From 010255d2c19bbd06ca7e5ce92d8f98f3c12e0ad3 Mon Sep 17 00:00:00 2001
From: Nick White <git@njw.name>
Date: Thu, 24 Jan 2019 17:26:36 +0000
Subject: Add -d to -hocr tool, and improve documentation

---
 bucket-lines-hocr/bucket-lines-hocr.go           | 91 ++++++++++++++++++++++++
 line-conf-buckets-tess/line-conf-buckets-tess.go | 86 ----------------------
 2 files changed, 91 insertions(+), 86 deletions(-)
 create mode 100644 bucket-lines-hocr/bucket-lines-hocr.go
 delete mode 100644 line-conf-buckets-tess/line-conf-buckets-tess.go

diff --git a/bucket-lines-hocr/bucket-lines-hocr.go b/bucket-lines-hocr/bucket-lines-hocr.go
new file mode 100644
index 0000000..b35c824
--- /dev/null
+++ b/bucket-lines-hocr/bucket-lines-hocr.go
@@ -0,0 +1,91 @@
+package main
+
+// TODO: merge with -prob, using filename extension to determine what to do for each file
+
+import (
+	"flag"
+	"fmt"
+	"image/png"
+	"io/ioutil"
+	"log"
+	"os"
+	"path/filepath"
+	"strings"
+
+	"git.rescribe.xyz/testingtools/parse"
+	"git.rescribe.xyz/testingtools/parse/hocr"
+)
+
+func detailsFromFile(f string) (parse.LineDetails, error) {
+	var newlines parse.LineDetails
+
+	file, err := ioutil.ReadFile(f)
+	if err != nil {
+		return newlines, err
+	}
+
+	h, err := hocr.Parse(file)
+	if err != nil {
+		return newlines, err
+	}
+
+	pngfn := strings.Replace(f, ".hocr", ".png", 1)
+	pngf, err := os.Open(pngfn)
+	if err != nil {
+		return newlines, err
+	}
+	defer pngf.Close()
+	img, err := png.Decode(pngf)
+	if err != nil {
+		return newlines, err
+	}
+
+	n := strings.Replace(filepath.Base(f), ".hocr", "", 1)
+	return hocr.GetLineDetails(h, img, n)
+}
+
+func main() {
+	b := parse.BucketSpecs{
+		// minimum confidence, name
+		{ 0, "bad" },
+		{ 0.95, "95to98" },
+		{ 0.98, "98plus" },
+	}
+
+	flag.Usage = func() {
+		fmt.Fprintf(os.Stderr, "Usage: bucket-lines-hocr [-d dir] hocr1 [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, "This uses the x_wconf data in .hocr files, which it assumes will be.\n")
+		fmt.Fprintf(os.Stderr, "in the same directory as the line's image and text files. It can\n")
+		fmt.Fprintf(os.Stderr, "handle hocr where each character is tagged separately and hocr where\n")
+		fmt.Fprintf(os.Stderr, "only whole words are tagged.\n")
+		flag.PrintDefaults()
+	}
+	dir := flag.String("d", "buckets", "Directory to store the buckets")
+	flag.Parse()
+	if flag.NArg() < 1 {
+		flag.Usage()
+		os.Exit(1)
+	}
+
+	lines := make(parse.LineDetails, 0)
+
+	for _, f := range flag.Args() {
+		newlines, err := detailsFromFile(f)
+		if err != nil {
+			log.Fatal(err)
+		}
+
+		for _, l := range newlines {
+			lines = append(lines, l)
+		}
+	}
+
+	stats, err := parse.BucketUp(lines, b, *dir)
+	if err != nil {
+		log.Fatal(err)
+	}
+
+	parse.PrintBucketStats(os.Stdout, stats)
+}
diff --git a/line-conf-buckets-tess/line-conf-buckets-tess.go b/line-conf-buckets-tess/line-conf-buckets-tess.go
deleted file mode 100644
index 38dec15..0000000
--- a/line-conf-buckets-tess/line-conf-buckets-tess.go
+++ /dev/null
@@ -1,86 +0,0 @@
-package main
-
-// TODO: rename
-// TODO: set bucket dirname from cmdline
-
-import (
-	"flag"
-	"fmt"
-	"image/png"
-	"io/ioutil"
-	"log"
-	"os"
-	"path/filepath"
-	"strings"
-
-	"git.rescribe.xyz/testingtools/parse"
-	"git.rescribe.xyz/testingtools/parse/hocr"
-)
-
-func detailsFromFile(f string) (parse.LineDetails, error) {
-	var newlines parse.LineDetails
-
-	file, err := ioutil.ReadFile(f)
-	if err != nil {
-		return newlines, err
-	}
-
-	h, err := hocr.Parse(file)
-	if err != nil {
-		return newlines, err
-	}
-
-	pngfn := strings.Replace(f, ".hocr", ".png", 1)
-	pngf, err := os.Open(pngfn)
-	if err != nil {
-		return newlines, err
-	}
-	defer pngf.Close()
-	img, err := png.Decode(pngf)
-	if err != nil {
-		return newlines, err
-	}
-
-	n := strings.Replace(filepath.Base(f), ".hocr", "", 1)
-	return hocr.GetLineDetails(h, img, n)
-}
-
-func main() {
-	flag.Usage = func() {
-		fmt.Fprintf(os.Stderr, "Usage: line-conf-buckets hocr1 [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")
-		flag.PrintDefaults()
-	}
-	flag.Parse()
-	if flag.NArg() < 1 {
-		flag.Usage()
-		os.Exit(1)
-	}
-
-	lines := make(parse.LineDetails, 0)
-
-	for _, f := range flag.Args() {
-		newlines, err := detailsFromFile(f)
-		if err != nil {
-			log.Fatal(err)
-		}
-
-		for _, l := range newlines {
-			lines = append(lines, l)
-		}
-	}
-
-	b := parse.BucketSpecs{
-		{ 0, "bad" },
-		{ 0.95, "95to98" },
-		{ 0.98, "98plus" },
-	}
-
-	stats, err := parse.BucketUp(lines, b, "newbuckets")
-	if err != nil {
-		log.Fatal(err)
-	}
-
-	parse.PrintBucketStats(os.Stdout, stats)
-}
-- 
cgit v1.2.1-24-ge1ad