Merge branch 'development' of https://github.com/unidoc/unipdf into c…

…olumns
unidoc · Jun 25, 2020 · b39f205 · b39f205
2 parents 3cca581 + 4f96762
commit b39f205
Show file tree

Hide file tree

Showing 4 changed files with 145 additions and 4 deletions.
diff --git a/creator/creator_test.go b/creator/creator_test.go
@@ -3138,6 +3138,74 @@ func TestReferencedPageDestinations(t *testing.T) {
 	testPages(buf, 6, 5)
 }
 
+func TestExtractTextColor(t *testing.T) {
+	red := ColorRGBFrom8bit(255, 0, 0)
+	green := ColorRGBFrom8bit(0, 255, 0)
+	blue := ColorRGBFrom8bit(0, 0, 255)
+
+	// Test data.
+	type textMark struct {
+		text  string
+		color Color
+	}
+
+	lines := [][]textMark{
+		[]textMark{
+			textMark{text: "a", color: red},
+			textMark{text: "b", color: green},
+			textMark{text: "c", color: blue},
+		},
+		[]textMark{
+			textMark{text: "x", color: green},
+			textMark{text: "y", color: blue},
+			textMark{text: "z", color: red},
+		},
+	}
+
+	// Create output file.
+	c := New()
+
+	for _, line := range lines {
+		p := c.NewStyledParagraph()
+		for _, mark := range line {
+			p.Append(mark.text).Style.Color = mark.color
+		}
+		require.NoError(t, c.Draw(p))
+	}
+
+	buf := bytes.NewBuffer(nil)
+	require.NoError(t, c.Write(buf))
+
+	// Extract output file.
+	reader, err := model.NewPdfReader(bytes.NewReader(buf.Bytes()))
+	require.NoError(t, err)
+
+	for _, page := range reader.PageList {
+		ex, err := extractor.New(page)
+		require.NoError(t, err)
+
+		pageText, _, _, err := ex.ExtractPageText()
+		require.NoError(t, err)
+		marks := pageText.Marks().Elements()
+
+		for i, line := range lines {
+			lenLine := len(line)
+			for j, inMark := range line {
+				outMark := marks[i*lenLine+i+j]
+				outR, outG, outB, _ := outMark.FillColor.RGBA()
+
+				// Compare the fill color of the input mark with the one
+				// of the extracted mark.
+				inR, inG, inB := inMark.color.ToRGB()
+				require.Equal(t, inMark.text, outMark.Text)
+				require.Equal(t, uint32(inR*255), outR>>8)
+				require.Equal(t, uint32(inG*255), outG>>8)
+				require.Equal(t, uint32(inB*255), outB>>8)
+			}
+		}
+	}
+}
+
 var errRenderNotSupported = errors.New("rendering pdf is not supported on this system")
 
 // renderPDFToPNGs uses ghostscript (gs) to render specified PDF file into a set of PNG images (one per page).

diff --git a/extractor/text.go b/extractor/text.go
@@ -9,6 +9,7 @@ import (
 	"bytes"
 	"errors"
 	"fmt"
+	"image/color"
 	"math"
 	"sort"
 	"strings"
@@ -94,7 +95,6 @@ func (e *Extractor) extractPageText(contents string, resources *model.PdfPageRes
 	processor.AddHandler(contentstream.HandlerConditionEnumAllOperands, "",
 		func(op *contentstream.ContentStreamOperation, gs contentstream.GraphicsState,
 			resources *model.PdfPageResources) error {
-
 			operand := op.Operand
 
 			if verboseGeom {
@@ -353,6 +353,14 @@ func (e *Extractor) extractPageText(contents string, resources *model.PdfPageRes
 				pageText.marks = append(pageText.marks, formResult.pageText.marks...)
 				state.numChars += formResult.numChars
 				state.numMisses += formResult.numMisses
+			case "rg", "g", "k", "cs", "sc", "scn":
+				// Set non-stroking color/colorspace.
+				to.gs.ColorspaceNonStroking = gs.ColorspaceNonStroking
+				to.gs.ColorNonStroking = gs.ColorNonStroking
+			case "RG", "G", "K", "CS", "SC", "SCN":
+				// Set stroking color/colorspace.
+				to.gs.ColorspaceStroking = gs.ColorspaceStroking
+				to.gs.ColorStroking = gs.ColorStroking
 			}
 			return nil
 		})
@@ -726,6 +734,16 @@ func (to *textObject) reset() {
 	to.marks = nil
 }
 
+// getFillColor returns the fill color of the text object.
+func (to *textObject) getFillColor() color.Color {
+	return pdfColorToGoColor(to.gs.ColorspaceNonStroking, to.gs.ColorNonStroking)
+}
+
+// getStrokeColor returns the stroke color of the text object.
+func (to *textObject) getStrokeColor() color.Color {
+	return pdfColorToGoColor(to.gs.ColorspaceStroking, to.gs.ColorStroking)
+}
+
 // renderText processes and renders byte array `data` for extraction purposes.
 // It extracts textMarks based the charcodes in `data` and the currect text and graphics states
 // are tracked in `to`.
@@ -767,6 +785,9 @@ func (to *textObject) renderText(data []byte) error {
 
 	common.Log.Trace("renderText: %d codes=%+v runes=%q", len(charcodes), charcodes, len(texts))
 
+	fillColor := to.getFillColor()
+	strokeColor := to.getStrokeColor()
+
 	for i, text := range texts {
 		r := []rune(text)
 		if len(r) == 1 && r[0] == '\x00' {
@@ -826,12 +847,15 @@ func (to *textObject) renderText(data []byte) error {
 			translation(end),
 			math.Abs(spaceWidth*trm.ScalingFactorX()),
 			font,
-			to.state.tc)
+			to.state.tc,
+			fillColor,
+			strokeColor)
+
 		if !onPage {
 			common.Log.Debug("Text mark outside page. Skipping")
 			continue
 		}
-		if font == nil {
+if font == nil {
 			common.Log.Debug("ERROR: No font.")
 		} else if font.Encoder() == nil {
 			common.Log.Debug("ERROR: No encoding. font=%s", font)
@@ -875,6 +899,14 @@ func (to *textObject) moveTo(tx, ty float64) {
 	to.tm = to.tlm
 }
 
+
+
+
+
+
+
+
+
 // PageText represents the layout of text on a device page.
 type PageText struct {
 	marks      []*textMark        // Texts and their positions on a PDF page.
@@ -1084,6 +1116,12 @@ type TextMark struct {
 	// spaces (line breaks) when we see characters that are over a threshold horizontal (vertical)
 	//  distance  apart. See wordJoiner (lineJoiner) in PageText.computeViews().
 	Meta bool
+	// FillColor is the fill color of the text.
+	// The color is nil for spaces and line breaks (i.e. the Meta field is true).
+	FillColor color.Color
+	// StrokeColor is the stroke color of the text.
+	// The color is nil for spaces and line breaks (i.e. the Meta field is true).
+	StrokeColor color.Color
 }
 
 // String returns a string describing `tm`.

diff --git a/extractor/text_mark.go b/extractor/text_mark.go
@@ -7,6 +7,7 @@ package extractor
 
 import (
 	"fmt"
+	"image/color"
 	"math"
 
 	"github.com/unidoc/unipdf/v3/common"
@@ -27,13 +28,16 @@ type textMark struct {
 	trm                transform.Matrix   // The current text rendering matrix (TRM above).
 	end                transform.Point    // The end of character device coordinates.
 	originaBBox        model.PdfRectangle // Bounding box without orientation correction.
+	fillColor          color.Color        // Text fill color.
+	strokeColor        color.Color        // Text stroke color.
 }
 
 // newTextMark returns a textMark for text `text` rendered with text rendering matrix (TRM) `trm`
 // and end of character device coordinates `end`. `spaceWidth` is our best guess at the width of a
 // space in the font the text is rendered in device coordinates.
 func (to *textObject) newTextMark(text string, trm transform.Matrix, end transform.Point,
-	spaceWidth float64, font *model.PdfFont, charspacing float64) (textMark, bool) {
+	spaceWidth float64, font *model.PdfFont, charspacing float64,
+	fillColor, strokeColor color.Color) (textMark, bool) {
 	theta := trm.Angle()
 	orient := nearestMultiple(theta, orientationGranularity)
 	var height float64
@@ -117,6 +121,8 @@ func (to *textObject) newTextMark(text string, trm transform.Matrix, end transfo
 		trm:          trm,
 		end:          end,
 		orient:       orient,
+		fillColor:    fillColor,
+		strokeColor:  strokeColor,
 	}
 	if verboseGeom {
 		common.Log.Info("newTextMark: start=%.2f end=%.2f %s", start, end, tm.String())

diff --git a/extractor/utils.go b/extractor/utils.go
@@ -8,9 +8,12 @@ package extractor
 import (
 	"bytes"
 	"fmt"
+	"image/color"
 
+	"github.com/unidoc/unipdf/v3/common"
 	"github.com/unidoc/unipdf/v3/common/license"
 	"github.com/unidoc/unipdf/v3/core"
+	"github.com/unidoc/unipdf/v3/model"
 )
 
 // RenderMode specifies the text rendering mode (Tmode), which determines whether showing text shall cause
@@ -76,3 +79,29 @@ func truncate(s string, n int) string {
 	}
 	return s[:n]
 }
+
+// pdfColorToGoColor converts the specified color to a Go color, using the
+// provided colorspace. If unsuccessful, color.Black is returned.
+func pdfColorToGoColor(space model.PdfColorspace, c model.PdfColor) color.Color {
+	if space == nil || c == nil {
+		return color.Black
+	}
+
+	conv, err := space.ColorToRGB(c)
+	if err != nil {
+		common.Log.Debug("WARN: could not convert color %v (%v) to RGB: %s", c, space, err)
+		return color.Black
+	}
+	rgb, ok := conv.(*model.PdfColorDeviceRGB)
+	if !ok {
+		common.Log.Debug("WARN: converted color is not in the RGB colorspace: %v", conv)
+		return color.Black
+	}
+
+	return color.NRGBA{
+		R: uint8(rgb.R() * 255),
+		G: uint8(rgb.G() * 255),
+		B: uint8(rgb.B() * 255),
+		A: uint8(255),
+	}
+}