Separate out bookpipeline from catch-all go.git repo, and rename to rescribe.xyz/bookpipeline

The dependencies from the go.git repo will follow in due course.

4 files changed, 0 insertions, 386 deletions

diff --git a/lib/hocr/hocr.go b/lib/hocr/hocr.go
deleted file mode 100644
index dcd0494..0000000
--- a/lib/hocr/hocr.go
+++ /dev/null
@@ -1,129 +0,0 @@
-package hocr
-
-import (
-	"encoding/xml"
-	"errors"
-	"io/ioutil"
-	"regexp"
-	"strconv"
-	"strings"
-)
-
-type Hocr struct {
-	Lines []OcrLine `xml:"body>div>div>p>span"`
-}
-
-type OcrLine struct {
-	Class string    `xml:"class,attr"`
-	Id    string    `xml:"id,attr"`
-	Title string    `xml:"title,attr"`
-	Words []OcrWord `xml:"span"`
-	Text  string    `xml:",chardata"`
-}
-
-type OcrWord struct {
-	Class string    `xml:"class,attr"`
-	Id    string    `xml:"id,attr"`
-	Title string    `xml:"title,attr"`
-	Chars []OcrChar `xml:"span"`
-	Text  string    `xml:",chardata"`
-}
-
-type OcrChar struct {
-	Class string    `xml:"class,attr"`
-	Id    string    `xml:"id,attr"`
-	Title string    `xml:"title,attr"`
-	Chars []OcrChar `xml:"span"`
-	Text  string    `xml:",chardata"`
-}
-
-// Returns the confidence for a word based on its x_wconf value
-func wordConf(s string) (float64, error) {
-	re, err := regexp.Compile(`x_wconf ([0-9.]+)`)
-	if err != nil {
-		return 0.0, err
-	}
-	conf := re.FindStringSubmatch(s)
-	return strconv.ParseFloat(conf[1], 64)
-}
-
-func boxCoords(s string) ([4]int, error) {
-	var coords [4]int
-	re, err := regexp.Compile(`bbox ([0-9]+) ([0-9]+) ([0-9]+) ([0-9]+)`)
-	if err != nil {
-		return coords, err
-	}
-	coordstr := re.FindStringSubmatch(s)
-	for i := range coords {
-		c, err := strconv.Atoi(coordstr[i+1])
-		if err != nil {
-			return coords, err
-		}
-		coords[i] = c
-	}
-	return coords, nil
-}
-
-func noText(s string) bool {
-	t := strings.Trim(s, " \n")
-	return len(t) == 0
-}
-
-func Parse(b []byte) (Hocr, error) {
-	var hocr Hocr
-
-	err := xml.Unmarshal(b, &hocr)
-	if err != nil {
-		return hocr, err
-	}
-
-	return hocr, nil
-}
-
-func GetText(hocrfn string) (string, error) {
-	var s string
-
-	file, err := ioutil.ReadFile(hocrfn)
-	if err != nil {
-		return s, err
-	}
-
-	h, err := Parse(file)
-	if err != nil {
-		return s, err
-	}
-
-
-	for _, l := range h.Lines {
-		s += getLineText(l)
-	}
-	return s, nil
-}
-
-func GetAvgConf(hocrfn string) (float64, error) {
-	file, err := ioutil.ReadFile(hocrfn)
-	if err != nil {
-		return 0, err
-	}
-
-	h, err := Parse(file)
-	if err != nil {
-		return 0, err
-	}
-
-	var total, num float64
-	for _, l := range h.Lines {
-		for _, w := range l.Words {
-			c, err := wordConf(w.Title)
-			if err != nil {
-				return 0, err
-			}
-			total += c
-			num++
-		}
-	}
-	if num == 0 {
-		return 0, errors.New("No words found")
-	}
-	return total / num, nil
-}
diff --git a/lib/hocr/lines.go b/lib/hocr/lines.go
deleted file mode 100644
index 74e8f9a..0000000
--- a/lib/hocr/lines.go
+++ /dev/null
@@ -1,131 +0,0 @@
-package hocr
-
-// TODO: Parse line name to zero pad line numbers, so they can
-//       be sorted easily
-
-import (
-	"image"
-	"image/png"
-	"io/ioutil"
-	"log"
-	"os"
-	"path/filepath"
-	"strings"
-
-	"rescribe.xyz/go.git/lib/line"
-)
-
-func getLineText(l OcrLine) (string) {
-	linetext := ""
-
-	linetext = l.Text
-	if noText(linetext) {
-		linetext = ""
-		for _, w := range l.Words {
-			if w.Class != "ocrx_word" {
-				continue
-			}
-			linetext += w.Text + " "
-		}
-	}
-	if noText(linetext) {
-		linetext = ""
-		for _, w := range l.Words {
-			if w.Class != "ocrx_word" {
-				continue
-			}
-			for _, c := range w.Chars {
-				if c.Class != "ocrx_cinfo" {
-					continue
-				}
-				linetext += c.Text
-			}
-			linetext += " "
-		}
-	}
-	linetext = strings.TrimRight(linetext, " ")
-	linetext += "\n"
-	return linetext
-}
-
-func parseLineDetails(h Hocr, i image.Image, name string) (line.Details, error) {
-	lines := make(line.Details, 0)
-
-	for _, l := range h.Lines {
-		totalconf := float64(0)
-		num := 0
-		for _, w := range l.Words {
-			c, err := wordConf(w.Title)
-			if err != nil {
-				return lines, err
-			}
-			num++
-			totalconf += c
-		}
-
-		coords, err := boxCoords(l.Title)
-		if err != nil {
-			return lines, err
-		}
-
-		var ln line.Detail
-		ln.Name = l.Id
-		ln.Avgconf = (totalconf / float64(num)) / 100
-		ln.Text = getLineText(l)
-		ln.OcrName = name
-		if i != nil {
-			var imgd line.ImgDirect
-			imgd.Img = i.(*image.Gray).SubImage(image.Rect(coords[0], coords[1], coords[2], coords[3]))
-			ln.Img = imgd
-		}
-		lines = append(lines, ln)
-	}
-	return lines, nil
-}
-
-func GetLineDetails(hocrfn string) (line.Details, error) {
-	var newlines line.Details
-
-	file, err := ioutil.ReadFile(hocrfn)
-	if err != nil {
-		return newlines, err
-	}
-
-	h, err := Parse(file)
-	if err != nil {
-		return newlines, err
-	}
-
-	var img image.Image
-	pngfn := strings.Replace(hocrfn, ".hocr", ".png", 1)
-	pngf, err := os.Open(pngfn)
-	if err != nil {
-		log.Println("Warning: can't open image %s\n", pngfn)
-	} else {
-		defer pngf.Close()
-		img, err = png.Decode(pngf)
-		if err != nil {
-			log.Println("Warning: can't load image %s\n", pngfn)
-		}
-	}
-
-	n := strings.Replace(filepath.Base(hocrfn), ".hocr", "", 1)
-	return parseLineDetails(h, img, n)
-}
-
-func GetLineBasics(hocrfn string) (line.Details, error) {
-	var newlines line.Details
-
-	file, err := ioutil.ReadFile(hocrfn)
-	if err != nil {
-		return newlines, err
-	}
-
-	h, err := Parse(file)
-	if err != nil {
-		return newlines, err
-	}
-
-	n := strings.Replace(filepath.Base(hocrfn), ".hocr", "", 1)
-	return parseLineDetails(h, nil, n)
-}
diff --git a/lib/line/line.go b/lib/line/line.go
deleted file mode 100644
index d4e3e44..0000000
--- a/lib/line/line.go
+++ /dev/null
@@ -1,57 +0,0 @@
-package line
-
-import (
-	"image"
-	"image/png"
-	"io"
-	"os"
-)
-
-type Detail struct {
-	Name    string
-	Avgconf float64
-	Img     CopyableImg
-	Text    string
-	OcrName string
-}
-
-type CopyableImg interface {
-	CopyLineTo(io.Writer) error
-}
-
-type Details []Detail
-
-func (l Details) Len() int           { return len(l) }
-func (l Details) Less(i, j int) bool { return l[i].Avgconf < l[j].Avgconf }
-func (l Details) Swap(i, j int)      { l[i], l[j] = l[j], l[i] }
-
-// This is an implementation of the CopyableImg interface that
-// stores the image directly as an image.Image
-type ImgDirect struct {
-	Img image.Image
-}
-
-func (i ImgDirect) CopyLineTo(w io.Writer) error {
-	err := png.Encode(w, i.Img)
-	if err != nil {
-		return err
-	}
-	return nil
-}
-
-// This is an implementation of the CopyableImg interface that
-// stores the path of an image
-type ImgPath struct {
-	Path string
-}
-
-func (i ImgPath) CopyLineTo(w io.Writer) error {
-	f, err := os.Open(i.Path)
-	if err != nil {
-		return err
-	}
-	defer f.Close()
-
-	_, err = io.Copy(w, f)
-	return err
-}
diff --git a/lib/prob/prob.go b/lib/prob/prob.go
deleted file mode 100644
index 31a484d..0000000
--- a/lib/prob/prob.go
+++ /dev/null
@@ -1,69 +0,0 @@
-package prob
-
-import (
-	"io/ioutil"
-	"path/filepath"
-	"strconv"
-	"strings"
-
-	"rescribe.xyz/go.git/lib/line"
-)
-
-func getLineAvg(f string) (float64, error) {
-	totalconf := float64(0)
-	num := 0
-
-	prob, err := ioutil.ReadFile(f)
-	if err != nil {
-		return 0, err
-	}
-
-	for _, l := range strings.Split(string(prob), "\n") {
-		fields := strings.Fields(l)
-
-		if len(fields) == 2 {
-			conf, err := strconv.ParseFloat(fields[1], 64)
-			if err != nil {
-				continue
-			}
-			totalconf += conf
-			num += 1
-		}
-	}
-	if num <= 0 {
-		return 0, nil
-	}
-	avg := totalconf / float64(num)
-	return avg, nil
-}
-
-// Note this only processes one line at a time
-func GetLineDetails(probfn string) (line.Details, error) {
-	var l line.Detail
-	lines := make(line.Details, 0)
-
-	avg, err := getLineAvg(probfn)
-	if err != nil {
-		return lines, err
-	}
-
-	filebase := strings.Replace(probfn, ".prob", "", 1)
-
-	txt, err := ioutil.ReadFile(filebase + ".txt")
-	if err != nil {
-		return lines, err
-	}
-
-	l.Name = filepath.Base(filebase)
-	l.Avgconf = avg
-	l.Text = string(txt)
-	l.OcrName = filepath.Base(filepath.Dir(filebase))
-
-	var imgfn line.ImgPath
-	imgfn.Path = filebase + ".bin.png"
-	l.Img = imgfn
-
-	lines = append(lines, l)
-
-	return lines, nil
-}