summaryrefslogtreecommitdiff
path: root/parse
diff options
context:
space:
mode:
authorNick White <git@njw.name>2019-01-23 23:17:13 +0000
committerNick White <git@njw.name>2019-01-23 23:17:13 +0000
commit367e95adee6a7efbbe85468ac180f06402873a2f (patch)
tree4782e1fee0f515a7e132de42e6ef6776ad6f343a /parse
parentfb77852acbdbbcedcdb9771770cb6771da002851 (diff)
Create general BucketUp function, use it for line-conf-buckets
Diffstat (limited to 'parse')
-rw-r--r--parse/bucket.go82
-rw-r--r--parse/prob/prob.go2
2 files changed, 83 insertions, 1 deletions
diff --git a/parse/bucket.go b/parse/bucket.go
new file mode 100644
index 0000000..3a4cd6a
--- /dev/null
+++ b/parse/bucket.go
@@ -0,0 +1,82 @@
+package parse
+
+import (
+ "io"
+ "path/filepath"
+ "os"
+ "sort"
+ "strconv"
+)
+
+type BucketSpec struct {
+ Min float64
+ Name string
+}
+type BucketSpecs []BucketSpec
+func (b BucketSpecs) Len() int { return len(b) }
+func (b BucketSpecs) Swap(i, j int) { b[i], b[j] = b[j], b[i] }
+func (b BucketSpecs) Less(i, j int) bool { return b[i].Min < b[j].Min }
+
+func bucketLine(l LineDetail, buckets BucketSpecs, dirname string) error {
+ todir := ""
+ for _, b := range buckets {
+ if l.Avgconf >= b.Min {
+ todir = b.Name
+ }
+ }
+
+ if todir == "" {
+ return nil
+ }
+
+ avgstr := strconv.FormatFloat(l.Avgconf, 'G', -1, 64)
+ if len(avgstr) > 2 {
+ avgstr = avgstr[2:]
+ }
+
+ base := filepath.Join(dirname, todir, filepath.Base(l.OcrName) + "_" + l.Name + "_" + avgstr)
+
+ err := os.MkdirAll(filepath.Join(dirname, todir), 0700)
+ if err != nil {
+ return err
+ }
+
+ f, err := os.Create(base + ".png")
+ if err != nil {
+ return err
+ }
+ defer f.Close()
+
+ err = l.Img.CopyLineTo(f)
+ if err != nil {
+ return err
+ }
+
+ f, err = os.Create(base + ".txt")
+ if err != nil {
+ return err
+ }
+ defer f.Close()
+
+ _, err = io.WriteString(f, l.Text)
+ if err != nil {
+ return err
+ }
+
+ return nil
+}
+
+// Copies line images and text into directories based on their
+// confidence, as defined by the buckets struct
+func BucketUp(lines LineDetails, buckets BucketSpecs, dirname string) error {
+ sort.Sort(buckets)
+ // TODO: record and print out summary of % in each bucket category (see how tools did it)
+ for _, l := range lines {
+ err := bucketLine(l, buckets, dirname)
+ if err != nil {
+ return err
+ }
+ }
+
+ return nil
+}
diff --git a/parse/prob/prob.go b/parse/prob/prob.go
index 5a84567..6781a1a 100644
--- a/parse/prob/prob.go
+++ b/parse/prob/prob.go
@@ -57,7 +57,7 @@ func GetLineDetails(name string, r *bufio.Reader) (parse.LineDetails, error) {
return lines, err
}
- line.Name = name
+ line.Name = filepath.Base(filebase)
line.Avgconf = avg
line.Text = string(txt)
line.OcrName = filepath.Dir(filebase)