Changed error handling. Allow partial encoding maps. Don't continue processing unsupported fonts

pdf/extractor/text.go

@@ -12,6 +12,8 @@ package extractor
 import (
 	"bytes"
 	"errors"
+	"fmt"
+	"os"
 
 	"github.com/unidoc/unidoc/common"
 	"github.com/unidoc/unidoc/pdf/contentstream"
@@ -148,10 +150,10 @@ func (e *Extractor) ExtractXYText() (*TextList, error) {
 					return err
 				}
 				err = to.setFont(name, size)
-				if err == model.ErrUnsupportedFont {
-					common.Log.Debug("Swallow error. err=%v", err)
-					err = nil
-				}
+				// if err == model.ErrUnsupportedFont {
+				// 	common.Log.Debug("Swallow error. err=%v", err)
+				// 	err = nil
+				// }
 				if err != nil {
 					return err
 				}
@@ -206,10 +208,10 @@ func (e *Extractor) ExtractXYText() (*TextList, error) {
 		})
 
 	err = processor.Process(e.resources)
-	if err == model.ErrUnsupportedFont {
-		common.Log.Debug("Swallow error. err=%v", err)
-		err = nil
-	}
+	// if err == model.ErrUnsupportedFont {
+	// 	common.Log.Debug("Swallow error. err=%v", err)
+	// 	err = nil
+	// }
 	if err != nil {
 		common.Log.Error("ERROR: Processing: err=%v", err)
 		return textList, err
@@ -329,8 +331,8 @@ func (to *TextObject) setFont(name string, size float64) error {
 	if err == nil {
 		to.State.Tf = font
 	} else if err == ErrFontNotSupported {
-		// XXX: HACK !@#$ This is not correct. Fix it.
-		to.State.Tf = nil
+		return err
+		// to.State.Tf = nil
 	} else {
 		return err
 	}

pdf/model/font.go

@@ -8,6 +8,7 @@ package model
 import (
 	"errors"
 	"fmt"
+	"strings"
 
 	"github.com/unidoc/unidoc/common"
 	. "github.com/unidoc/unidoc/pdf/core"
@@ -112,7 +113,7 @@ func newPdfFontFromPdfObject(fontObj PdfObject, allowType0 bool) (*PdfFont, erro
 		font.context = cidfont
 	default:
 		common.Log.Debug("ERROR: Unsupported font type: font=%s", font)
-		return nil, ErrUnsupportedFont
+		return nil, fmt.Errorf("Unsupported font type: font=%s", font)
 	}
 
 	return font, nil
@@ -159,8 +160,8 @@ func (font PdfFont) CharcodeBytesToUnicode(data []byte) (string, error) {
 		for _, code := range charcodes {
 			r, ok := encoder.CharcodeToRune(code)
 			if !ok {
-				common.Log.Debug("ERROR: No rune. code=0x%04x font=%s encoding=%s data = [% 02x]=%#q",
-					code, font, encoder, data, data)
+				common.Log.Debug("ERROR: No rune. code=0x%04x data = [% 02x]=%#q\nfont=%s\nencoding=%s ",
+					code, data, data, font, encoder)
 				r = cmap.MissingCodeRune
 				return string(data), ErrBadText
 			}
@@ -311,7 +312,7 @@ func (skel fontSkeleton) toDict(subtype string) *PdfObjectDictionary {
 func (skel fontSkeleton) String() string {
 	descriptor := ""
 	if skel.fontDescriptor != nil {
-		descriptor = "(has descriptor)"
+		descriptor = skel.fontDescriptor.String()
 	}
 	return fmt.Sprintf("FONT{%#q %#q %s}", skel.subtype, skel.basefont, descriptor)
 }
@@ -458,6 +459,20 @@ type PdfFontDescriptor struct {
 	container *PdfIndirectObject
 }
 
+func (descriptor *PdfFontDescriptor) String() string {
+	parts := []string{}
+	if descriptor.FontName != nil {
+		parts = append(parts, descriptor.FontName.String())
+	}
+	if descriptor.FontFamily != nil {
+		parts = append(parts, descriptor.FontFamily.String())
+	}
+	parts = append(parts, fmt.Sprintf("FontFile=%t", descriptor.FontFile != nil))
+	parts = append(parts, fmt.Sprintf("FontFile2=%t", descriptor.FontFile2 != nil))
+	parts = append(parts, fmt.Sprintf("FontFile3=%t", descriptor.FontFile3 != nil))
+	return fmt.Sprintf("FONT_DESCRIPTON{%s}", strings.Join(parts, ", "))
+}
+
 // newPdfFontDescriptorFromPdfObject loads the font descriptor from a PdfObject.  Can either be a
 // *PdfIndirectObject or a *PdfObjectDictionary.
 func newPdfFontDescriptorFromPdfObject(obj PdfObject) (*PdfFontDescriptor, error) {
@@ -490,7 +505,6 @@ func newPdfFontDescriptorFromPdfObject(obj PdfObject) (*PdfFontDescriptor, error
 	} else {
 		common.Log.Trace("Incompatibility: Type (Required) missing. font=%q %T",
 			fontname, descriptor.FontName)
-		// return nil, errors.New("$$$$$")
 	}
 
 	descriptor.FontFamily = d.Get("FontFamily")

pdf/model/fontfile.go

@@ -21,7 +21,11 @@ type fontFile struct {
 }
 
 func (fontfile *fontFile) String() string {
-	return fmt.Sprintf("FONTFILE{%#q encoder=%s}", fontfile.name, fontfile.encoder)
+	encoding := "[None]"
+	if fontfile.encoder != nil {
+		encoding = fontfile.encoder.String()
+	}
+	return fmt.Sprintf("FONTFILE{%#q encoder=%s}", fontfile.name, encoding)
 }
 
 // newFontFileFromPdfObject loads a FontFile from a PdfObject.  Can either be a
@@ -107,6 +111,9 @@ func (fontfile *fontFile) loadFromSegments(segment1, segment2 []byte) error {
 // parseAsciiPart parses the ASCII part of the FontFile.
 func (fontfile *fontFile) parseAsciiPart(data []byte) error {
 	common.Log.Debug("parseAsciiPart: %d ", len(data))
+	// fmt.Println("~~~~~~~~~~~~~~~~~~~~~~~^^^~~~~~~~~~~~~~~~~~~~~~~~")
+	// fmt.Printf("data=%s\n", string(data))
+	// fmt.Println("~~~~~~~~~~~~~~~~~~~~~~~!!!~~~~~~~~~~~~~~~~~~~~~~~")
 	// The start of a FontFile looks like
 	//     %!PS-AdobeFont-1.0: MyArial 003.002
 	//     %%Title: MyArial

pdf/model/textencoding/simple.go

@@ -47,7 +47,7 @@ func NewCustomSimpleTextEncoder(encoding map[uint16]string, differences map[byte
 		r, ok := GlyphToRune(glyph)
 		if !ok {
 			common.Log.Debug("ERROR: Unknown glyph. %q", glyph)
-			return SimpleEncoder{}, ErrTypeError
+			// return SimpleEncoder{}, ErrTypeError
 		}
 		baseEncoding[code] = r
 	}
@@ -87,7 +87,7 @@ func newSimpleTextEncoder(baseEncoding map[uint16]rune, baseName string,
 }
 
 // simpleEncoderNumEntries is the maximum number of encoding entries shown in SimpleEncoder.String()
-const simpleEncoderNumEntries = 0
+const simpleEncoderNumEntries = 1000
 
 // String returns a string that describes `se`.
 func (se SimpleEncoder) String() string {