summaryrefslogtreecommitdiff
path: root/cmd/bucket-lines/main.go
diff options
context:
space:
mode:
Diffstat (limited to 'cmd/bucket-lines/main.go')
-rw-r--r--cmd/bucket-lines/main.go89
1 files changed, 89 insertions, 0 deletions
diff --git a/cmd/bucket-lines/main.go b/cmd/bucket-lines/main.go
new file mode 100644
index 0000000..af81b44
--- /dev/null
+++ b/cmd/bucket-lines/main.go
@@ -0,0 +1,89 @@
+package main
+
+import (
+ "encoding/json"
+ "flag"
+ "fmt"
+ "io/ioutil"
+ "log"
+ "os"
+ "path/filepath"
+
+ "rescribe.xyz/utils/pkg/hocr"
+ "rescribe.xyz/utils/pkg/line"
+ "rescribe.xyz/utils/pkg/prob"
+)
+
+func main() {
+ b := BucketSpecs{
+ // minimum confidence, name
+ {0, "bad"},
+ {0.95, "95to98"},
+ {0.98, "98plus"},
+ }
+
+ flag.Usage = func() {
+ fmt.Fprintf(os.Stderr, "Usage: bucket-lines [-d dir] [-s specs.json] [hocr1] [prob1] [hocr2] [...]\n")
+ fmt.Fprintf(os.Stderr, "Copies image-text line pairs into different directories according\n")
+ fmt.Fprintf(os.Stderr, "to the average character probability for the line.\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")
+ fmt.Fprintf(os.Stderr, "The .prob and .hocr files are assumed to be in the same directory\n")
+ fmt.Fprintf(os.Stderr, "as the line's image and text files.\n\n")
+ flag.PrintDefaults()
+ fmt.Fprintf(os.Stderr, "\nAn example specs.json file would be the following:\n")
+ fmt.Fprintf(os.Stderr, "[{\"min\": 0, \"name\": \"terrible\"}, {\"min\": 0.80, \"name\": \"ok\"}, {\"min\": 0.98, \"name\": \"great\"}]\n")
+ }
+ dir := flag.String("d", "buckets", "Directory to store the buckets")
+ specs := flag.String("s", "", "JSON file describing specs to bucket into")
+ flag.Parse()
+ if flag.NArg() < 1 {
+ flag.Usage()
+ os.Exit(1)
+ }
+
+ if *specs != "" {
+ js, err := ioutil.ReadFile(*specs)
+ if err != nil {
+ log.Fatal(err)
+ }
+ err = json.Unmarshal(js, &b)
+ if err != nil {
+ log.Fatal(err)
+ }
+ }
+
+ 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 {
+ if l.Img != nil {
+ lines = append(lines, l)
+ }
+ }
+ }
+
+ stats, err := BucketUp(lines, b, *dir)
+ if err != nil {
+ log.Fatal(err)
+ }
+
+ PrintBucketStats(os.Stdout, stats)
+}