OrgGo/unipdf
1
0
Fork 0
mirror of https://github.com/unidoc/unipdf.git synced 2025-04-30 13:48:51 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fixed landscape rotation for text extraction.

Browse Source 
Also compute metrics for standard 14 fonts when not created from dict.
This commit is contained in:
Peter Williams 2018-11-19 16:50:28 +11:00
parent ea8a26a7dc
commit 2f8b50af75
2 changed files with 20 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
pdf/extractor/text.go
View File

@ -805,8 +805,8 @@ func (tl *TextList) SortPosition() {
xi, xj := ti.X, tj.X xi, xj := ti.X, tj.X
yi, yj := ti.Y, tj.Y yi, yj := ti.Y, tj.Y
if ti.Orient == contentstream.OrientationLandscape { if ti.Orient == contentstream.OrientationLandscape {
xi, yi = yi, xi xi, yi = yi, -xi
xj, yj = yj, xj xj, yj = yj, -xj
} }
if yi != yj { if yi != yj {
@ -833,17 +833,21 @@ func (tl *TextList) toLines() []Line {
if t.Orient == contentstream.OrientationPortrait { if t.Orient == contentstream.OrientationPortrait {
portText = append(portText, t) portText = append(portText, t)
} else { } 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) landText = append(landText, t)
} }
} }
common.Log.Debug("toLines: portrait ^^^^^^^")
portLines := portText.toLinesOrient() portLines := portText.toLinesOrient()
common.Log.Debug("toLines: landscape &&&&&&&")
landLines := landText.toLinesOrient() landLines := landText.toLinesOrient()
common.Log.Debug("portText=%d landText=%d", len(portText), len(landText)) common.Log.Debug("portText=%d landText=%d", len(portText), len(landText))
return append(portLines, landLines...) 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. // NOTE: Caller must sort the text list top-to-bottom, left-to-write before calling this function.
func (tl *TextList) toLinesOrient() []Line { func (tl *TextList) toLinesOrient() []Line {
tl.printTexts("toLines: before") tl.printTexts("toLines: before")

21
pdf/model/font.go
View File

@ -115,12 +115,8 @@ func DefaultFont() *PdfFont {
// NewStandard14Font returns the standard 14 font named `basefont` as a *PdfFont, or an error if it // 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. // `basefont` is not one of the standard 14 font names.
func NewStandard14Font(basefont Standard14Font) (*PdfFont, error) { func NewStandard14Font(basefont Standard14Font) (*PdfFont, error) {
std, ok := loadStandard14Font(basefont) font, _, err := NewStandard14FontWithEncoding(basefont, nil)
if !ok { return font, err
common.Log.Debug("ERROR: Invalid standard 14 font name %#q", basefont)
return nil, ErrFontNotSupported
}
return &PdfFont{context: &std}, nil
} }
// NewStandard14FontMustCompile returns the standard 14 font named `basefont` as a *PdfFont. // 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 // 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. // 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. // 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" baseEncoder := "MacRomanEncoding"
common.Log.Trace("NewStandard14FontWithEncoding: basefont=%#q baseEncoder=%#q alphabet=%q", common.Log.Trace("NewStandard14FontWithEncoding: basefont=%#q baseEncoder=%#q alphabet=%q",
basefont, baseEncoder, string(sortedAlphabet(alphabet))) basefont, baseEncoder, string(sortedAlphabet(alphabet)))
@ -201,9 +198,15 @@ func NewStandard14FontWithEncoding(basefont Standard14Font, alphabet map[rune]in
slotIdx++ 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`. // GetAlphabet returns a map of the runes in `text`.