From 2f8b50af758cf09052b4eb1de556e956a19090cd Mon Sep 17 00:00:00 2001
From: Peter Williams <peter.williams@papercut.com>
Date: Mon, 19 Nov 2018 16:50:28 +1100
Subject: [PATCH] Fixed landscape rotation for text extraction.

Also compute metrics for standard 14 fonts when not created from dict.
---
 pdf/extractor/text.go | 12 ++++++++----
 pdf/model/font.go     | 21 ++++++++++++---------
 2 files changed, 20 insertions(+), 13 deletions(-)

diff --git a/pdf/extractor/text.go b/pdf/extractor/text.go
index e1694861..eb4a93a4 100644
--- a/pdf/extractor/text.go
+++ b/pdf/extractor/text.go
@@ -805,8 +805,8 @@ func (tl *TextList) SortPosition() {
 		xi, xj := ti.X, tj.X
 		yi, yj := ti.Y, tj.Y
 		if ti.Orient == contentstream.OrientationLandscape {
-			xi, yi = yi, xi
-			xj, yj = yj, xj
+			xi, yi = yi, -xi
+			xj, yj = yj, -xj
 		}
 
 		if yi != yj {
@@ -833,17 +833,21 @@ func (tl *TextList) toLines() []Line {
 		if t.Orient == contentstream.OrientationPortrait {
 			portText = append(portText, t)
 		} else {
-			t.X, t.Y = t.Y, t.X
+			t.X, t.Y = t.Y, -t.X
+			t.End.X, t.End.Y = t.End.Y, -t.End.X
+			t.Orient = contentstream.OrientationPortrait
 			landText = append(landText, t)
 		}
 	}
+	common.Log.Debug("toLines: portrait  ^^^^^^^")
 	portLines := portText.toLinesOrient()
+	common.Log.Debug("toLines: landscape &&&&&&&")
 	landLines := landText.toLinesOrient()
 	common.Log.Debug("portText=%d landText=%d", len(portText), len(landText))
 	return append(portLines, landLines...)
 }
 
-// toLinesOrient return the text and positions in `tl` as a slice of Line.
+// toLinesOrient returns the text and positions in `tl` as a slice of Line.
 // NOTE: Caller must sort the text list top-to-bottom, left-to-write before calling this function.
 func (tl *TextList) toLinesOrient() []Line {
 	tl.printTexts("toLines: before")
diff --git a/pdf/model/font.go b/pdf/model/font.go
index c57425fa..0af4fe99 100644
--- a/pdf/model/font.go
+++ b/pdf/model/font.go
@@ -115,12 +115,8 @@ func DefaultFont() *PdfFont {
 // NewStandard14Font returns the standard 14 font named `basefont` as a *PdfFont, or an error if it
 // `basefont` is not one of the standard 14 font names.
 func NewStandard14Font(basefont Standard14Font) (*PdfFont, error) {
-	std, ok := loadStandard14Font(basefont)
-	if !ok {
-		common.Log.Debug("ERROR: Invalid standard 14 font name %#q", basefont)
-		return nil, ErrFontNotSupported
-	}
-	return &PdfFont{context: &std}, nil
+	font, _, err := NewStandard14FontWithEncoding(basefont, nil)
+	return font, err
 }
 
 // NewStandard14FontMustCompile returns the standard 14 font named `basefont` as a *PdfFont.
@@ -137,7 +133,8 @@ func NewStandard14FontMustCompile(basefont Standard14Font) *PdfFont {
 // NewStandard14FontWithEncoding returns the standard 14 font named `basefont` as a *PdfFont and
 // a SimpleEncoder that encodes all the runes in `alphabet`, or an error if this is not possible.
 // An error can occur if`basefont` is not one the standard 14 font names.
-func NewStandard14FontWithEncoding(basefont Standard14Font, alphabet map[rune]int) (*PdfFont, *textencoding.SimpleEncoder, error) {
+func NewStandard14FontWithEncoding(basefont Standard14Font, alphabet map[rune]int) (*PdfFont,
+	*textencoding.SimpleEncoder, error) {
 	baseEncoder := "MacRomanEncoding"
 	common.Log.Trace("NewStandard14FontWithEncoding: basefont=%#q baseEncoder=%#q alphabet=%q",
 		basefont, baseEncoder, string(sortedAlphabet(alphabet)))
@@ -201,9 +198,15 @@ func NewStandard14FontWithEncoding(basefont Standard14Font, alphabet map[rune]in
 			slotIdx++
 		}
 	}
-	encoder, err = textencoding.NewSimpleTextEncoder(baseEncoder, differences)
 
-	return &PdfFont{context: &std}, encoder, err
+	encoder, err = textencoding.NewSimpleTextEncoder(baseEncoder, differences)
+	if err != nil {
+		return nil, nil, err
+	}
+	std.std14Encoder = encoder
+	std.updateStandard14Font()
+
+	return &PdfFont{context: &std}, encoder, nil
 }
 
 // GetAlphabet returns a map of the runes in `text`.