From 561d8461cbe19316762489cd7b04f95b9014bcda Mon Sep 17 00:00:00 2001
From: Nick White <git@njw.name>
Date: Thu, 5 Sep 2019 22:24:37 +0100
Subject: Handle no words found error in a better way so any page that is
actually 0 confidence is recognised
---
bookpipeline/cmd/bookpipeline/main.go | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
(limited to 'bookpipeline/cmd')
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]
--
cgit v1.2.1-24-ge1ad