Rewrite pgconf to be more accurate by measuring average word confidence rather than average line confidence

2 files changed, 29 insertions, 26 deletions

diff --git a/lib/hocr/hocr.go b/lib/hocr/hocr.go
index f6316d8..bfc6a98 100644
--- a/lib/hocr/hocr.go
+++ b/lib/hocr/hocr.go
@@ -98,3 +98,28 @@ func GetText(hocrfn string) (string, error) {
 	}
 	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++
+		}
+	}
+	return total / num, nil
+}
diff --git a/pgconf/main.go b/pgconf/main.go
index d254f42..bc09c23 100644
--- a/pgconf/main.go
+++ b/pgconf/main.go
@@ -7,7 +7,6 @@ import (
 	"os"
 
 	"rescribe.xyz/go.git/lib/hocr"
-	"rescribe.xyz/go.git/lib/line"
 )
 
 func main() {
@@ -22,31 +21,10 @@ func main() {
 		os.Exit(1)
 	}
 
-	var err error
-	lines := make(line.Details, 0)
-
-	for _, f := range flag.Args() {
-		var newlines line.Details
-		newlines, err = hocr.GetLineBasics(f)
-		if err != nil {
-			log.Fatal(err)
-		}
-
-		for _, l := range newlines {
-			lines = append(lines, l)
-		}
-	}
-
-	if len(lines) == 0 {
-		fmt.Printf("No lines found\n")
-		os.Exit(0)
-	}
-
-	var total float64
-	for _, l := range lines {
-		total += l.Avgconf
+	avg, err := hocr.GetAvgConf(flag.Arg(0))
+	if err != nil {
+		log.Fatalf("Error retreiving confidence for %s: %v\n", flag.Arg(0), err)
 	}
-	avg := total / float64(len(lines))
 
-	fmt.Printf("%0.0f\n", avg * 100)
+	fmt.Printf("%0.0f\n", avg)
 }