2 files changed, 130 insertions, 0 deletions
diff --git a/cmd/avg-lines/html.go b/cmd/avg-lines/html.go
new file mode 100644
index 0000000..97d8ec9
--- /dev/null
+++ b/cmd/avg-lines/html.go
@@ -0,0 +1,61 @@
+package main
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+
+	"rescribe.xyz/utils/pkg/line"
+)
+
+func copylineimg(fn string, l line.Detail) error {
+	f, err := os.Create(fn)
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+
+	return l.Img.CopyLineTo(f)
+}
+
+func htmlout(dir string, lines line.Details) error {
+	err := os.MkdirAll(dir, 0700)
+	if err != nil {
+		return err
+	}
+
+	fn := filepath.Join(dir, "index.html")
+	f, err := os.Create(fn)
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+
+	_, err = fmt.Fprintf(f, "<!DOCTYPE html><html><head><meta charset='UTF-8'><title></title>"+
+		"<style>td {border: 1px solid #444}</style></head><body>\n<table>\n")
+	if err != nil {
+		return err
+	}
+	for _, l := range lines {
+		fn = filepath.Base(l.OcrName) + "_" + l.Name + ".png"
+		err = copylineimg(filepath.Join(dir, fn), l)
+		if err != nil {
+			return err
+		}
+		_, err = fmt.Fprintf(f, "<tr>\n"+
+			"<td><h1>%.4f%%</h1></td>\n"+
+			"<td>%s %s</td>\n"+
+			"<td><img src='%s' width='100%%' /><br />%s</td>\n"+
+			"</tr>\n",
+			l.Avgconf, l.OcrName, l.Name, fn, l.Text)
+		if err != nil {
+			return err
+		}
+	}
+	_, err = fmt.Fprintf(f, "</table>\n</body></html>\n")
+	if err != nil {
+		return err
+	}
+
+	return nil
+}
diff --git a/cmd/avg-lines/main.go b/cmd/avg-lines/main.go
new file mode 100644
index 0000000..f7cedab
--- /dev/null
+++ b/cmd/avg-lines/main.go
@@ -0,0 +1,69 @@
+package main
+
+import (
+	"flag"
+	"fmt"
+	"log"
+	"os"
+	"path/filepath"
+	"sort"
+
+	"rescribe.xyz/utils/pkg/hocr"
+	"rescribe.xyz/utils/pkg/line"
+	"rescribe.xyz/utils/pkg/prob"
+)
+
+func main() {
+	flag.Usage = func() {
+		fmt.Fprintf(os.Stderr, "Usage: avg-lines [-html dir] [-nosort] [prob1] [hocr1] [prob2] [...]\n")
+		fmt.Fprintf(os.Stderr, "Prints a report of the average confidence for each line, sorted\n")
+		fmt.Fprintf(os.Stderr, "from worst to best.\n")
+		fmt.Fprintf(os.Stderr, "Both .hocr and .prob files can be processed.\n")
+		fmt.Fprintf(os.Stderr, "For .hocr files, the x_wconf data is used to calculate confidence.\n")
+		fmt.Fprintf(os.Stderr, "The .prob files are generated using ocropy-rpred's --probabilities\n")
+		fmt.Fprintf(os.Stderr, "option.\n\n")
+		flag.PrintDefaults()
+	}
+	var html = flag.String("html", "", "Output in html format to the specified directory")
+	var nosort = flag.Bool("nosort", false, "Don't sort lines by confidence")
+	flag.Parse()
+	if flag.NArg() < 1 {
+		flag.Usage()
+		os.Exit(1)
+	}
+
+	var err error
+	lines := make(line.Details, 0)
+
+	for _, f := range flag.Args() {
+		var newlines line.Details
+		switch ext := filepath.Ext(f); ext {
+		case ".prob":
+			newlines, err = prob.GetLineDetails(f)
+		case ".hocr":
+			newlines, err = hocr.GetLineDetails(f)
+		default:
+			log.Printf("Skipping file '%s' as it isn't a .prob or .hocr\n", f)
+			continue
+		}
+		if err != nil {
+			log.Fatal(err)
+		}
+
+		for _, l := range newlines {
+			lines = append(lines, l)
+		}
+	}
+
+	if *nosort == false {
+		sort.Sort(lines)
+	}
+
+	if *html == "" {
+		for _, l := range lines {
+			fmt.Printf("%s %s: %.2f%%\n", l.OcrName, l.Name, l.Avgconf)
+		}
+	} else {
+		htmlout(*html, lines)
+	}
+}