌，仅用于解码无命名空间属性
+type graphicAttributeTokens struct {
+	start xml.StartElement
+	step  int
+}
+
+// Token 返回起始节点和结束节点，不读取图形子节点
+// 返回: xml.Token XML令牌, error 读取错误
+func (s *graphicAttributeTokens) Token() (xml.Token, error) {
+	s.step++
+	switch s.step {
+	case 1:
+		return s.start, nil
+	case 2:
+		return s.start.End(), nil
+	}
+	return nil, io.EOF
+}
+
+// decodeGraphicAttributes 按结构字段解析图形属性，忽略带命名空间的扩展属性
+// 入参: start 起始节点, value 不含自身解码方法的接收结构
+// 返回: error 解码错误
+func decodeGraphicAttributes(start xml.StartElement, value any) error {
+	return xml.NewTokenDecoder(&graphicAttributeTokens{start: ofdXMLStart(start)}).Decode(value)
+}
+
 // Actions 返回对象动作的独立副本，不修改原对象
 // 返回: []Action 动作列表
 func (o GraphicObject) Actions() []Action {
@@ -57,17 +84,45 @@ func (clips *Clips) usesObjectMatrix() bool {
 // 入参: d XML解码器, start 起始节点
 // 返回: error 解码错误
 func (clips *Clips) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
-	type plain Clips
-	var value plain
+	*clips = Clips{}
 	if classifyOFDNamespace(start.Name.Space) == ofdXMLLegacy {
 		transform := true
-		value.TransFlag = &transform
+		clips.TransFlag = &transform
 	}
-	if err := d.DecodeElement(&value, &start); err != nil {
+	for _, attr := range start.Attr {
+		if attr.Name.Space == "" && attr.Name.Local == "TransFlag" {
+			transform, err := strconv.ParseBool(strings.TrimSpace(attr.Value))
+			if err != nil {
 				return err
 			}
-	*clips = Clips(value)
-	return nil
+			clips.TransFlag = &transform
+		}
+	}
+	return decodeOFDCollection(d, start, "Clip", &clips.Clip)
+}
+
+// UnmarshalXML 解析裁剪分区，不将外部同名节点作为裁剪区域
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解码错误
+func (clip *Clip) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	*clip = Clip{}
+	return decodeOFDCollection(d, start, "Area", &clip.Area)
+}
+
+// UnmarshalXML 解析裁剪区域的变换和路径或文字内容
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解码错误
+func (area *ClipArea) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	*area = ClipArea{DrawParam: attrValue(start, "DrawParam"), CTM: attrValue(start, "CTM")}
+	return decodeOFDChildren(d, start, func(d *xml.Decoder, node xml.StartElement) error {
+		switch node.Name.Local {
+		case "Path":
+			return d.DecodeElement(&area.Path, &node)
+		case "Text":
+			return d.DecodeElement(&area.Text, &node)
+		}
+		return d.Skip()
+	})
 }
 
 // UnmarshalXML 解析路径并区分省略样式、显式实线与零值描边参数
@@ -75,27 +130,77 @@ func (clips *Clips) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
 // 返回: error 错误信息
 func (p *PathObject) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
 	type plain PathObject
-	var value struct {
-		plain
-		DashPattern *string `xml:"DashPattern,attr"`
-	}
-	if err := d.DecodeElement(&value, &start); err != nil {
+	var value plain
+	if err := decodeGraphicAttributes(start, &value); err != nil {
 		return err
 	}
-	*p = PathObject(value.plain)
-	for _, attr := range start.Attr {
-		if attr.Name.Local == "LineWidth" {
+	*p = PathObject(value)
+	for _, attr := range ofdXMLStart(start).Attr {
+		switch attr.Name.Local {
+		case "LineWidth":
 			p.LineWidthSet = p.LineWidth == 0
-		}
-		if attr.Name.Local == "MiterLimit" {
+		case "MiterLimit":
 			p.MiterLimitSet = p.MiterLimit == 0
-		}
-	}
-	if value.DashPattern != nil {
-		p.DashPattern = *value.DashPattern
+		case "DashPattern":
 			p.dashPatternSet = p.DashPattern == ""
 		}
-	return nil
+	}
+	return decodeOFDChildren(d, start, func(d *xml.Decoder, node xml.StartElement) error {
+		switch node.Name.Local {
+		case "FillColor":
+			return d.DecodeElement(&p.FillColor, &node)
+		case "StrokeColor":
+			return d.DecodeElement(&p.StrokeColor, &node)
+		case "AbbreviatedData":
+			return d.DecodeElement(&p.AbbreviatedData, &node)
+		case "Clips":
+			return d.DecodeElement(&p.Clips, &node)
+		case "Actions":
+			return decodeOFDActions(d, node, &p.Actions)
+		}
+		return d.Skip()
+	})
+}
+
+// UnmarshalXML 解析图片及其边框、裁剪和动作，忽略外部同名内容
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解码错误
+func (image *ImageObject) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	type plain ImageObject
+	var value plain
+	if err := decodeGraphicAttributes(start, &value); err != nil {
+		return err
+	}
+	*image = ImageObject(value)
+	return decodeOFDChildren(d, start, func(d *xml.Decoder, node xml.StartElement) error {
+		switch node.Name.Local {
+		case "Border":
+			return d.DecodeElement(&image.Border, &node)
+		case "Clips":
+			return d.DecodeElement(&image.Clips, &node)
+		case "Actions":
+			return decodeOFDActions(d, node, &image.Actions)
+		}
+		return d.Skip()
+	})
+}
+
+// UnmarshalXML 解析图片边框，保留省略线宽与显式零值的区别
+// 入参: d XML解码器, start 起始节点
+// 返回: error 解码错误
+func (border *ImageBorder) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
+	type plain ImageBorder
+	var value plain
+	if err := decodeGraphicAttributes(start, &value); err != nil {
+		return err
+	}
+	*border = ImageBorder(value)
+	return decodeOFDChildren(d, start, func(d *xml.Decoder, node xml.StartElement) error {
+		if node.Name.Local == "BorderColor" {
+			return d.DecodeElement(&border.BorderColor, &node)
+		}
+		return d.Skip()
+	})
 }
 
 // UnmarshalXML 解析绘制参数并保留显式实线和零值参数对基础参数的覆盖

ofdgo_text_xml.go

@@ -14,74 +14,101 @@
 
 package ofdgo
 
-import "encoding/xml"
-
-// textObjectTokens 记录字形变换所属TextCode，供解码后转换为对象内索引
-type textObjectTokens struct {
-	decoder     *xml.Decoder
-	start       *xml.StartElement
-	depth, code int
-	transforms  []int
-}
+import (
+	"encoding/xml"
+	"strconv"
+	"strings"
+)
 
 // UnmarshalXML 解析字形变换，省略的字符数和字形数均取1
 // 入参: decoder XML解码器, start 变换起始节点
 // 返回: error 解码错误
 func (transform *CGTransform) UnmarshalXML(decoder *xml.Decoder, start xml.StartElement) error {
-	type plain CGTransform
-	value := plain{CodeCount: 1, GlyphCount: 1}
+	*transform = CGTransform{CodeCount: 1, GlyphCount: 1}
+	for _, attr := range start.Attr {
+		if attr.Name.Space != "" {
+			continue
+		}
+		var target *int
+		switch attr.Name.Local {
+		case "CodePosition":
+			target = &transform.CodePosition
+		case "CodeCount":
+			target = &transform.CodeCount
+		case "GlyphCount":
+			target = &transform.GlyphCount
+		}
+		if target != nil {
+			value, err := strconv.Atoi(strings.TrimSpace(attr.Value))
+			if err != nil {
+				return err
+			}
+			*target = value
+		}
+	}
+	return decodeOFDChildren(decoder, start, func(d *xml.Decoder, node xml.StartElement) error {
+		if node.Name.Local == "Glyphs" {
+			return d.DecodeElement(&transform.Glyphs, &node)
+		}
+		return d.Skip()
+	})
+}
+
+// UnmarshalXML 解析文字定位及原文，不使用带外部命名空间的定位属性
+// 入参: decoder XML解码器, start 文字节点
+// 返回: error 解码错误
+func (code *TextCode) UnmarshalXML(decoder *xml.Decoder, start xml.StartElement) error {
+	type plain TextCode
+	var value plain
+	start = ofdXMLStart(start)
 	if err := decoder.DecodeElement(&value, &start); err != nil {
 		return err
 	}
-	*transform = CGTransform(value)
+	*code = TextCode(value)
 	return nil
 }
 
-// Token 转发XML令牌并记录直接子节点顺序
-// 返回: xml.Token XML令牌, error 解码错误
-func (s *textObjectTokens) Token() (xml.Token, error) {
-	if s.start != nil {
-		start := *s.start
-		s.start = nil
-		return start, nil
-	}
-	token, err := s.decoder.Token()
-	switch value := token.(type) {
-	case xml.StartElement:
-		if s.depth == 0 && value.Name.Local == "CGTransform" {
-			s.transforms = append(s.transforms, s.code)
-		}
-		s.depth++
-	case xml.EndElement:
-		s.depth--
-		if s.depth == 0 && value.Name.Local == "TextCode" {
-			s.code++
-		}
-	}
-	return token, err
-}
-
-// UnmarshalXML 将各TextCode前的局部字形位置统一为对象内索引
+// UnmarshalXML 解析标准文字节点，将各TextCode前的局部字形位置统一为对象内索引
 // 入参: decoder XML解码器, start 对象起始节点
 // 返回: error 解码错误
 func (obj *TextObject) UnmarshalXML(decoder *xml.Decoder, start xml.StartElement) error {
 	type plain TextObject
-	stream := &textObjectTokens{decoder: decoder, start: &start}
 	var value plain
-	if err := xml.NewTokenDecoder(stream).Decode(&value); err != nil {
+	if err := decodeGraphicAttributes(start, &value); err != nil {
 		return err
 	}
-	offsets := make([]int, len(value.TextCode))
-	for i := 1; i < len(offsets); i++ {
-		offsets[i] = offsets[i-1] + len(textCodeRunes(value.TextCode[i-1].Value))
+	offset, pending := 0, 0
+	if err := decodeOFDChildren(decoder, start, func(d *xml.Decoder, node xml.StartElement) error {
+		switch node.Name.Local {
+		case "FillColor":
+			return d.DecodeElement(&value.FillColor, &node)
+		case "StrokeColor":
+			return d.DecodeElement(&value.StrokeColor, &node)
+		case "Clips":
+			return d.DecodeElement(&value.Clips, &node)
+		case "Actions":
+			return decodeOFDActions(d, node, &value.Actions)
+		case "CGTransform":
+			return d.DecodeElement(&value.CGTransform, &node)
+		case "TextCode":
+			var code TextCode
+			if err := d.DecodeElement(&code, &node); err != nil {
+				return err
 			}
-	for i, code := range stream.transforms {
-		if code < len(offsets) {
-			value.CGTransform[i].CodePosition += offsets[code]
+			for i := pending; i < len(value.CGTransform); i++ {
+				value.CGTransform[i].CodePosition += offset
 			}
+			pending = len(value.CGTransform)
+			offset += len(textCodeRunes(code.Value))
+			value.TextCode = append(value.TextCode, code)
+			return nil
+		}
+		return d.Skip()
+	}); err != nil {
+		return err
 	}
 	*obj = TextObject(value)
-	for _, attr := range start.Attr {
+	for _, attr := range ofdXMLStart(start).Attr {
 		if attr.Name.Local == "LineWidth" {
 			obj.LineWidthSet = obj.LineWidth == 0
 		}