From 40c1e4956c0f4cd2c5f139aba7cb7bc04c57fc0f Mon Sep 17 00:00:00 2001
From: Nick White <git@njw.name>
Date: Fri, 25 Jan 2019 17:26:43 +0000
Subject: Add html output including all images, by writing them to an html
directory
---
avg-lines/avg-lines.go | 81 --------------------------------------------------
avg-lines/html.go | 61 +++++++++++++++++++++++++++++++++++++
avg-lines/main.go | 69 ++++++++++++++++++++++++++++++++++++++++++
3 files changed, 130 insertions(+), 81 deletions(-)
delete mode 100644 avg-lines/avg-lines.go
create mode 100644 avg-lines/html.go
create mode 100644 avg-lines/main.go
(limited to 'avg-lines')
diff --git a/avg-lines/avg-lines.go b/avg-lines/avg-lines.go
deleted file mode 100644
index 160dd7c..0000000
--- a/avg-lines/avg-lines.go
+++ /dev/null
@@ -1,81 +0,0 @@
-package main
-
-import (
- "flag"
- "fmt"
- "log"
- "os"
- "path/filepath"
- "sort"
-
- "rescribe.xyz/go.git/lib/line"
- "rescribe.xyz/go.git/lib/hocr"
- "rescribe.xyz/go.git/lib/prob"
-)
-
-func main() {
- flag.Usage = func() {
- fmt.Fprintf(os.Stderr, "Usage: avg-lines [-html] [-nosort] [prob1] [hocr1] [prob2] [...]\n")
- fmt.Fprintf(os.Stderr, "Prints a report of the average confidence for each line, sorted\n")
- fmt.Fprintf(os.Stderr, "from worst to best.\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\n")
- flag.PrintDefaults()
- }
- var usehtml = flag.Bool("html", false, "Output html page")
- var nosort = flag.Bool("nosort", false, "Don't sort lines by confidence")
- flag.Parse()
- if flag.NArg() < 1 {
- flag.Usage()
- os.Exit(1)
- }
-
- 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 {
- lines = append(lines, l)
- }
- }
-
- if *nosort == false {
- sort.Sort(lines)
- }
-
- if *usehtml == false {
- for _, l := range lines {
- fmt.Printf("%s %s: %.2f%%\n", l.OcrName, l.Name, l.Avgconf)
- }
- } else {
- fmt.Printf("<!DOCTYPE html><html><head><meta charset='UTF-8'><title></title><style>td {border: 1px solid #444}</style></head><body>\n")
- fmt.Printf("<table>\n")
- for _, l := range lines {
- fmt.Printf("<tr>\n")
- fmt.Printf("<td><h1>%.4f%%</h1></td>\n", l.Avgconf)
- fmt.Printf("<td>%s %s</td>\n", l.OcrName, l.Name)
- // TODO: think about this, what do we want to do here? if showing imgs is important,
- // will need to copy them somewhere, so works with hocr too
- //fmt.Printf("<td><img src='%s' /><br />%s</td>\n", l.Filebase + ".bin.png", l.Fulltext)
- fmt.Printf("</tr>\n")
- }
- fmt.Printf("</table>\n")
- fmt.Printf("</body></html>\n")
- }
-}
diff --git a/avg-lines/html.go b/avg-lines/html.go
new file mode 100644
index 0000000..f299830
--- /dev/null
+++ b/avg-lines/html.go
@@ -0,0 +1,61 @@
+package main
+
+import (
+ "fmt"
+ "os"
+ "path/filepath"
+
+ "rescribe.xyz/go.git/lib/line"
+)
+
+func copylineimg(fn string, l line.Detail) error {
+ f, err := os.Create(fn)
+ if err != nil {
+ return err
+ }
+ defer f.Close()
+
+ return l.Img.CopyLineTo(f)
+}
+
+func htmlout(dir string, lines line.Details) error {
+ err := os.MkdirAll(dir, 0700)
+ if err != nil {
+ return err
+ }
+
+ fn := filepath.Join(dir, "index.html")
+ f, err := os.Create(fn)
+ if err != nil {
+ return err
+ }
+ defer f.Close()
+
+ _, err = fmt.Fprintf(f, "<!DOCTYPE html><html><head><meta charset='UTF-8'><title></title>" +
+ "<style>td {border: 1px solid #444}</style></head><body>\n<table>\n")
+ if err != nil {
+ return err
+ }
+ for _, l := range lines {
+ fn = filepath.Base(l.OcrName) + "_" + l.Name + ".png"
+ err = copylineimg(filepath.Join(dir, fn), l)
+ if err != nil {
+ return err
+ }
+ _, err = fmt.Fprintf(f, "<tr>\n" +
+ "<td><h1>%.4f%%</h1></td>\n" +
+ "<td>%s %s</td>\n" +
+ "<td><img src='%s' width='100%%' /><br />%s</td>\n" +
+ "</tr>\n",
+ l.Avgconf, l.OcrName, l.Name, fn, l.Text)
+ if err != nil {
+ return err
+ }
+ }
+ _, err = fmt.Fprintf(f, "</table>\n</body></html>\n")
+ if err != nil {
+ return err
+ }
+
+ return nil
+}
diff --git a/avg-lines/main.go b/avg-lines/main.go
new file mode 100644
index 0000000..a953598
--- /dev/null
+++ b/avg-lines/main.go
@@ -0,0 +1,69 @@
+package main
+
+import (
+ "flag"
+ "fmt"
+ "log"
+ "os"
+ "path/filepath"
+ "sort"
+
+ "rescribe.xyz/go.git/lib/line"
+ "rescribe.xyz/go.git/lib/hocr"
+ "rescribe.xyz/go.git/lib/prob"
+)
+
+func main() {
+ flag.Usage = func() {
+ fmt.Fprintf(os.Stderr, "Usage: avg-lines [-html dir] [-nosort] [prob1] [hocr1] [prob2] [...]\n")
+ fmt.Fprintf(os.Stderr, "Prints a report of the average confidence for each line, sorted\n")
+ fmt.Fprintf(os.Stderr, "from worst to best.\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\n")
+ flag.PrintDefaults()
+ }
+ var html = flag.String("html", "", "Output in html format to the specified directory")
+ var nosort = flag.Bool("nosort", false, "Don't sort lines by confidence")
+ flag.Parse()
+ if flag.NArg() < 1 {
+ flag.Usage()
+ os.Exit(1)
+ }
+
+ 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 {
+ lines = append(lines, l)
+ }
+ }
+
+ if *nosort == false {
+ sort.Sort(lines)
+ }
+
+ if *html == "" {
+ for _, l := range lines {
+ fmt.Printf("%s %s: %.2f%%\n", l.OcrName, l.Name, l.Avgconf)
+ }
+ } else {
+ htmlout(*html, lines)
+ }
+}
--
cgit v1.2.1-24-ge1ad