Handle no words found error in a better way so any page that is actually 0 confidence is recognised

2 files changed, 4 insertions, 4 deletions

diff --git a/bookpipeline/cmd/bookpipeline/main.go b/bookpipeline/cmd/bookpipeline/main.go
index 97f4109..ec6a08d 100644
--- a/bookpipeline/cmd/bookpipeline/main.go
+++ b/bookpipeline/cmd/bookpipeline/main.go
@@ -149,6 +149,9 @@ func analyse(toanalyse chan string, up chan string, errc chan error, logger *log
 		}
 		logger.Println("Calculating confidence for", path)
 		avg, err := hocr.GetAvgConf(path)
+		if err != nil && err.Error() == "No words found" {
+			continue
+		}
 		if err != nil {
 			for range toanalyse {
 			} // consume the rest of the receiving channel so it isn't blocked
@@ -156,9 +159,6 @@ func analyse(toanalyse chan string, up chan string, errc chan error, logger *log
 			errc <- errors.New(fmt.Sprintf("Error retreiving confidence for %s: %s", path, err))
 			return
 		}
-		if avg == 0 {
-			continue
-		}
 		base := filepath.Base(path)
 		codestart := strings.Index(base, "_bin")
 		name := base[0:codestart]
diff --git a/lib/hocr/hocr.go b/lib/hocr/hocr.go
index 052aa19..dcd0494 100644
--- a/lib/hocr/hocr.go
+++ b/lib/hocr/hocr.go
@@ -123,7 +123,7 @@ func GetAvgConf(hocrfn string) (float64, error) {
 		}
 	}
 	if num == 0 {
-		return 0, nil
+		return 0, errors.New("No words found")
 	}
 	return total / num, nil
 }